Category: 3. Business

  • FinTech adoption for ESG integration through robo advisors, personalization, and perceived trust

    FinTech adoption for ESG integration through robo advisors, personalization, and perceived trust

  • Sánchez, M. A. A multi-level perspective on financial technology transitions. Technol. Forecast. Soc. Chang. 181, 121766 (2022).

    Google Scholar 

  • Arner, D. W. et al. Sustainability, fintech and financial inclusion. Eur. Bus. Organ. Law Rev. 21, 7–35 (2020).

    Google Scholar 

  • Macchiavello, E. & Siri, M. Sustainable finance and fintech: can technology contribute to achieving environmental goals? A preliminary assessment of green fintech and sustainable digital finance. Eur. Co. Financial Law Rev. 19 (1), 128–174 (2022).

    Google Scholar 

  • Shanmuganathan, M. Behavioural finance in an era of artificial intelligence: longitudinal case study of robo-advisors in investment decisions. J. Behav. Experimental Finance. 27, 100297 (2020).

    Google Scholar 

  • Alkaraan, F. et al. Sustainable strategic investment decision-making practices in UK companies: the influence of governance mechanisms on synergy between industry 4.0 and circular economy. Technol. Forecast. Soc. Chang. 187, 122187 (2023).

    Google Scholar 

  • Tan, G. K. S. Robo-advisors and the financialization of Lay investors. Geoforum 117, 46–60 (2020).

    PubMed 
    PubMed Central 

    Google Scholar 

  • Bei, J. & Wang, C. Renewable energy resources and sustainable development goals: evidence based on green finance, clean energy and environmentally friendly investment. Resour. Policy. 80, 103194 (2023).

    Google Scholar 

  • Folqué, M., Escrig-Olmedo, E., Corzo, T. & Santamaría Sustainable development and financial system: integrating ESG risks through sustainable investment strategies in a climate change context. Sustain. Dev. 29 (5), 876–890 (2021).

    Google Scholar 

  • Chen, L. & Aklikokou, A. K. Determinants of E-government adoption: testing the mediating effects of perceived usefulness and perceived ease of use. Int. J. Public. Adm. 43 (10), 850–865 (2020).

    Google Scholar 

  • Nguyen, T. P. L. et al. Factors influencing acceptance of Robo-Advisors for wealth management in Malaysia. Cogent Eng. 10 (1), 2188992 (2023).

    MathSciNet 

    Google Scholar 

  • Tan, Y. & Zhu, Z. The effect of ESG rating events on corporate green innovation in china: the mediating role of financial constraints and managers’ environmental awareness. Technol. Soc. 68, 101906 (2022).

    Google Scholar 

  • Asif, M. & Sarwar, F. Impact of customer relationship management, financial literacy and social influence on online banking adoption: the moderating role of personal innovativeness. Asia-Pacific J. Bus. Adm. 17 (3), 744–764 (2024).

    Google Scholar 

  • Kasmon, B. et al. FinTech application in Islamic social finance in Asia region: a systematic literature review. Int. J. Ethics Syst. 41 (1), 213–237 (2025).

    Google Scholar 

  • Ho, S. M., Ocasio-Velázquez, M. & Booth, C. Trust or consequences? Causal effects of perceived risk and subjective norms on cloud technology adoption. Computers Secur. 70, 581–595 (2017).

    Google Scholar 

  • Namahoot, K. S. & Laohavichien, T. Assessing the intentions to use internet banking: the role of perceived risk and trust as mediating factors. Int. J. Bank. Mark. 36 (2), 256–276 (2018).

    Google Scholar 

  • Gupta, S. & Dey, D. K. Risk perception and adoption of digital innovation in mobile stock trading. J. Consumer Behav. 23 (2), 639–654 (2024).

    Google Scholar 

  • Galeone, G., Ranaldo, S. & Fusco, A. ESG and fintech: are they connected? Res. Int. Bus. Finance. 69, 102225 (2024).

    Google Scholar 

  • Knewtson, H. S. & Rosenbaum, Z. A. Toward Understanding fintech and its industry. Managerial Finance. 46 (8), 1043–1060 (2020).

    Google Scholar 

  • Sciarelli, M. et al. Socially responsible investment strategies for the transition towards sustainable development: the importance of integrating and communicating ESG. TQM J. 33 (7), 39–56 (2021).

    Google Scholar 

  • Roh, T., Park, B. I. & Xiao, S. S. Adoption of AI-enabled Robo-advisors in fintech: simultaneous employment of UTAUT and the theory of reasoned action. J. Electron. Commer. Res. 24 (1), 29–47 (2023).

    Google Scholar 

  • Gigante, G. & Zago, A. DARQ technologies in the financial sector: artificial intelligence applications in personalized banking. Qualitative Res. Financial Markets. 15 (1), 29–57 (2023).

    Google Scholar 

  • Amnas, M. B. et al. Understanding the determinants of fintech adoption: integrating UTAUT2 with trust theoretic model. J. Risk Financial Manage. 16 (12), 505 (2023).

    Google Scholar 

  • Macchiavello, E. & Siri, M. Sustainable finance and fintech: Can technology contribute to achieving environmental goals? A preliminary assessment of ‘green fintech’and ‘sustainable digital finance’. European Company and Financial Law Review, 19 (1), 128–174 (2022).

  • Saqib, N. et al. Harnessing digital solutions for sustainable development: a quantile-based framework for designing an SDG framework for green transition. Environ. Sci. Pollut. Res. 30 (51), 110851–110868 (2023).

    Google Scholar 

  • Shan, S., Umar, M. & Mirza, N. Can robo advisors expedite carbon transitions? Evidence from automated funds. Technol. Forecast. Soc. Chang. 180, 121694 (2022).

    Google Scholar 

  • Naseer, M. M. et al. Sustainable investments in volatile times: nexus of climate change risk, ESG practices, and market volatility. Int. Rev. Financial Anal. 95, 103492 (2024).

    Google Scholar 

  • Chakraborty, D. et al. Corporate governance and investment decisions of retail investors in equity: do group affiliation and firm age matter? Managerial Auditing J. 38 (1), 1–34 (2023).

    MathSciNet 

    Google Scholar 

  • Davis, F. D. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 13, 319–340 (1989).

    Google Scholar 

  • Wang, J. S. Exploring biometric identification in fintech applications based on the modified TAM. Financial Innov. 7 (1), 42 (2021).

    Google Scholar 

  • Bagozzi, R. P. The legacy of the technology acceptance model and a proposal for a paradigm shift. J. Association Inform. Syst. 8 (4), 243–254 (2007).

    Google Scholar 

  • Schierz, P. G., Schilke, O. & Wirtz, B. W. Understanding consumer acceptance of mobile payment services: an empirical analysis. Electron. Commer. Res. Appl. 9 (3), 209–216 (2010).

    Google Scholar 

  • Pavlou, P. A. Consumer acceptance of electronic commerce: integrating trust and risk with the technology acceptance model. Int. J. Electron. Commer. 7 (3), 101–134 (2003).

    Google Scholar 

  • Hung, N. T. & Hang, V. T. T. A studying on factors affecting decision to use smart tourism applications using extended TAM. WSEAS Trans. Bus. Econ. 17, 288–299 (2020).

    Google Scholar 

  • Balaskas, S. et al. FinTech services adoption in greece: the roles of trust, government support, and technology acceptance factors. FinTech 3 (1), 83–101 (2024).

    Google Scholar 

  • Chueca, C. & Ferruz Agudo, L. Fintech and sustainability: do they affect each other? Sustainability 13 (13), 7012 (2021).

    Google Scholar 

  • Bajunaied, K., Hussin, N. & Kamarudin, S. Behavioral intention to adopt fintech services: an extension of unified theory of acceptance and use of technology. J. Open. Innovation: Technol. Market Complex. 9 (1), 100010 (2023).

    Google Scholar 

  • Belanche, D., Casaló, L. V. & Flavián, C. Artificial intelligence in fintech: Understanding robo-advisors adoption among customers. Industrial Manage. Data Syst. 119 (7), 1411–1430 (2019).

    Google Scholar 

  • Faradynawati, I. A. A. & Söderberg, I. L. Sustainable investment preferences among robo-advisor clients. Sustainability 14 (19), 12636 (2022).

    Google Scholar 

  • Au, C. D. et al. Business model of sustainable robo-advisors: empirical insights for practical implementation. Sustainability 13 (23), 13009 (2021).

    Google Scholar 

  • Khalil, M. K. & Khalil, R. Leveraging buyers’ interest in ESG investments through sustainability awareness. Sustainability 14 (21), 14278 (2022).

    Google Scholar 

  • Karwowski, M. & Raulinajtys-Grzybek, M. The application of corporate social responsibility (CSR) actions for mitigation of environmental, social, corporate governance (ESG) and reputational risk in integrated reports. Corp. Soc. Responsib. Environ. Manag. 28 (4), 1270–1284 (2021).

    Google Scholar 

  • Orgeldinger, J. Digital transformation with robo-advising and AI in asset management: A critical appraisal. J. AI Rob. Workplace Autom. 3 (2), 151–167 (2024).

    Google Scholar 

  • Bohn, J., Goldberg, L. R. & Ulucam, S. Sustainable investing from a practitioner’s viewpoint: what’s in your ESG portfolio. J. Invest. Manage. 20 (2), 14–29 (2022).

    Google Scholar 

  • Rousseau, S. et al. ESG tech: attractions and challenges for fintechs in the age of COVID-19. Bank. Finance Law Rev. 37 (1), 57–96 (2021).

    Google Scholar 

  • Jo, H. Understanding the key antecedents of users’ continuance intention in the context of smart factory. Technol. Anal. Strateg. Manag. 35 (2), 153–166 (2023).

    Google Scholar 

  • Yi, T. Z. et al. The adoption of robo-advisory among millennials in the 21st century: trust, usability and knowledge perception. Sustainability 15 (7), 6016 (2023).

    Google Scholar 

  • Asl, M. G. et al. Fintech’s impact on conventional and Islamic sustainable equities: Short-and long-term contributions of the digital financial ecosystem. Glob. Financ. J. 62, 101022 (2024).

    Google Scholar 

  • Islam, H. et al. Nexus between perception, purpose of use, technical challenges and satisfaction for mobile financial services: theory and empirical evidence from Bangladesh. Technological Sustain. 3 (2), 147–170 (2024).

    Google Scholar 

  • Sabir, A. A. et al. Consumer acceptance and adoption of AI robo-advisors in fintech industry. Mathematics 11 (6), 1311 (2023).

    Google Scholar 

  • Paetzold, F. & Busch, T. Unleashing the powerful few: sustainable investing behaviour of wealthy private investors. Organ. Environ. 27 (4), 347–367 (2014).

    Google Scholar 

  • Akhmedova, A., Vila-Brunet, N. & Mas-Machuca, M. Building trust in sharing economy platforms: trust antecedents and their configurations. Internet Res. 31 (4), 1463–1490 (2021).

    Google Scholar 

  • Ng, E. & Pan, S. L. Competitive strategies for ensuring fintech platform performance: evidence from multiple case studies. Inform. Syst. J. 34 (3), 616–641 (2024).

    Google Scholar 

  • Addy, W. A. et al. Data-driven sustainability: how fintech innovations are supporting green finance. Eng. Sci. Technol. J. 5 (3), 760–773 (2024).

    Google Scholar 

  • Al-Somali, S. A., Gholami, R. & Clegg, B. An investigation into the acceptance of online banking in Saudi Arabia. Technovation 29 (2), 130–141 (2009).

    Google Scholar 

  • Tyrväinen, O., Karjaluoto, H. & Saarijärvi, H. Personalization and hedonic motivation in creating customer experiences and loyalty in omnichannel retail. J. Retailing Consumer Serv. 57, 102233 (2020).

    Google Scholar 

  • Li, M. W., Teng, H. Y. & Chen, C. Y. Unlocking the customer engagement-brand loyalty relationship in tourism social media: the roles of brand attachment and customer trust. J. Hospitality Tourism Manage. 44, 184–192 (2020).

