Automated temporal landslide detection: Testing whether
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
Automated temporal landslide detection: Testing whether human-in-the-loop machine learning can bootstrap generalizability - Sustainability Accelerator
sustainability-accelerator.stanford.edu
November 8, 2026 at 12:09 PM
Automated temporal landslide detection: Testing whether
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
the replication of results in clinical trials is important for informing treatment guidelines. Observational data improve generalizability beyond individuals who are eligible for and consent for clinical trials.
November 9, 2026 at 8:27 PM
the replication of results in clinical trials is important for informing treatment guidelines. Observational data improve generalizability beyond individuals who are eligible for and consent for clinical trials.
Toward LLM Self-Play Under Data Scarcity: Effectiveness, Generalizability, and Trustworthiness
Zhiyin Yu et al.
#arXiv #cs.AI
Zhiyin Yu et al.
#arXiv #cs.AI
Toward LLM Self-Play Under Data Scarcity: Effectiveness, Generalizability, and Trustworthiness
The rapid development of large language models (LLMs) has been driven largely by scaling model size and human-curated training data, yet this paradigm is increasingly constrained by data scarcity and diminishing returns from human supervision. Self-play offers a promising pathway for autonomous dat…
arxiv.org
November 10, 2026 at 1:57 AM
Finally, I consider a number of potential objections to the arguments (and analyses) in the paper, including the idea that interpretability is actually idiographic, not nomothetic, as well as the idea that generalizability is not a concern if we can run automated interpretability suites.
November 10, 2026 at 8:08 PM
Finally, I consider a number of potential objections to the arguments (and analyses) in the paper, including the idea that interpretability is actually idiographic, not nomothetic, as well as the idea that generalizability is not a concern if we can run automated interpretability suites.
A theory of generalizability is also itself of scientific interest: even ostensibly "idiographic" research on a given LLM is often informed by insights from the study of other LLMs. Knowing which population a given LLM sample stands in for, and how, is a helpful sign of progress in the field.
November 10, 2026 at 8:08 PM
A theory of generalizability is also itself of scientific interest: even ostensibly "idiographic" research on a given LLM is often informed by insights from the study of other LLMs. Knowing which population a given LLM sample stands in for, and how, is a helpful sign of progress in the field.
17/ This generalizability concern is just one small part of the paper. Check out the thread by @mbneff.bsky.social for a great summary of the other arguments.
November 4, 2026 at 5:42 PM
17/ This generalizability concern is just one small part of the paper. Check out the thread by @mbneff.bsky.social for a great summary of the other arguments.
Zhiyin Yu, Yuchen Mou, Bingzhe Wu, Zhonghai Wu, Shiyang Feng, Xiao Luo: Toward LLM Self-Play Under Data Scarcity: Effectiveness, Generalizability, and Trustworthiness https://arxiv.org/abs/2611.08908 https://arxiv.org/pdf/2611.08908 https://arxiv.org/html/2611.08908
November 10, 2026 at 6:39 AM
Zhiyin Yu, Yuchen Mou, Bingzhe Wu, Zhonghai Wu, Shiyang Feng, Xiao Luo: Toward LLM Self-Play Under Data Scarcity: Effectiveness, Generalizability, and Trustworthiness https://arxiv.org/abs/2611.08908 https://arxiv.org/pdf/2611.08908 https://arxiv.org/html/2611.08908
Next theme: generalizability - can a framework built for one artifact type be applied to another? -- @nelmqvist.bsky.social -- "most popular coding language now is English" can we reimplement just paper + images (no code) @alex-lex.bsky.social videos are much better documentation for reproducibility
November 10, 2026 at 4:17 PM
Next theme: generalizability - can a framework built for one artifact type be applied to another? -- @nelmqvist.bsky.social -- "most popular coding language now is English" can we reimplement just paper + images (no code) @alex-lex.bsky.social videos are much better documentation for reproducibility
Zhiyin Yu, Yuchen Mou, Bingzhe Wu, Zhonghai Wu, Shiyang Feng, Xiao Luo: Toward LLM Self-Play Under Data Scarcity: Effectiveness, Generalizability, and Trustworthiness https://arxiv.org/abs/2611.08908 https://arxiv.org/pdf/2611.08908 https://arxiv.org/html/2611.08908
November 9, 2026 at 6:38 AM
Zhiyin Yu, Yuchen Mou, Bingzhe Wu, Zhonghai Wu, Shiyang Feng, Xiao Luo: Toward LLM Self-Play Under Data Scarcity: Effectiveness, Generalizability, and Trustworthiness https://arxiv.org/abs/2611.08908 https://arxiv.org/pdf/2611.08908 https://arxiv.org/html/2611.08908
From the latest issue: Yahyavi & colleagues compare CREST-2 participants with patients undergoing carotid revascularization in real-world practice, finding substantial alignment in baseline characteristics and GDMT-related measures, supporting the generalizability of CREST-2.
doi.org/10.1177/1358...
doi.org/10.1177/1358...
Generalizability of the CREST‑2 trial: A real‑world analysis - Ashkan Yahyavi, Mehdi H Shishehbor, 2026
Alternatively, explore:
doi.org
November 6, 2026 at 5:01 PM
From the latest issue: Yahyavi & colleagues compare CREST-2 participants with patients undergoing carotid revascularization in real-world practice, finding substantial alignment in baseline characteristics and GDMT-related measures, supporting the generalizability of CREST-2.
doi.org/10.1177/1358...
doi.org/10.1177/1358...
3/ (b) during piloting, study design features (