    Google Scholar 

  • Hassan, M. S. et al. Drivers influencing the adoption intention towards mobile fintech services: a study on the emerging Bangladesh market. Information 13 (7), 349 (2022).

    Google Scholar 

  • Davis, F. D., Bagozzi, R. P. & Warshaw, P. R. User acceptance of computer technology: A comparison of two theoretical models. Manage. Sci. 35 (8), 982–1003 (1989).

    Google Scholar 

  • Saif, M. A. et al. Determinants of the intention to adopt digital-only banks in malaysia: the extension of environmental concern. Sustainability 14 (17), 1–32 (2022).

    Google Scholar 

  • Cheng, J. C. et al. Building customer satisfaction with tour leaders: the roles of customer trust, justice perception, and Cooperation in group package tours. Asia Pac. J. Tourism Res. 22 (4), 395–407 (2017).

    Google Scholar 

  • Malär, L. et al. Emotional brand attachment and brand personality: the relative importance of the actual and the ideal self. J. Mark. 75 (4), 35–52 (2011).

    Google Scholar 

  • Samala, N. et al. Investigating the role of participation and customer-engagement with tourism brands (CETB) on social media. Acad. Mark. Stud. J. 23 (1), 1–16 (2019).

    MathSciNet 

    Google Scholar 

  • Touni, R. et al. Antecedents and an outcome of customer engagement with hotel brand community on Facebook. J. Hospitality Tourism Res. 44 (2), 278–299 (2020).

    Google Scholar 

  • Hair, J. et al. An updated and expanded assessment of PLS-SEM in information systems research. Industrial Manage. Data Syst. 117 (3), 442–458 (2017).

    Google Scholar 

  • Henseler, J., Ringle, C. M. & Sinkovics, R. R. The use of partial least squares path modeling in international marketing, in new challenges to international marketingp. 277–319 (Emerald Group Publishing Limited, 2009).

  • D’Souza, C. et al. CSR investments and innovation–Aligning and creating shared value. J. Clean. Prod. 481, 144189 (2024).

    Google Scholar 

  • Rodríguez, N. G., Pérez, M. J. S. & Gutiérrez, J. A. T. Can a good organizational climate compensate for a lack of top management commitment to new product development? J. Bus. Res. 61 (2), 118–131 (2008).

    Google Scholar 

  • Hair, J. F. et al. When to use and how to report the results of PLS-SEM. Eur. Bus. Rev. 31 (1), 2–24 (2019).

    Google Scholar 

  • Fornell, C. & Larcker, D. F. Evaluating structural equation models with unobservable variables and measurement error. J. Mark. Res. 18 (1), 39–50 (1981).

    Google Scholar 

  • Henseler, J., Ringle, C. M. & Sarstedt, M. A new criterion for assessing discriminant validity in variance-based structural equation modeling. J. Acad. Mark. Sci. 43, 115–135 (2015).

    Google Scholar 

  • Hair, J. F. et al. An assessment of the use of partial least squares structural equation modeling in marketing research. J. Acad. Mark. Sci. 40, 414–433 (2012).

    Google Scholar 

  • Tenenhaus, M. et al. PLS path modeling. Comput. Stat. Data Anal. 48 (1), 159–205 (2005).

    MathSciNet 

    Google Scholar 

  • Voorhees, C. M. et al. Discriminant validity testing in marketing: an analysis, causes for concern, and proposed remedies. J. Acad. Mark. Sci. 44, 119–134 (2016).

    Google Scholar 

  • Shahzad, M. F. et al. Effect of stakeholder pressure on environmental performance: do virtual CSR, green credit, environmental and social reputation matter? J. Environ. Manage. 368, 122223 (2024).

    PubMed 

    Google Scholar 

  • Bhatia, A. et al. Role of financial literacy and trust in perceived usefulness of robo advisory services. in international conference on information and communication technology for competitive strategies. Springer. (2023).

  • Kumar, J. & Rani, V. Financial innovation and gender dynamics: a comparative study of male and female FinTech adoption in emerging economies. In. J. Account. Inf. Ma. ahead-of-print(ahead-of-print) (2024).

  • Adeoye, O. B. et al. Integrating artificial intelligence in personalized insurance products: a pathway to enhanced customer engagement. Int. J. Manage. Entrepreneurship Res. 6 (3), 502–511 (2024).

    Google Scholar 

  • Wang, M., Cho, S. & Denton, T. The impact of personalization and compatibility with past experience on e-banking usage. Int. J. Bank. Mark. 35 (1), 45–55 (2017).

    Google Scholar 

  • Asif, M. & Sarwar, F. Customer relationship management, financial literacy, and online banking adoption: social influence moderation-insights from technology adoption. RADS J. Bus. Manage. 5 (2), 129–144 (2023).

    Google Scholar 

Continue Reading

  • Eli Lilly’s obesity pill will rival Novo Nordisk’s oral Wegovy drug

    Eli Lilly’s obesity pill will rival Novo Nordisk’s oral Wegovy drug

    A sign with the company logo sits outside of the headquarters of Eli Lilly in Indianapolis, Indiana, on March 17, 2024.

    Scott Olson | Getty Images

    Eli Lilly‘s stock is still recovering after the drugmaker released trial data earlier this month on its closely watched obesity pill that underwhelmed Wall Street.

    In a key late-stage trial, Eli Lilly’s pill, orforglipron, caused less weight loss and had higher side effects than what analysts were expecting. The pill’s efficacy also appeared to come in slightly below that of Novo Nordisk‘s oral semaglutide for obesity, which showed strong data in a separate study.

    Shares of Eli Lilly fell about 13% on the day the trial results were released, although they’re up about 12% since then.

    But some analysts say Eli Lilly’s daily pill, if approved, could still be a viable competitor in the weight loss drug space — even if it will likely be second to enter the market. It’s a highly lucrative area that is eager for more convenient options that could ease the supply shortfalls and access hurdles created by the pricey weekly injections currently dominating it.

    Analysts note that Eli Lilly’s pill could have a few advantages over the daily oral version of Novo Nordisk’s weight loss drug semaglutide, which is on track to become the first needle-free alternative for obesity to win approval in the U.S. later this year. Eli Lilly hopes to launch its pill globally “this time next year,” CEO David Ricks told CNBC in early August.

    Both drugs work by mimicking a gut hormone called GLP-1 to suppress appetite and regulate blood sugar. But while Novo Nordisk’s pill is a peptide medication, orforglipron is a small-molecule drug.

    That means Eli Lilly’s pill is absorbed more easily in the body and doesn’t require dietary restrictions like Novo Nordisk’s does. Orforglipron will also be easier to manufacture at scale, which is crucial as demand for obesity and diabetes injections outpaces supply.

    Neither company has released prices for its respective pill, but some analysts said Eli Lilly’s drug could potentially have a lower price than Novo Nordisk’s pill. That would be a notable edge, as many health plans in the U.S. still don’t cover obesity treatments.

    “It’s a little bit of an apples and oranges comparison because Novo Nordisk could have difficulty manufacturing enough of the product, given the high cost and requirements to manufacture oral semaglutide,” Leerink Partners analyst David Risinger said in an interview. 

    “Whereas Lilly plans to blanket the world with orforglipron, and very quickly it will generate dramatically more sales,” he continued. “It can launch globally in an extraordinary manner with lower prices and with no food intake consideration.”

    Goldman Sachs analysts seem to agree, based on a note in August. They forecast daily oral pills will capture 24% share — or around $22 billion — of the 2030 global weight loss drug market, which they expect to be worth $95 billion. 

    The Goldman analysts said they expect Eli Lilly’s pill to have a 60% share — or roughly $13.6 billion — of the daily oral segment of the market in 2030. They expect Novo Nordisk’s oral semaglutide to have a 21% share — or around $4 billion — of that segment. The remaining 19% slice will go to other emerging pills, the analysts said.

    The race to develop a more convenient obesity pill has been fraught, as companies such as Pfizer have had to scrap previous contenders and bring forward new ones. Novo Nordisk and Eli Lilly are also exploring other experimental oral drugs, along with a slate of other companies such as Viking Therapeutics, Structure Therapeutics, AstraZeneca and Roche. 

    In a statement, Novo Nordisk CEO Mike Doustar said “we strongly believe in the efficacy” of the oral drug. The Danish company added it will be “laser-focused on getting this product to patients without supply constraints” in the U.S. 

    Dr. Mihail “Misha” Zilbermint, director of endocrine hospitalists at Johns Hopkins Community Physicians, said it’s hard to crown a winner between Eli Lilly and Novo Nordisk without knowing how their respective pills will be priced and whether insurance will cover them. 

    “I think both of the drugs are going to be gamechangers,” he said. “When it comes to which company is going to win the game — cost is the biggest issue.”

    Weight loss, side effect comparisons

    It’s difficult to directly compare the results of separate clinical trials, especially as investors wait for Eli Lilly and Novo Nordisk to release the full data from their phase three studies.

    Eli Lilly’s ATTAIN-1 trial also followed 3,000 patients, while Novo Nordisk’s OASIS 4 study evaluated a much smaller group of roughly 300. There are currently no studies directly comparing the two drugs, a Novo Nordisk spokesperson said.

    But Novo Nordisk’s oral semaglutide appears to cause a greater level of weight loss than Eli Lilly’s pill based on the available data, said BMO Capital Markets analyst Evan Seigerman. 

    In the trial, the highest dose of Eli Lilly’s pill helped patients lose 12.4% of their body weight on average at 72 weeks. The pill’s weight loss was 11.2% when analyzing all patients regardless of discontinuations.

    Wall Street had hoped Eli Lilly’s pill would generate weight loss of around 15%, the same level as Novo Nordisk’s blockbuster weight loss injection Wegovy. Semaglutide is the active ingredient in Wegovy and its diabetes counterpart Ozempic. 

    Novo Nordisk flags flutter outside its office in Bagsvaerd, on the outskirts of Copenhagen, Denmark, on July 14, 2025.

    Tom Little | Reuters

    Meanwhile, the 25-milligram dose of Novo Nordisk’s oral semaglutide helped patients lose up to 16.6% of their weight on average at 64 weeks, according to results from the trial presented at a medical conference in 2024. That weight loss was 13.6% when the company analyzed all patients regardless of whether they stopped the drug. 

    A Novo Nordisk spokesperson added that 20% of weight loss was observed in nearly one-third of patients in the trial.

    Still, the slightly lower efficacy of Eli Lilly’s pill may not be significant enough to deter patients from taking it. 

    “For many patients, 12% is a really great number,” said Seigerman. “There’s definitely a market there” for orforglipron.

    In a note earlier this month, Bank of America analysts shared a similar sentiment. 

    “Yes, weight loss fell a bit short, but ask 100 prescribers whether this new data will really make a difference in who they’d put on orforglipron, and our belief is the vast majority would say, ‘not really,’” they wrote, referring to Eli Lilly’s trial data. 

    Some investors raised concerns about the side effects and discontinuation rates in the trial of Eli Lilly’s pill. But Seigerman said the drug’s tolerability data — how well patients tolerate it — appears to be relatively in line with that of Novo Nordisk’s oral semaglutide. 

    About 10.3% of patients who took the highest dose of Eli Lilly’s pill — 36 milligrams — discontinued treatment due to side effects, compared with around 2.6% of those who took a placebo.

    Those side effects were mainly gastrointestinal, such as nausea and vomiting, and mild to moderate in severity. An estimated 24% of those who took the highest dose of Eli Lilly’s pill reported vomiting, while 33.7% had nausea. 

    Leerink’s Risinger said he is watching to see how persistent those gastrointestinal issues are once Eli Lilly presents the full data. 

    The side effects in the trial on Novo Nordisk’s pill were mostly gastrointestinal-related: 30.9% of those who took oral semaglutide reported vomiting and 46.6% reported nausea, according to the trial results. 

    Johns Hopkins’ Zilbermint said it’s difficult for him to decide which one has a better safety and tolerability profile based on the available data. 

    Meanwhile, Seigerman pointed to a different factor “that will also matter a lot”: dietary requirements. 

    Food requirements, manufacturing, price 

    Unlike Eli Lilly’s pill, patients need to take Novo Nordisk’s oral semaglutide in the morning on an empty stomach with no more than four ounces of plain water. They’re instructed to wait 30 minutes before eating, drinking or taking other oral medicines.

    Seigerman said that could be a hurdle for some patients. 

    For example, “if you’re a parent with kids and you have to take this drug and wait half an hour before you can drink your coffee, you’re going to drive yourself crazy, especially if you have to take this every day,” he said. “I try to think about the real-world use of these drugs in a market like this. It’s going to matter.” 

    Leerink’s Risinger said oral semaglutide will also be “extremely expensive to manufacture” since it is a peptide medication, and “is likely going to have to be priced higher than orforlipgron.”

    A Novo Nordisk spokesperson said the pill will be made mostly in the U.S., and the company is excited about the potential the pill “provides millions of Americans living with obesity.”

    “Currently, all typical launch readiness activities [for the pill] are fully underway and building momentum,” the spokesperson said. They added that over the past decade, the company has invested $24 billion in the U.S. to expand manufacturing capacity and fuel research and development. That includes investments aimed at increasing manufacturing of active pharmaceutical ingredients and capacity for the final stages of production for both current and future injectable and oral products. 

    Small molecules are chemically simpler and easier to produce at scale, making them generally cheaper for companies to formulate. But it is still unclear how Eli Lilly will price orforglipron. 

    During an earnings call in August, Eli Lilly’s Ricks said the pricing will be based on the value orforglipron brings, considering health-care savings and the comorbidities it can address.

    In the note earlier this month, Goldman Sachs analysts said they expect the pill to be “priced at parity” to Eli Lilly’s tirzepatide, the active ingredient in the company’s obesity injection Zepbound and diabetes counterpart Mounjaro, which list for just over $1,000 for a month’s supply. 

    “They should be cheaper than injections because they are easier to produce. But it does not mean they will be cheaper,” Johns Hopkins’ Zilbermint said. “We just don’t know — for example, we don’t know how much went into research and development.”

    Seigerman said commercialization strategies will also be key when the pills compete on the market. 

    He questioned whether Novo Nordisk will lean into the deal it recently struck with CVS‘s pharmacy benefit manager, Caremark. Under the deal, Caremark started to prioritize Novo Nordisk’s Wegovy on its standard formularies on July 1, making that weekly injection the preferred GLP-1 drug for obesity over Zepbound. 

    But it is unclear whether oral semaglutide could receive a similar preferential status.

    Seigerman also questioned whether Eli Lilly will offer orforglipron through its direct-to-consumer pharmacy, LillyDirect. That offering bypasses insurers and pharmacy benefit managers, allowing patients to directly purchase Zepbound and some of Eli Lilly’s other drugs from the company. 

    Seigerman said he expects “a lot of nuances in the go-to-market campaign for these drugs,” adding “that’s going to matter.”

    Other competitors trail behind

    Other obesity pills are in earlier stages of development, making it difficult to directly compare them to the drugs from Eli Lilly and Novo Nordisk without longer and larger trials. 

    But so far, some experts think they pale in comparison.

    For example, Viking Therapeutics on Tuesday released mid-stage trial data that disappointed investors, sending its stock down as much as 40%. 

    Jared Holz, Mizuho health care equity strategist, said in an email Tuesday that the results on Viking’s drug “look inferior” to those of Eli Lilly’s pill “on almost all metrics.” 

    Viking’s once-daily pill helped patients lose up to 12.2% of their weight at around three months, with no plateau, which means patients could lose even more in a longer-term study.

    Holz pointed to the high rate of patients who discontinued Viking’s drug for any reason over 13 weeks, which was around 28%. Meanwhile, around a quarter of people discontinued Eli Lilly’s pill, orforglipron, for any reason over 72 weeks.

    That’s “a much longer trial and therefore [Lilly] looks far better head-to-head,” Holz said.

    Continue Reading

  • The Implementation of a New Inpatient Admission Sheet in the Department of Pediatrics at the Atbara Teaching Hospital: A Quality Improvement Project

    The Implementation of a New Inpatient Admission Sheet in the Department of Pediatrics at the Atbara Teaching Hospital: A Quality Improvement Project


    Continue Reading

  • Efficient self-healing framework for smart distribution networks

    Efficient self-healing framework for smart distribution networks

    Post-fault self-healing

    Case (A): application of self-healing methodology on the IEEE 33-bus test system using CYME software

    The self-healing methodology was applied to the IEEE 33-bus test system using CYME software, with a fault simulated in Sect. 8 due to its suitability for fault isolation and multiple reconfiguration options to restore power with minimal switch operations. The self-healing protection system responds by opening switches Sw-8 and Sw-9 to isolate the fault, resulting in a power disruption to Sects. 8 through 17, as shown in Fig. 7. This outage affects a total unserved consumer load of 815 kW.

    Network reconfiguration was effectively carried out using self-healing techniques to strategically manage switch operations, ensuring optimal load transfer and power restoration for consumers affected by the outage in Sect. 8. The optimal setup for Sect. 8 involves opening switch Sw-9 and closing the normally open tie switch Sw-35, as outlined in Table 4, for scenarios both with and without distributed generation (DG). This configuration facilitates the transfer of 815 kW through the tie line switches after fault isolation. The resulting network layouts are illustrated in Fig. 8.

    Case (B): application of the self-healing system using matlab/simulink

    In this study, the simulation framework for evaluating self-healing mechanisms in power distribution networks is developed using MATLAB/Simulink, with the IEEE 33-bus test system serving as the foundational topology. This section delineates the configuration of the simulation environment, encompassing the hardware and software infrastructure, alongside a comprehensive description of the model components, including bus configurations, load profiles, power sources, and switching mechanisms.

    Central to the simulation is the control unit model, precisely designed in MATLAB/Simulink to coordinate the operation of smart protection devices. The Simulink model represents the dynamic behavior of these devices, offering a graphical representation of their interactions within the distribution network. Key components include bus signal blocks, which emulate protective relays with configurable parameters such as overcurrent settings, time delays, and trip characteristics; circuit breaker blocks, which model the response to fault conditions through controlled opening and closing actions; and switch blocks, which simulate network reconfiguration by isolating faulty segments or redirecting power flow. The control unit block integrates decision-making algorithms, processing inputs from sensors and relays to execute coordinated actions, such as fault isolation or power restoration.

    Additionally, a measurement unit block collects and processes real-time system data, providing critical inputs for the control unit’s decision-making process. This model facilitates the simulation of diverse fault scenarios, enabling thorough analysis of device coordination, response times, and the overall efficacy of the smart protection system in enhancing the reliability and stability of the distribution network. Figure 10 shows the Simulink control unit, integrating relays, breakers, and switches.

    Fig. 10

    The configuration of the system components in the Simulink model. (a) The Simulink model with integrating components, system units, relays, breakers, and switches. (b) The bus current and voltage monitoring components in the measurement system unit. (c) The logic unit model diagram in the control system.

    The self-healing approach was implemented in MATLAB/Simulink using protective devices capable of receiving remote commands to respond rapidly to faults. In this scenario, a fault is simulated in Sect. 8, as shown in Fig. 11. When the fault occurs, the current exceeds its rated limit, prompting the monitoring unit to detect the abnormal current surge. This triggers a signal to the comparison unit, which evaluates the currents in each feeder segment against the relay settings. To isolate the fault, the comparison unit sends a signal to switch Sw-8, which opens within 0.02 s. Consequently, the power supply to Sects. 8 through 17 is disrupted, resulting in an unserved consumer load of 815 kW.

    Fig. 11
    figure 11

    The IEEE 33 bus test system is supported with 2 DGs Simulink mode.

    Figure 12 demonstrates that when implementing the self-healing process without distributed generators (DGs), the minimum voltage of 0.90 p.u. presented at bus 17, and the system losses calculated to be 144.97 kW. With the integration of DGs, the minimum voltage improved to 0.973 p.u. at bus 17, and the losses were substantially reduced to 48.6 kW as detailed in Table 4. Figure 12 presents the post-fault voltage profile, while Fig. 13 compares branch currents.

    Fig. 12
    figure 12

    The IEEE 33-bus system voltage profile with and without 2 DGs in MATLAB.

    Fig. 13
    figure 13

    Voltage profile before/after fault isolation with DG.

    Figure 13 offers the voltage profile before/after activation of the fault event in the self-healing system. The pre-fault voltage at bus 9, steady at 0.984 p.u., demonstrates the enhanced steady-state condition achieved through optimal DGs placement. Following the fault, the voltage stabilizes at 0.972 p.u. The slight voltage reduction at bus 9 is due to load redistribution and line impedance effects post-reconfiguration, yet it remains within the acceptable range of 0.95–1.05 p.u. Additionally, Fig. 13 indicates that the post-restoration minimum system voltage is 0.94 p.u. at bus 32. This robust dynamic response highlights the system’s capability to maintain voltage stability during fault conditions, reinforcing its reliability as a key feature of a smart grid.

    Table 4 The optimal switching plan after isolating the fault at sect. 8.

    Figure 14 illustrates the branch current loading in the IEEE 33-bus system under two conditions: normal operation and post-fault isolation. During normal operation, the main feeder carries a load current of 106.2 A per phase. Following a fault in section 8 and the subsequent self-healing process with fault isolation, the total recovered load current across the 33-bus system reduces to 74.6 A per phase, as detailed in Table 5.

    Fig. 14
    figure 14

    Branch currents before/after fault isolation.

    Table 5 Simulation results before and after fault isolation and service restoration using 2DGs.

    Figure 15 displays a time-domain plot of the voltage magnitude at bus 9 during a fault in Sect. 8, simulated using MATLAB/Simulink. The x-axis represents time in seconds, and the y-axis shows the voltage value in per unit (p.u.). Before the fault, the voltage was stable at approximately 0.98 p.u., reflecting the improved voltage profile due to DGs placed at buses 6 and 32.

    Fig. 15
    figure 15

    Voltage on bus-bar 9 before and after the application of the self-healing process, using DGs, during a fault occurrence.

    At the fault onset (t = 0.1 s.), the voltage drops sharply due to the disturbance. Within 0.02 s., the fault is detected and isolated by opening switch Sw-8, resulting in a transient voltage dip. The self-healing process then activates, opening switch Sw-9 and closing switch Sw-35 to restore power. At t = 0.16 s, the voltage stabilizes at approximately 0.97 p.u., confirming successful power restoration. This rapid recovery, achieved within 60 ms, highlights the effectiveness of the control unit and DG support in mitigating voltage sags and ensuring system stability. During the fault event, the self-healing approach underscores the system’s capability to maintain stability and ensure reliable power restoration. The self-healing system significantly reduced overall power losses by 76% (from 202.66 kW to 48.6 kW).

    Although direct comparisons with other studies are limited by the lack of standardized benchmarks for the IEEE 33-bus system, Table 6 compares the quantitative performance of the proposed self-healing framework based on PSO for DG placement with related works in the literature. The comparison includes voltage-profile improvement, loss reduction, and service restoration time. Additionally, a comparison between the base case load conditions before and after optimal DG placement is presented in Table 3.

    Table 6 Comparison of the proposed work with other works on a 33-bus system.

    The proposed self-healing mechanism demonstrates robust performance under varying load conditions and fault scenarios, including multi-point faults and real-time load variations, through its adaptive control strategy and optimized network reconfiguration. It adapts to different load conditions through PSO-based DG placement and network reconfiguration, ensuring voltage stability and minimizing losses.

    In fault scenarios, the system effectively handles single-point faults and is also capable of managing multi-point faults using the same control strategy. It utilizes automated switching and centralized control functions for monitoring, fault detection, isolation, and reconfiguration, all managed by a central control unit through fully controlled switches. Specifically, only three switches are used per fault: one bus-bar switch is tripped to isolate the faulty section based on a protection relay signal; the next downstream bus-bar switch is automatically tripped to isolate the faulty line; and finally, the tie switch is closed automatically to restore service to the healthy sections.

    Continue Reading

  • The Hidden Ingredients Behind AI’s Creativity

    The Hidden Ingredients Behind AI’s Creativity

    The original version of this story appeared in Quanta Magazine.

    We were once promised self-driving cars and robot maids. Instead, we’ve seen the rise of artificial intelligence systems that can beat us in chess, analyze huge reams of text, and compose sonnets. This has been one of the great surprises of the modern era: physical tasks that are easy for humans turn out to be very difficult for robots, while algorithms are increasingly able to mimic our intellect.

    Another surprise that has long perplexed researchers is those algorithms’ knack for their own, strange kind of creativity.

    Diffusion models, the backbone of image-generating tools such as DALL·E, Imagen, and Stable Diffusion, are designed to generate carbon copies of the images on which they’ve been trained. In practice, however, they seem to improvise, blending elements within images to create something new—not just nonsensical blobs of color, but coherent images with semantic meaning. This is the “paradox” behind diffusion models, said Giulio Biroli, an AI researcher and physicist at the École Normale Supérieure in Paris: “If they worked perfectly, they should just memorize,” he said. “But they don’t—they’re actually able to produce new samples.”

    To generate images, diffusion models use a process known as denoising. They convert an image into digital noise (an incoherent collection of pixels), then reassemble it. It’s like repeatedly putting a painting through a shredder until all you have left is a pile of fine dust, then patching the pieces back together. For years, researchers have wondered: If the models are just reassembling, then how does novelty come into the picture? It’s like reassembling your shredded painting into a completely new work of art.

    Now two physicists have made a startling claim: It’s the technical imperfections in the denoising process itself that leads to the creativity of diffusion models. In a paper presented at the International Conference on Machine Learning 2025, the duo developed a mathematical model of trained diffusion models to show that their so-called creativity is in fact a deterministic process—a direct, inevitable consequence of their architecture.

    By illuminating the black box of diffusion models, the new research could have big implications for future AI research—and perhaps even for our understanding of human creativity. “The real strength of the paper is that it makes very accurate predictions of something very nontrivial,” said Luca Ambrogioni, a computer scientist at Radboud University in the Netherlands.

    Bottoms Up

    Mason Kamb, a graduate student studying applied physics at Stanford University and the lead author of the new paper, has long been fascinated by morphogenesis: the processes by which living systems self-assemble.

    One way to understand the development of embryos in humans and other animals is through what’s known as a Turing pattern, named after the 20th-century mathematician Alan Turing. Turing patterns explain how groups of cells can organize themselves into distinct organs and limbs. Crucially, this coordination all takes place at a local level. There’s no CEO overseeing the trillions of cells to make sure they all conform to a final body plan. Individual cells, in other words, don’t have some finished blueprint of a body on which to base their work. They’re just taking action and making corrections in response to signals from their neighbors. This bottom-up system usually runs smoothly, but every now and then it goes awry—producing hands with extra fingers, for example.

    Continue Reading

  • Microbiome characteristics associated with lymph node metastasis in laryngeal squamous cell carcinoma

    Microbiome characteristics associated with lymph node metastasis in laryngeal squamous cell carcinoma

    Patient characteristics

    This study included 36 LSCC patients with lymph node metastasis and 72 without lymph node metastasis. The demographic characteristics of the study cohort are presented in Table 1. There were no significant differences between the two groups in terms of age, sex, smoking status, or alcohol consumption (P > 0.05). The clinical data of the patients are summarized in Supplementary Table S2.

    Table 1 Demographic and clinical characteristics of participants.

    Comparison of microbial diversity between LSCC patients with and without lymph node metastasis

    We assessed the microbial diversity in the LN + and LN- groups of LSCC across different sample types. For α-diversity, compared to LN- patients, LN + patients exhibited lower microbial richness in tumor tissues, adjacent normal tissues, and lymph node tissues, whereas the inverse trend was observed in oral rinse samples, as measured by Shannon, Simpson, Chao1, and Observed OTUs indices. However, none of these differences reached statistical significance (Fig. 1a). For β-diversity, PCoA based on Bray-Curtis distances at the genus level revealed distinct clustering between LN + and LN- groups in tumor tissues (P = 0.0004, PERMANOVA), whereas adjacent normal tissues and lymph node tissues showed no significant compositional differences (Fig. 1b). Additionally, compared to tissue samples, microbial taxa in oral rinse samples from both groups exhibited a closer clustering pattern.

    Fig. 1

    Microbial diversity in tumor tissues, adjacent normal tissues, lymph node (LN) tissues, and oral rinses samples of LSCC patients with (LN+) and without (LN-) lymph node metastasis. (a) Comparison of α-diversity between groups. P values were calculated using the Wilcoxon rank-sum test. (b) Principal coordinate analysis (PCoA) plot. P values were derived from permutational multivariate analysis of variance (PERMANOVA). P < 0.05 was considered statistically significant. NT, tumor tissues from LN- patients; MT, tumor tissues from LN + patients; NC, adjacent normal tissues from LN- patients; MC, adjacent normal tissues from LN + patients; NL, non-metastatic lymph node tissues; ML, metastatic lymph node tissues; NW, oral rinses from LN- patients; MW, oral rinses from LN + patients.

    Microbial structural discrepancy in lymph node metastasis and non-metastasis groups of LSCC

    We compared the relative abundances of the top 15 genera across different sample types (tumor tissues, adjacent normal tissues, lymph node tissues, and oral rinses) in LSCC patients with and without lymph node metastasis (Fig. 2). In tumor tissues, significant differences in microbial composition were observed between the LN + and LN- groups. Specifically, Ralstonia accounted for a substantially higher proportion in LN + tumor tissues compared to LN- tumors (0.172 ± 0.212 (mean ± SD) vs. 0.034 ± 0.079, LN + vs. LN- tumor tissues), whereas Fusobacterium was more prevalent in LN- tumors (0.044 ± 0.077 vs. 0.138 ± 0.169). In contrast, adjacent normal tissues and lymph node tissues exhibited similar microbial profiles between groups, with Ralstonia remaining the most abundant genus in both sample types, regardless of metastatic statu. For oral rinse samples, the overall microbial composition was more similar between LN + and LN- patients. Streptococcus, Neisseria, and Prevotella_7 were consistently the top three abundant genera in both groups. Full statistical data, including percentage distributions and P values for all genera across comparative groups (Supplementary Table S3).

    Fig. 2
    figure 2

    The relative abundance of the top 15 genera in tumor tissue, adjacent normal tissue, lymph node (LN) tissue, and oral rinse samples. NT, tumor tissues from LN- patients; MT, tumor tissues from LN + patients; NC, adjacent normal tissues from LN- patients; MC, adjacent normal tissues from LN + patients; NL, non-metastatic lymph node tissues; ML, metastatic lymph node tissues; NW, oral rinses from LN- patients; MW, oral rinses from LN + patients.

    LEfSe analysis identified differentially abundant genera between LN + and LN- groups across various sample types, with heatmaps illustrating relative abundance patterns (Fig. 3). In tumor tissues, the relative abundance of Ralstonia, Methyloversatilis, Delftia, Lactobacillus, and Methylobacterium-Methylorubrum was significantly higher in the LN + group than in the LN- group. Conversely, Fusobacterium, Peptococcus, Sneathia, Moraxella, and [Eubacterium] saphenum group were significantly less abundant in LN + tumor tissues. Metastatic and non-metastatic lymph nodes also exhibited distinct microbial profiles. Bdellovibrio, 1174-901-12, and Ramlibacter were significantly enriched in metastatic lymph nodes, whereas Roseburia, TM7x, Coriobacteriaceae UCG-002, [Eubacterium] nodatum group, Bifidobacterium, and Monoglobus were more abundant in non-metastatic lymph nodes. For oral rinse samples, several genera displayed significant differences between LN + and LN- patients. In LN + patients, the relative abundance of Corynebacterium, Selenomonas, Centipeda, Peptoniphilus, F0058, and Lawsonella was elevated, while [Eubacterium] nodatum group and Oceanivirga were reduced.

    Fig. 3
    figure 3

    Differential microbial composition between LN + and LN- groups in LSCC patients across tumor tissue, adjacent normal tissue, lymph node tissue, and oral rinse samples. A linear discriminant analysis (LDA) score threshold of |LDA Score| ≥ 2 was used to identify genera with significant differences between LN + and LN − LSCC patients across different sample types. The bar plot displays these differentially abundant genera, while the heatmap illustrates their relative abundance. NT, tumor tissues from LN- patients; MT, tumor tissues from LN + patients; NC, adjacent normal tissues from LN- patients; MC, adjacent normal tissues from LN + patients; NL, non-metastatic lymph node tissues; ML, metastatic lymph node tissues; NW, oral rinses from LN- patients; MW, oral rinses from LN + patients.

    Venn diagram analysis of microbial community distribution revealed shared and unique microbial features across different sample types in LSCC (Fig. 4a). At the genus level, 227 genera were universally present in all four sample types, with no genus exclusively unique to any single sample type. Additionally, 29 genera, including 1174-901-12, Pandoraea, Hoeflea, and Sphingobium, were shared among tumor tissues, adjacent normal tissues, and lymph node tissues but were absent in oral rinse samples. And 7 genera (Pseudoalteromonas, Comamonas, Oceanivirga, CL500-29 marine group, Citrobacter, F0332, and Conservatibacter) were detected in tumor tissues, lymph node tissues, and oral rinse samples, but not in adjacent normal tissue. Notably, all genera identified in lymph node tissues were concurrently observed in tumor tissues.

    Fig. 4
    figure 4

    Venn diagrams illustrate the distribution of microbial communities in tumor tissue, adjacent normal tissue, lymph node tissue, and oral rinse samples. (a) Overlapping and unique genera across sample types are presented, with a heatmap depicting the relative abundances of genera. (b) Overlapping and distinct distributions of differentially abundant genera across sample types are shown, further visualized by a heatmap. NT, tumor tissues from LN- patients; MT, tumor tissues from LN + patients; NC, adjacent normal tissues from LN- patients; MC, adjacent normal tissues from LN + patients; NL, non-metastatic lymph node tissues; ML, metastatic lymph node tissues; NW, oral rinses from LN- patients; MW, oral rinses from LN + patients.

    We then examined the distribution of genera that exhibited significant differences in relative abundance between the LN + and LN- groups across different sample types (Fig. 4b). No genera exhibited significant abundance differences in all four sample types. However, 17 genera (Ralstonia, Bulleidia, Fusobacterium, etc.) showed differential abundance exclusively in tumor tissues between the LN + and LN − groups. Distinctly differentially abundant genera were also identified in adjacent normal tissues, lymph nodes, and oral rinses. Furthermore, [Eubacterium] nodatum group was shared between lymph node and oral rinse samples, showing consistent enrichment in LN- groups for both sample types.

    Functional analysis of metastatic and non-metastatic LSCC patients

    The relative abundances of the top 20 predicted metabolic pathways were compared between specimens from the LN + and LN- groups (Fig. 5). Tumor tissues were enriched in pathways related to nucleotide metabolism (e.g., UMP biosynthesis I, superpathway of adenosine nucleotides de novo biosynthesis II), carbohydrate utilization (e.g., Calvin-Benson-Bassham cycle, pentose phosphate pathway [non − oxidative branch] I), and lipid biosynthesis (e.g., phosphatidylglycerol biosynthesis II [non − plastidic], CDP − diacylglycerol biosynthesis I). In contrast, adjacent normal tissues and lymph node tissues exhibited more similar functional profiles, characterized by pathways associated with energy metabolism, fatty acid synthesis, and amino acid biosynthesis, such as aerobic respiration I (cytochrome c), palmitate biosynthesis (type II fatty acid synthase) and superpathway of branched chain amino acid biosynthesis. Oral rinse samples displayed a distinct functional profile, including pathways involving not only nucleotide and carbohydrate metabolism, but also peptidoglycan biosynthesis (peptidoglycan biosynthesis III). In addition, inter-individual variability in pathway distribution within each group was relatively low across all sample types, indicating functional consistency among patients with the same lymph node metastasis status.

    Fig. 5
    figure 5figure 5

    The relative abundance of the top 20 predicted microbial metabolic pathways in tumor tissue, adjacent normal tissue, lymph node (LN) tissue, and oral rinse samples. NT, tumor tissues from LN- patients; MT, tumor tissues from LN + patients; NC, adjacent normal tissues from LN- patients; MC, adjacent normal tissues from LN + patients; NL, non-metastatic lymph node tissues; ML, metastatic lymph node tissues; NW, oral rinses from LN- patients; MW, oral rinses from LN + patients.

    LEfSe analysis further revealed differentially abundant metabolic pathways between LN + and LN- groups across the four sample types (Fig. 6). In tumor tissues, pathways related to aerobic energy production (e.g., aerobic respiration I [cytochrome c]) and ubiquinol biosynthesis were enriched in LN + tumor tissues, while LN- tumors were characterized by pathways associated with basic biosynthetic functions, including pyruvate fermentation to acetate and lactate II and superpathway of L-aspartate and L-asparagine biosynthesis. In metastatic lymph node tissues, mycolate biosynthesis, which is a pathway associated with bacterial pathogenicity, was significantly enriched. And in oral rinse samples, the LN + group showed increased abundance of pathways involved in bacterial cell wall synthesis, such as poly (glycerol phosphate) wall teichoic acid biosynthesis and peptidoglycan biosynthesis II (staphylococci).

    Fig. 6
    figure 6

    Differential predicted microbial metabolic pathways between LN + and LN- groups in LSCC patients across tumor tissue, adjacent normal tissue, lymph node tissue, and oral rinse samples. A linear discriminant analysis (LDA) score threshold of |LDA Score| ≥ 1.5 was used to identify pathways with significant differences between LN + and LN- LSCC patients across different sample types. The bar plot displays these differentially abundant pathways, while the heatmap illustrates their relative abundance. Only pathways with LDA > 2.5 are shown in tumor tissues due to the large number of differentially abundant functions. NT, tumor tissues from LN- patients; MT, tumor tissues from LN + patients; NC, adjacent normal tissues from LN- patients; MC, adjacent normal tissues from LN + patients; NL, non-metastatic lymph node tissues; ML, metastatic lymph node tissues; NW, oral rinses from LN- patients; MW, oral rinses from LN + patients.

    Microbial classification models for stratifying LSCC patients by lymph node metastasis status

    To evaluate the potential of microbial features in distinguishing LSCC patients with and without lymph node metastasis, we developed random forest classifiers using microbial genera from tumor tissues, lymph node tissues, and oral rinses. For each sample type, 25 key genera were selected as classification features (Fig. 7a). The classifier based on lymph node tissues achieved an AUC of 84.31% (95% confidence interval [CI]: 81.76% − 86.85%), followed by tumor tissues (AUC = 84.11%, 95% CI: 81.75% − 86.46%) and oral rinses (AUC = 79.88%, 95% CI: 77.09% − 83.11%) (Fig. 7b). To evaluate biological relevance, PERMANOVA analyses were applied to the 25 key genera from each sample type. Significant differences in microbial community structure between LN + and LN- groups were identified in tumor tissues (P = 0.0001), lymph node tissues (P = 0.0006), and oral rinses (P = 0.005) (Supplementary Figure S2).

    Fig. 7
    figure 7

    Identification of microbial biomarkers for lymph node metastasis in LSCC patients by random forest models. (a) Mean Decrease Accuracy (MDA) values from random forest models based on 25 selected microbial biomarkers in tumor tissues, lymph nodes, and oral rinses. (b) Receiver operating characteristic (ROC) curves and area under the curve (AUC) values for tumor tissues, lymph nodes, and oral rinses. (c) MDA values of 17 tumor-specific differentially abundant genera in tumor tissues and lymph nodes. (d) ROC curves and AUC values based on the 17 tumor-specific differentially abundant genera in tumor tissues and lymph nodes.

    We further explored the classification potential of 17 tumor-specific differentially abundant genera (Ralstonia, Bulleidia, Fusobacterium, etc.) between the LN + and LN- groups (Fig. 7c). A classifier trained on these 17 genera effectively distinguished LN + from LN- patients in tumor tissues (AUC = 84.11%, 95% CI: 81.75% − 86.46%) but showed moderate performance in lymph node tissues (AUC = 63.15%, 95% CI: 57.59% − 68.71%) (Fig. 7d).

    Continue Reading

  • Aboveground biomass estimation using multimodal remote sensing observations and machine learning in mixed temperate forest

    Aboveground biomass estimation using multimodal remote sensing observations and machine learning in mixed temperate forest

  • Dixon, R. K. et al. Carbon pools and flux of global forest ecosystems. Science 263(5144), 185–190 (1994).

    PubMed 

    Google Scholar 

  • Baccini, A. G. S. J. et al. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Nature climate change 2(3), 182–185 (2012).

    Google Scholar 

  • Zaki, N. A., Mohd, Z. A., Latif & Mohd Zainee Zainal. and. Predicting above-ground biomass and carbon stocks by using geographically weighted regression (GWR). In 38th Asian Conf Remote Sens–Sp Appl Touching Hum Lives, ACRS  (2017).

  • Oswalt, S. N., Brad Smith, W., Miles, P. D., Scott, A. & Pugh Forest resources of the United States. In General Technical Report-US Department of Agriculture, Forest Service. Forest Service (2019). (2017).

  • FS-130 & Update, R. Forests of Connecticut. (2016).

  • Wang, L. P., Basu, S. & Zhang, Z. M. Direct and indirect methods for calculating thermal emission from layered structures with nonuniform temperatures. 072701. (2011).

  • Shi, L. & Liu, S. Methods of estimating forest biomass: A review. Biomass Volume Estimation Valorization Energy. 10, 65733 (2017).

    Google Scholar 

  • Wang, M., Im, J., Zhao, Y. & Zhen, Z. Multi-Platform lidar for Non-Destructive individual aboveground biomass Estimation for Changbai larch (Larix olgensis Henry) using a hierarchical bayesian approach. Remote Sens. 14 (17), 4361 (2022).

    Google Scholar 

  • Jenkins, J. C., Chojnacky, D. C., Heath, L. S. & Birdsey, R. A. National-scale biomass estimators for united States tree species. For. Sci. 49 (1), 12–35 (2003).

    Google Scholar 

  • Somogyi, Z. et al. Indirect methods of large-scale forest biomass Estimation. European J. For. Research. 126, 197–207 (2007).

    Google Scholar 

  • Smith, W. B. Forest inventory and analysis: a National inventory and monitoring program. Environ. Pollut. 116, S233–S242 (2002).

    PubMed 

    Google Scholar 

  • Woudenberg, S. W. et al. And Karen L. Waddell. The Forest Inventory and Analysis Database: Database Description and User’s Manual Version 4.0 for Phase 2 (United States Department of Agriculture, Forest Service, Rocky Mountain Research Station, 2010).

  • Tamiminia, H., Salehi, B., Mahdianpari, M. & Goulden, T. State-wide forest canopy height and aboveground biomass map for new York with 10 m resolution, integrating GEDI, Sentinel-1, and Sentinel-2 data. Ecol. Inf. 79, 102404 (2024).

    Google Scholar 

  • Huang, H., Liu, C., Wang, X., Zhou, X. & Gong, P. Integration of multi-resource remotely sensed data and allometric models for forest aboveground biomass Estimation in China. Remote Sens. Environ. 221, 225–234 (2019).

    Google Scholar 

  • Li, C., Li, Y. & Li, M. Improving Forest aboveground biomass (AGB) estimation by incorporating crown density and using landsat 8 OLI images of a subtropical forest in Western Hunan in Central China. Forests 10(2), 104 (2019).

    Google Scholar 

  • Johnson, K. D. et al. Integrating forest inventory and analysis data into a LIDAR-based carbon monitoring system. Carbon Balance Manag. 9, 1–11 (2014).

    Google Scholar 

  • Johnson, K. D. et al. Integrating LIDAR and forest inventories to fill the trees outside forests data gap. Environ. Monit. Assess. 187, 1–8 (2015).

    Google Scholar 

  • Johnson, L. K. et al. Fine-resolution landscape-scale biomass mapping using a Spatiotemporal patchwork of lidar coverages. Int. J. Appl. Earth Obs. Geoinf. 114, 103059 (2022).

    Google Scholar 

  • Hu, T. et al. Mapping global forest aboveground biomass with spaceborne lidar, optical imagery, and forest inventory data. Remote Sens. 8 (7), 565 (2016).

    Google Scholar 

  • Hudak, A. T. et al. A carbon monitoring system for mapping regional, annual aboveground biomass across the Northwestern USA. Environmental Res. Letters. 15 (9), 095003 (2020).

    Google Scholar 

  • Tang, H. et al. and G. C. Hurtt. Lidar derived biomass, canopy height, and cover for new England region, USA. ORNL DAAC (2021).

  • Chen, H. et al. Mapping forest aboveground biomass with MODIS and Fengyun-3 C VIRR imageries in Yunnan province, Southwest China using linear regression, K-Nearest neighbor and random forest. Remote Sens. 14 (21), 5456 (2022).

    Google Scholar 

  • Dubayah, R. O. et al. Estimation of tropical forest height and biomass dynamics using lidar remote sensing at La selva, Costa Rica. J. Geophys. Res. Biogeosci. 115, (G2) (2010).

  • Ehlers, D. et al. Mapping forest aboveground biomass using multisource remotely sensed data. Remote Sens. 14 (5), 1115 (2022).

    Google Scholar 

  • Zheng, D., Heath, L. S. & Ducey, M. J. Spatial distribution of forest aboveground biomass estimated from remote sensing and forest inventory data in new england, USA. J. Appl. Remote Sens. 2 (1), 021502 (2008).

    Google Scholar 

  • Mancini, F. et al. An integrated procedure to assess the stability of coastal Rocky cliffs: from UAV close-range photogrammetry to Geomechanical finite element modeling. Remote Sens. 9 (12), 1235 (2017).

    Google Scholar 

  • Sheridan, R. D. et al. Modeling forest aboveground biomass and volume using airborne lidar metrics and forest inventory and analysis data in the Pacific Northwest. Remote Sens. 7 (1), 229–255 (2014).

    Google Scholar 

  • Urbazaev, M. et al. Estimation of forest aboveground biomass and uncertainties by integration of field measurements, airborne lidar, and SAR and optical satellite data in Mexico. Carbon Balance Manag. 13, 1–20 (2018).

    Google Scholar 

  • Duncanson, L. et al. Implications of allometric model selection for county-level biomass mapping. Carbon Balance Manag. 12, 1–11 (2017).

    Google Scholar 

  • Haralick, R. M., Shanmugam, K. & Its’ Hak Dinstein. Textural features for image classification. IEEE Trans. Syst. Man. Cybernetics. 6, 610–621 (1973).

    Google Scholar 

  • Csillik, O., Kumar, P., Mascaro, J., O’Shea, T. & Asner, G. P. Monitoring tropical forest carbon stocks and emissions using planet satellite data. Sci. Rep. 9 (1), 1–12 (2019).

    Google Scholar 

  • Nandy, S., Srinet, R. & Padalia, Hitendra. Mapping forest height and aboveground biomass by integrating ICESat-2, Sentinel‐1 and Sentinel‐2 data using Random Forest algorithm in northwest Himalayan foothills of India. Geophysical Research Letters 48(14), e2021GL093799 (2021).

    Google Scholar 

  • Naik, P., Dalponte, M. & Lorenzo Bruzzone. Prediction of forest aboveground biomass using multitemporal multispectral remote sensing data. Remote Sens. 13 (7), 1282 (2021).

    Google Scholar 

  • Luo, P., Liao, J. & Shen, G. Combining spectral and texture features for estimating leaf area index and biomass of maize using Sentinel-1/2, and Landsat-8 data. IEEE Access. 8, 53614–53626 (2020).

    Google Scholar 

  • Gao, Y. et al. Comparative analysis of modeling algorithms for forest aboveground biomass Estimation in a subtropical region. Remote Sens. 10 (4), 627 (2018).

    Google Scholar 

  • Breiman, L. Random forests. Mach. Learn. 45, 5–32 (2001).

    Google Scholar 

  • Han, S., Williamson, B. D. & Youyi Fong. Improving random forest predictions in small datasets from two-phase sampling designs. BMC Med. Inf. Decis. Mak. 21, 1–9 (2021).

    Google Scholar 

  • Sivasankar, T., Lone, J. M., Sarma, K. K., Qadir, A. & Raju, P. L. N. Estimation of above ground biomass using support vector. Vietnam Journal of Earth Sciences 41(2), 95–104 (2013).

    Google Scholar 

  • Wolpert, D. H. Stacked generalization. Neural Netw. 5 (2), 241–259 (1992).

    Google Scholar 

  • Ma, L. et al. Mapping US forest biomass using nationwide forest inventory data and moderate resolution information. Remote sensing of Environment 112(4), 1658–1677 (2022).

    Google Scholar 

  • Menlove, J. & Healey, S. CMS: Forest Aboveground Biomass from FIA Plots across the Conterminous USA 2009–2019 (ORNL DAAC, 2021).

  • Fox, E. W., Jay, M., Ver Hoef, Anthony, R. & Olsen Comparing Spatial regression to random forests for large environmental data sets. PloS ONE. 15 (3), e0229509 (2020).

    PubMed 
    PubMed Central 

    Google Scholar 

  • Torre-Tojal, L., Bastarrika, A. & Boyano, A. Above-ground biomass estimation from LiDAR data using random forest algorithms. J. Comput. Sci. 58, 101517 (2022).

  • Luna, Soriano et al. Determinants of above-ground biomass and its spatial variability in a temperate forest managed for timber production. Forests 9(8), 490 (2018).

    Google Scholar 

  • Riemann, R., Wilson, B. T., Lister, A. & Parks, S. An effective assessment protocol for continuous Geospatial datasets of forest characteristics using USFS forest inventory and analysis (FIA) data. Remote Sens. Environ. 114 (10), 2337–2352 (2010).

    Google Scholar 

  • Food and Agriculture Organization of the United Nations (FAO). Global Forest Resources Assessment 2020: Main Report (FAO, 2020).

  • EPA. Level III and IV Ecoregions of the Continental United States. U.S. Environmental Protection Agency. (Accessed 15 May 2025). https://www.epa.gov/eco-research/ecoregions (2013).

  • Liu, Z., Luong, P., Boley, M. & Schmidt, D. F. Improving random forests by smoothing. https://arXiv.org/abs/2505.06852. (2025).

  • USDA Forest Service. Forests of Connecticut, 2020. Resource Update FS-334. Madison 2. https://doi.org/10.2737/FS-RU-334 (U.S. Department of Agriculture, Forest Service, 2021).

  • Brand, G. J., Nelson, M. D., Wendt, D. G. & Kevin, K. Nimerfro. The Hexagon/panel System for Selecting FIA Plots Under an Annual Inventory ( USFS FIA research, 2003).

  • Lu, D. et al. A survey of remote sensing-based aboveground biomass Estimation methods in forest ecosystems. Int. J. Digit. Earth. 9 (1), 63–105 (2016).

    Google Scholar 

  • Tinkham, W. T. et al. Applications of the united States forest inventory and analysis dataset: a review and future directions. Can. J. For. Res. 48 (11), 1251–1268 (2018).

    Google Scholar 

  • Burkman, B. Forest inventory and analysis: sampling and plot design. FIA Fact. Sheet Ser. (2005).

  • Butler, B. J. & Connecticut Forests of 2017. Resource Update FS-159. Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 3. https://doi.org/10.2737/FS-RU-159 (2018).

  • Hoppus, M. and Andrew Lister. The status of accurately locating forest inventory and analysis plots using the global positioning system. In Proceedings of the Seventh Annual Forest Inventory and Analysis Symposium, Portland, OR, USA 36, 179184. (2005).

  • Lister, A. et al., Strategies for preserving owner privacy in the national information management system of the USDA Forest Service’s Forest Inventory and Analysis unit. United States department of agriculture forest service general technical report NC 352 163 (2005).

  • Woodall, C. W., Linda, S., Heath, G. M., Domke & Nichols, M. C. Methods and equations for estimating aboveground volume, biomass, and carbon for trees in the US forest inventory. Gen. Tech. Rep. NRS-88. Newtown Square, PA: US Department of Agriculture, Forest Service, Northern Research Station 30 (2011).

  • Burrill, E. A. et al. The Forest Inventory and Analysis Database: Database Description and User Guide Version 9.0.1 for Phase 2. U.S. Department of Agriculture, Forest Service. 1026. (Accessed 03 March 2022). https://research.fs.usda.gov/programs/fia#data-and-tools (2021).

  • Chen, Q., Laurin, G. V. & Valentini, R. Uncertainty of remotely sensed aboveground biomass over an African tropical forest: propagating errors from trees to plots to pixels. Remote Sens. Environ. 160, 134–143 (2015).

    Google Scholar 

  • Chen, Q., Laurin, G. V., Battles, J. J. & Saah, D. Integration of airborne lidar and vegetation types derived from aerial photography for mapping aboveground live biomass. Remote Sens. Environ. 121, 108–117 (2012).

    Google Scholar 

  • CT department of energy and environmental protection. (n.d.). Connecticut environmental conditions online. Connecticut Environmental Conditions Online Maps & Geospatial Data for Everyone. https://cteco.uconn.edu/guides/Soils.htm (2024).

  • Shao, G. et al. Improving Lidar-based aboveground biomass Estimation of temperate hardwood forests with varying site productivity. Remote Sens. Environ. 204, 872–882 (2018).

    Google Scholar 

  • McPherson, E., Gregory, Natalie, S., van Doorn & Peper, P. J. Urban tree database and allometric equations. Gen. Tech. Rep. PSW-GTR-253. Albany, CA: US department of agriculture, forest service. Pac. Southwest. Res. Stn. 86, 253 (2016).

    Google Scholar 

  • Hayashi, M., Saigusa, N., Yamagata, Y. & Hirano, T. Regional forest biomass Estimation using icesat/glas spaceborne lidar over Borneo. Carbon Manag. 6 (1–2), 19–33 (2015).

    Google Scholar 

  • Chen, L., Ren, C., Zhang, B., Wang, Z. & Xi, Y. Estimation of forest above-ground biomass by geographically weighted regression and machine learning with Sentinel imagery. Forests 9 (10), 582 (2018).

    Google Scholar 

  • Bright, B. C., Hicke, J. A. & Hudak, A. T. Estimating aboveground carbon stocks of a forest affected by mountain pine beetle in Idaho using lidar and multispectral imagery. Remote Sens. Environ. 124, 270–281 (2012).

    Google Scholar 

  • Pandit, S., Tsuyuki, S. & Dube, T. Estimating above-ground biomass in sub-tropical buffer zone community forests, nepal, using Sentinel 2 data. Remote Sens. 10 (4), 601 (2018).

    Google Scholar 

  • Moradi, F., Darvishsefat, A. A., Pourrahmati, M. R. & Deljouei, A. Stelian Alexandru Borz, Estimating aboveground biomass in dense Hyrcanian forests by the use of Sentinel-2 data. Forests 13(1), 104 (2022).

    Google Scholar 

  • Parent, J. R., Arthur, J., Gold, E., Vogler & Kelly Addy Lowder Guiding decisions on the future of dams: A GIS database characterizing ecological and social considerations of dam decisions. J. Environ. Manage. 351, 119683 (2024).

    PubMed 

    Google Scholar 

  • Huete, A. R. A soil-adjusted vegetation index (SAVI). Remote sensing of environment 25(3), 295–309 (1988).

    Google Scholar 

  • Mohammadpour, P., Viegas, D. X. & Carlos Viegas. Vegetation mapping with random forest using Sentinel 2 and GLCM texture feature – A case study for lousã region, Portugal. Remote Sens. 14 (18), 4585 (2022).

    Google Scholar 

  • Genuer, R., Poggi, J. M. & Tuleau-Malot, C. Variable selection using random forests. Pattern Recognit. Lett. 31 (14), 2225–2236 (2010).

    Google Scholar 

  • Wang, Y., Zhang, X. & Guo, Z. Estimation of tree height and aboveground biomass of coniferous forests in North China using stereo ZY-3, multispectral Sentinel-2, and DEM data. Ecol. Ind. 126, 107645 (2021).

    Google Scholar 

  • Breiman, L. Classification and Regression Trees (Routledge, 2017).

  • Wongchai, W., Onsree, T., Sukkam, N., Promwungkwa, A. & Nakorn Tippayawong. Machine learning models for estimating above ground biomass of fast-growing trees. Expert Syst. Appl. 199, 117186 (2022).

    Google Scholar 

  • Biau, G. Analysis of a random forests model. J. Mach. Learn. Res. 13 (1), 1063–1095 (2012).

    MathSciNet 

    Google Scholar 

  • Sekulić, A., Kilibarda, M., Heuvelink, G. B. M., Nikolić, M. & Branislav Bajat. Random forest Spatial interpolation. Remote Sens. 12 (10), 1687 (2020).

    Google Scholar 

  • Liaw, A. & Matthew Wiener. Classification and regression by randomforest. R News. 2 (3), 18–22 (2002).

    Google Scholar 

  • Dewi, C. & Rung-Ching Chen Random forest and support vector machine on features selection for regression analysis. Int. J. Innov. Comput. Inf. Control. 15 (6), 2027–2037 (2019).

    Google Scholar 

  • Xu, D., Wang, H., Xu, W., Luan, Z. & Xu, X. LiDAR applications to estimate forest biomass at individual tree scale: Opportunities, challenges and future perspectives. Forests 12(5), 550 (2021).

    Google Scholar 

  • Hong, Y. et al. Combining multisource data and machine learning approaches for multiscale Estimation of forest biomass. Forests 14 (11), 2248 (2023).

    Google Scholar 

  • Tang, Z., Xia, X., Huang, Y., Lu, Y. & Guo, Z. Estimation of National forest aboveground biomass from multi-source remotely sensed dataset with machine learning algorithms in China. Remote Sens. 14 (21), 5487 (2022).

    Google Scholar 

  • U.S. Forest Service. National Scale Volume and Biomass Estimators (NSVB). Forest Inventory and Analysis Program. (Accessed 27 October 2024). https://research.fs.usda.gov/programs/fia/nsvb

  • Asner, G. P. et al. James Jacobson, Ty Kennedy-Bowdoin et al. High-resolution forest carbon stocks and emissions in the Amazon. Proc. Natl. Acad. Sci. 107 (38), 16738–16742. https://doi.org/10.1073/pnas.1004875107 (2010).

  • Shorten, C. & Khoshgoftaar, T. M. A survey on image data augmentation for deep learning. J. Big Data. 6 (1), 1–48 (2019).

    Google Scholar 

  • Lu, Y. et al. Machine learning for synthetic data generation: a review. (2023). arXiv preprint arXiv:2302.04062.

  • Qadeer, A., Shakir, M., Wang, L. & Talha, S. M. Evaluating machine learning approaches for aboveground biomass prediction in fragmented high-elevated forests using multi-sensor satellite data. Remote Sens. Applications: Soc. Environ. 36, 101291 (2024).

    Google Scholar 

  • White, J. et al. Coops. A model development and application guide for generating an enhanced forest inventory using airborne laser scanning data and an area-based approach. (2017).

  • Friedman, J. H., Bogdan, E. & Popescu Predictive learning via rule ensembles. 916–954. (2008).

  • Riggins, J. J., Tullis, J. A. & Stephen, F. M. Per-segment aboveground forest biomass Estimation using LIDAR-derived height percentile statistics. GIScience Remote Sens. 46 (2), 232–248 (2009).

    Google Scholar 

  • Holmgren, J. Prediction of tree height, basal area and stem volume in forest stands using airborne laser scanning. Scand. J. For. Res. 19 (6), 543–553 (2004).

    Google Scholar 

  • Lim, K. S. & Treitz, P. M. Estimation of above ground forest biomass from airborne discrete return laser scanner data using canopy-based quantile estimators. Scand. J. For. Res. 19 (6), 558–570 (2004).

    Google Scholar 

  • Næsset, E. Airborne laser scanning as a method in operational forest inventory: status of accuracy assessments accomplished in Scandinavia. Scand. J. For. Res. 22 (5), 433–442 (2007).

    Google Scholar 

  • Chen, L., Wang, Y., Ren, C., Zhang, B. & Wang, Z. Optimal combination of predictors and algorithms for forest above-ground biomass mapping from Sentinel and SRTM data. Remote Sens. 11 (4), 414 (2019).

    Google Scholar 

  • Wai, P., Su, H. & Li, M. Estimating aboveground biomass of two different forest types in Myanmar from sentinel-2 data with machine learning and Geostatistical algorithms. Remote Sens. 14 (9), 2146 (2022).

    Google Scholar 

  • Dang, A. T. N. et al. Forest aboveground biomass Estimation using machine learning regression algorithm in Yok don National park, Vietnam. Ecol. Inf. 50, 24–32 (2019).

    Google Scholar 

  • Huete, A. et al. Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens. Environ. 83 (1–2), 195–213 (2002).

    Google Scholar 

  • Liu, Z., Luong, P., Boley, M., Daniel, F. & Schmidt Improv. Random Forests Smoothing https://arXiv.org/abs/250506852. (2025).

  • He, Q., Chen, E., An, R. & Li, Y. Above-ground biomass and biomass components Estimation using lidar data in a coniferous forest. Forests 4 (4), 984–1002 (2013).

    Google Scholar 

  • Chen, L. et al. Evaluating the transferability of spectral variables and prediction models for mapping forest aboveground biomass using transfer learning methods. Remote Sens. 15 (22), 5358 (2023).

    Google Scholar 

  • Schmude, J. et al. Prithvi wxc: foundation model for weather and climate. https://arXiv.org/abs/2409.13598 (2024).

  • Klemmer, K., Rolf, E., Robinson, C., Mackey, L. & Rußwurm, M. April. Satclip: Global, general-purpose location embeddings with satellite imagery. In Proceedings of the AAAI Conference on Artificial Intelligence. 39 (4), 4347–4355. (2025).

  • Biau, G. Analysis of a random forests model. J. Mach. Learn. Res. 13, 1063–1095 (2012).

    MathSciNet 

    Google Scholar 

    • Continue Reading

  • Multi-fault diagnosis and damage assessment of rolling bearings based on IDBO-VMD and CNN-BiLSTM

    Multi-fault diagnosis and damage assessment of rolling bearings based on IDBO-VMD and CNN-BiLSTM

  • Chen, X., Yang, R. & Xue, Y. Deep transfer learning for bearing fault diagnosis: A systematic review since 2016. IEEE Trans. 72, 1–21. https://doi.org/10.1109/TIM.2023.3244237 (2023).

    Article 

    Google Scholar 

  • Salunkhe, V. G. et al. A novel incipient fault detection technique for roller bearing using deep independent component analysis and variational modal decomposition. ASME J. Tribol. 145, 7. https://doi.org/10.1115/1.4056899 (2023).

    Article 

    Google Scholar 

  • Zhou, X. et al. Fault diagnosis method of rolling bearing based on improved VMD spectrogram and FCM. Mach. Tools Hydraul. 51, 206–211. https://doi.org/10.1007/s11042-020-09534-w (2023).

    Article 

    Google Scholar 

  • Cheng, X. et al. Gearbox fault diagnosis method based on lightweight channel attention mechanism and transfer learning. Sci. Rep. 14, 743. https://doi.org/10.1038/s41598-023-50826-6 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Jin, X. H., Zhao, M. B., Chow, T. W. S. & Pecht, M. Motor bearing fault diagnosis using trace ratio linear discriminant analysis. 61, 2441–2451. https://doi.org/10.1109/TIE.2013.2273471 (2014).

  • Salunkhe, V. G. & Desavale, R. G. An intelligent prediction for detecting bearing vibration characteristics using a machine learning model. ASME J. Nondestructive Evaluation. 4, 3. https://doi.org/10.1115/1.4049938 (2021).

    Article 

    Google Scholar 

  • Wang, Q., Hu, S. & Wang, X. Detection of incipient rotor unbalance fault based on the RIME-VMD and modified-WKN. Sci. Rep. 14, 4683. https://doi.org/10.1038/s41598-024-54984-z (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Mao, M. et al. Application of FCEEMD-TSMFDE and adaptive catboost in fault diagnosis of complex variable condition bearings. Sci. Rep. 14, 30448. https://doi.org/10.1038/s41598-024-78845-x (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Yan, R. Q., Shang, Z. G., Xu, H. & Wen, J. C. Wavelet transform for rotary machine fault diagnosis:10 years revisited. Mech. Syst. Signal. Proc. 00, 110545. https://doi.org/10.1016/j.ymssp.2023.110545 (2023).

    Article 

    Google Scholar 

  • Randall, R. B. & Antoni, J. Why EMD and similar decompositions are of little benefit for bearing diagnostics. Mech. Syst. Signal. Proc. 192, 110207. https://doi.org/10.1016/j.ymssp.2023.110545 (2023).

    Article 

    Google Scholar 

  • Cheng, Y. & Zou, D. Complementary ensemble local means decomposition method and its application to rolling element bearings fault diagnosis. Proc. Inst. Mech. Eng. O 233, 868–880 (2019).

  • Wang, Z., Yang, J. & Guo, Y. Unknown fault feature extraction of rolling bearings under variable speed conditions based on statistical complexity measures. Mech. Syst. Signal. Proc. 72, 108964. https://doi.org/10.1016/j.ymssp.2022.108964 (2022).

    Article 

    Google Scholar 

  • Dragomiretskiy, K. & Zosso, D. Variational mode decomposition. IEEE Trans. Signal. Process. 62, 531–544. https://doi.org/10.1109/TSP.2013.2288675 (2013).

    Article 
    MathSciNet 

    Google Scholar 

  • Zhang, S., Wang, Y. & He, S. Bearing fault diagnosis based on variational mode decomposition and total variation denoising. Meas. Sci. Technol. 27, 075101. https://doi.org/10.1088/0957-0233/27/7/075101 (2016).

    Article 

    Google Scholar 

  • Li, K., Su, L., Wu, J. & Wang, H. A. Rolling bearing fault diagnosis method based on variational mode decomposition and an improved kernel extreme learning machine. Appl. Sci. 7, 1004. https://doi.org/10.3390/app7101004 (2017).

    Article 

    Google Scholar 

  • Yuan, Y. et al. Noise reduction and feature enhancement of hob vibration signal based on parameter adaptive VMD and autocorrelation analysis. Meas. Sci. Technol. 33, 125116. https://doi.org/10.1088/1361-6501/ac8e23 (2022).

    Article 

    Google Scholar 

  • Xing, Y. & Jian, R. Features method for selecting VMD parameters based on spectrum without modal overlap. J. Phys. Conf. Ser. 1605, 012002. https://doi.org/10.1088/1742-6596/1605/1/012002 (2020).

    Article 

    Google Scholar 

  • Zhong, X., Xia, T. & Mei, Q. An effective centre frequency selection scheme based on variational mode extraction and its application to gear fault diagnosis. Insight-Non-Destructive Test. Condition Monit. 64, 155–163. https://doi.org/10.1784/insi.2022.64.3.155 (2022).

    Article 

    Google Scholar 

  • Li, H., Liu, T., Wu, X. & Chen, Q. An optimized VMD method and its applications in bearing fault diagnosis. Meas 166, 108185. https://doi.org/10.1016/j.measurement.2020.108 (2020).

    Article 

    Google Scholar 

  • Zhang, X., Miao, Q., Zhang, H. & Wang, L. A parameter-adaptive VMD method based on grasshopper optimization algorithm to analyze vibration signals from rotating machinery. Mech. Syst. Signal. Proc. 108, 58–72. https://doi.org/10.1016/j.ymssp.2017.11.029 (2018).

    Article 

    Google Scholar 

  • Jin, Z., He, D. & Wei, Z. Intelligent fault diagnosis of train axle box bearing based on parameter optimization VMD and improved DBN. Eng. Appl. Artif. Intell. 110, 104713. https://doi.org/10.1016/j.engappai.2022.104713 (2022).

    Article 

    Google Scholar 

  • Wang, X. et al. Fault diagnosis method of rolling bearing based on SSA-VMD and RCMDE. Sci. Rep. 14, 30637. https://doi.org/10.1038/s41598-024-81262-9 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Tang, J., Liu, G. & Pan, Q. A review on representative swarm intelligence algorithms for solving optimization problems: applications and trends. IEEE-CAA J. AUTOMATIC. 8 (10), 1627–1643. https://doi.org/10.1109/JAS.2021.1004129 (2021).

    Article 
    MathSciNet 

    Google Scholar 

  • Xue, J. & Shen, B. Dung beetle optimizer: A new meta-heuristic algorithm for global optimization. J. Supercomput. 79, 7305–7336. https://doi.org/10.1007/s11227-022-04959-6 (2023).

    Article 

    Google Scholar 

  • Kong, W. et al. PID control algorithm based on multistrate-gy enhanced Dung beetle optimizer and back propagation neural network for DC motor control. Sci. Rep. 14, 28276. https://doi.org/10.1038/s41598-024-79653-z (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Tang, X., He, Z. & Jia, C. Multi-strategy cooperative enhancement Dung beetle optimizer and its application in obstacle avoidance navigation. Sci. Rep. 14, 28041. https://doi.org/10.1038/s41598-024-79420-0 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Li, X. et al. A review on convolutional neural network in rolling bearing fault diagnosis. Meas. Sci. Technol. 35, 1. https://doi.org/10.1088/1361-6501/ad356e (2024).

    Article 

    Google Scholar 

  • Salunkhe, V. G. et al. Vibration dynamic analysis of the bearing parameters in steam turbine bearing systems in sugar refinery. ASME J. Tribol. 148, 13. https://doi.org/10.1115/1.4068559 (2025).

    Article 

    Google Scholar 

  • Ruan, D. W., Han, J. Z., Yan, J. P. & Guehmann, C. Light convolutional neural network by neural architecture search and model pruning for bearing fault diagnosis and remaining useful life prediction. Sci. Rep. 13, 1. https://doi.org/10.1038/s41598-023-31532 (2023).

    Article 

    Google Scholar 

  • LeCun, Y., Bengio, Y. & Hinton, G. Deep learning. Nature 521, 436–444. https://doi.org/10.1038/nature14539 (2015).

    Article 
    PubMed 

    Google Scholar 

  • Choudhary, A., Mian, T. & Fatima, S. Convolutional neural network based bearing fault diagnosis of rotating machine using thermal images. Meas 176, 109196. https://doi.org/10.1016/j.measurement.2021.109196 (2021).

    Article 

    Google Scholar 

  • Tang, G., Zhou, Y., Wang, H. & Li, G. Prediction of bearing performance degradation with bottleneck feature based on LSTM network. IEEE Int. Inst. Meas. Tec. Conf. 1–6. https://doi.org/10.1109/I2MTC.2018.8409564 (2018).

  • Guo, Y., Mao, J. & Zhao, M. Rolling bearing fault diagnosis method based on attention CNN and BiLSTM network. Neural Process. Lett. 55, 3377–3410. https://doi.org/10.1007/s11063-022-11013-2 (2023).

    Article 

    Google Scholar 

  • Zhang, Y. et al. Attention activation network for bearing fault diagnosis under various noise environments. Sci. Rep. 15, 977. https://doi.org/10.1038/s41598-025-85275-w (2025).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Salunkhe, V. G. et al. Rolling element bearing fault diagnosis by the implementation of Elman neural networks with long Short-Term memory strategy. ASME J. Tribol. 147, 8. https://doi.org/10.1115/1.4067382 (2025).

    Article 

    Google Scholar 

  • Saeed, A., Khan, A. & Akram, M. Deep learning based approaches for intelligent industrial machinery health management and fault diagnosis in resource-constrained environments. Sci. Rep. 15, 1114. https://doi.org/10.1038/s41598-024-79151-2 (2025).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Chen, X., Zhang, B. & Gao, D. Bearing fault diagnosis base on multi-scale CNN and LSTM model. J. Intell. Manuf. 32, 971–987. https://doi.org/10.1007/s10845-020-01600-2 (2021).

    Article 

    Google Scholar 

  • Nacer, S. M. et al. A novel method for bearing fault diagnosis based on BiLSTM neural networks. Int. J. Adv. Manuf. Technol. 125, 1477–1492. https://doi.org/10.1007/s00170-022-10792-1 (2023).

    Article 

    Google Scholar 

  • Salunkhe, V. G. et al. Unbalance bearing fault identification using highly accurate Hilbert–Huang transform approach. J. Nondestr Eval Diag. 6, 3. https://doi.org/10.1115/1.4062929 (2023).

    Article 

    Google Scholar 

  • Yıldız, B. S., Kumar, S., Pholdee, N. & Yildiz, A. R. A new chaotic lévy flight distribution optimization algorithm for solving constrained engineering problems. Expert Syst. Appl. 39, 12992. https://doi.org/10.1111/exsy.12992 (2022).

    Article 

    Google Scholar 

  • Tanyildizi, E. & Demir, G. Golden sine algorithm: A novel Math-Inspired algorithm. Adv. Electr. Comput. Eng. 17, 71–78. https://doi.org/10.4316/AECE.2017.02010 (2017).

    Article 

    Google Scholar 

  • Seyedali, M. SCA: A sine cosine algorithm for solving optimization problems. KNOWL-BASED SYST. 96, 120–133. https://doi.org/10.1016/j.knosys.2015.12.022 (2016).

    Article 

    Google Scholar 

  • Ghasemi, M., Zare, M. & Trojovský, P. Optimization based on the smart behavior of plants with its engineering applications: Ivy algorithm. Knowl.-Based Syst. 295, 111850. https://doi.org/10.1016/j.knosys.2024.111850 (2024).

    Article 

    Google Scholar 

  • Amiri, M. H. et al. Hippopotamus optimization algorithm: a novel nature-inspired optimization algorithm. Sci. Rep. 14, 5032. https://doi.org/10.1038/s41598-024-54910-3 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Xin, Y., Yong, L. & Guang, L. Evolutionary programming made faster. 3, 102. https://doi.org/10.1109/4235.771163 (1999).

  • Wang, Y. H. et al. Degradation trend prediction of hydropower units based on a comprehensive deterioration index and LSTM. Energies 15, 6273. https://doi.org/10.3390/en15176273 (2022).

  • Jiang, Z. et al. A fault detection of aero-engine rolling bearings based on CNN-BiLSTM network integrated cross-attention. Meas. Sci. Technol. 35, 12 (2014).

  • Dao, F. et al. Wear fault diagnosis in hydro-turbine via the incorporation of the IWSO algorithm optimized CNN-LSTM neural network. Sci. Rep. 14, 25278. https://doi.org/10.1038/s41598-024-77251-7 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Li, X. et al. An improved sparrow search algorithm and CNN-BiLSTM neural network for predicting sea level height. Sci. Rep. 14, 4560. https://doi.org/10.1038/s41598-024-55266-4 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Salunkhe, V. G. et al. An integrated dimension theory and modulation signal bispectrum technique for analyzing bearing fault in industrial fibrizer. J. Nondestr Eval Diag. 7, 3. https://doi.org/10.1115/1.4065545 (2024).

    Article 

    Google Scholar 

  • Song, Q. et al. Fault diagnosis of HVCB via the Subtraction average based optimizer algorithm optimized multi channel CNN-SABO-SVM network. Sci. Rep. 14, 29507. https://doi.org/10.1038/s41598-024-80954-6 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Xiao, Q. et al. Feature decoupling integrated domain generalization network for bearing fault diagnosis under unknown operating conditions. Sci. Rep. 14, 30848. https://doi.org/10.1038/s41598-024-81489-6 (2024).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Zhao, C., Zio, E. & Shen, W. Domain generalization for Cross-Domain fault diagnosis: an Application-oriented perspective and a benchmark study. Reliab. Eng. Syst. Safe. 245, 109964. https://doi.org/10.1016/j.ress.2024.109964 (2024).

    Article 

    Google Scholar 

  • Jagadeesha, T. et al. Investigation of Crack Detection Technique in a Rotating Shaft by Using Vibration Measurement 631–645. https://doi.org/10.1007/978-981-15-4739-3_54 (AIME, 2021).

  • Song, J. et al. Reliability analysis of gear-bearing drive systems considering gear manufacturing and installation errors. Sci. Rep. 15, 23301. https://doi.org/10.1038/s41598-025-06446-3 (2025).

    Article 
    PubMed 
    PubMed Central 

    Google Scholar 

  • Salunkhe, V. G., Desavale, R. G. & Jagadeesha, T. A numerical model for fault diagnosis in deep groove ball bearing using dimension theory. Mater. Today Proc. 47, 3077–3084. https://doi.org/10.1016/j.matpr.2021.06.072 (2021).

    • Continue Reading

  • Is It Ever Legal—or Ethical—to Remove DRM?

    Is It Ever Legal—or Ethical—to Remove DRM?

    Whatever you think about Digital Rights Management software, it’s hard to argue with the fact that it’s annoying. Such technology exists, in theory, to protect the intellectual property of the companies that create music, movies, and games, but it can also get in the way of you enjoying books, music, and videos the way you want to.

    Say, for example, that you bought a bunch of books on the Amazon Kindle platform but later decided you wanted to switch to a Kobo device (or vice versa). The DRM systems on both platforms are designed to prevent you from migrating your books from one platform to the other, meaning you might need to pay again just to read a book on the Kobo you already paid for on the Kindle.

    Software exists that can remove this DRM. It requires doing some research and jumping through some technical hoops, but it could save you from having to buy the same media a second time. But is it legal?

    Illegal, but Unlikely to Be Enforced

    Let’s rip off the Band-Aid: The act of removing DRM from any kind of copyrighted work is broadly illegal in the US under section 1201 of title 17, United States Code, which was passed with the Digital Millennium Copyright Act in 1998.

    The rule is pretty clearly written: “No person shall circumvent a technological measure that effectively controls access to a work protected under this title.”

    I asked to Derek Bambauer, a law professor at the University or Florida who specializes in internet law, cybersecurity, and intellectual property, whether there are any exceptions to this. Bambauer tells me there is very little legal ambiguity here.

    “Exceptions are in there, but they’re really narrow,” he says, emphasizing that First Amendment arguments tend not to work where intellectual property is concerned.

    But will removing DRM from a file you paid for end up with you going to jail or being sued? Maybe not, says Bambauer, as companies typically don’t bother pursing it.

    “We all kind of casually violate copyright constantly,” he says, mentioning that the photocopied comic strips on his office door are technically a violation. With finite resources, the owners of intellectual property simple can’t sue everyone. “Companies tend to only go after people who are distributing copyrighted works with others.”

    “If the DRM removal is just for personal consumption, it’s hard to detect,” Bambauer says, “and it’s not worth it for companies go after that. One, the music industry tried that, and it was a horrible failure. And two, it’s just a lot easier to go after the creators and distributors of the tools.”

    And that’s what companies do: attempt to shut down or block the distribution of software that removes DRM from files. Bambauer says this is the reason most of such software tends to be made by people outside the US and is distributed on websites outside US jurisdiction. He also emphasized that sharing files with others after removing the DRM is far more likely to attract lawsuits from companies.

    Continue Reading

  • How Payoneer’s Partnership With Stripe May Influence PAYO’s Cross-Border Payment Strategy

    How Payoneer’s Partnership With Stripe May Influence PAYO’s Cross-Border Payment Strategy

    • Earlier this month, Payoneer announced a partnership with Stripe to expand its Online Checkout solution for cross-border SMBs in Asia Pacific, enabling acceptance of Buy Now Pay Later options and major digital wallets through webstores.
    • This collaboration leverages Stripe’s technology to enhance payment acceptance, customer conversion rates, and fraud prevention, marking a move by Payoneer to broaden its SMB-focused financial solutions via high-impact partnerships.
    • We’ll explore how the expanded checkout offerings, including Buy Now Pay Later and digital wallets, may influence Payoneer’s investment narrative.

    Find companies with promising cash flow potential yet trading below their fair value.

    Payoneer Global Investment Narrative Recap

    To be a shareholder in Payoneer, you need to believe in the company’s ability to expand its platform for global SMB payments while navigating intensifying competition and advances in blockchain technology. The recent Stripe partnership could boost adoption of Payoneer’s checkout solution by broadening digital payment methods, but the most important near-term catalyst remains its ability to grow higher-margin B2B services. The biggest current risk, escalating competition from fintech and traditional banks, remains largely unaltered by this announcement, as top rivals continue to roll out similar offerings.

    Among recent announcements, Payoneer’s collaboration with Citi, enabling blockchain-based treasury transfers, stands out as especially relevant since it addresses one of the key threats: long-term disruption from blockchain-powered payment rails. While this supports Payoneer’s efforts to modernize its platform and potentially defend market share, the competitive pressure from larger integrated payment networks persists as a core challenge.

    Yet in contrast, investors should be aware of the ongoing risk that new competitors may compress Payoneer’s margins, especially as market standards and client…

    Read the full narrative on Payoneer Global (it’s free!)

    Payoneer Global’s narrative projects $1.3 billion in revenue and $130.7 million in earnings by 2028. This requires 8.2% yearly revenue growth and a $30.9 million earnings increase from $99.8 million today.

    Uncover how Payoneer Global’s forecasts yield a $9.81 fair value, a 40% upside to its current price.

    Exploring Other Perspectives

    PAYO Community Fair Values as at Aug 2025

    Simply Wall St Community members offer three fair value estimates for Payoneer stock, ranging from US$6.07 to US$11.02 per share. With competitive threats from both fintech start-ups and large incumbents, it is worth considering how varied these outlooks are before you form your own view.

    Explore 3 other fair value estimates on Payoneer Global – why the stock might be worth 13% less than the current price!

    Build Your Own Payoneer Global Narrative

    Disagree with existing narratives? Create your own in under 3 minutes – extraordinary investment returns rarely come from following the herd.

    • A great starting point for your Payoneer Global research is our analysis highlighting 2 key rewards that could impact your investment decision.
    • Our free Payoneer Global research report provides a comprehensive fundamental analysis summarized in a single visual – the Snowflake – making it easy to evaluate Payoneer Global’s overall financial health at a glance.

    Ready For A Different Approach?

    Right now could be the best entry point. These picks are fresh from our daily scans. Don’t delay:

    This article by Simply Wall St is general in nature. We provide commentary based on historical data
    and analyst forecasts only using an unbiased methodology and our articles are not intended to be financial advice.     It does not constitute a recommendation to buy or sell any stock, and does not take account of your objectives, or your
    financial situation. We aim to bring you long-term focused analysis driven by fundamental data.
    Note that our analysis may not factor in the latest price-sensitive company announcements or qualitative material.
    Simply Wall St has no position in any stocks mentioned.

    New: Manage All Your Stock Portfolios in One Place

    We’ve created the ultimate portfolio companion for stock investors, and it’s free.

    • Connect an unlimited number of Portfolios and see your total in one currency
    • Be alerted to new Warning Signs or Risks via email or mobile
    • Track the Fair Value of your stocks

    Try a Demo Portfolio for Free

    Have feedback on this article? Concerned about the content? Get in touch with us directly. Alternatively, email editorial-team@simplywallst.com

    Continue Reading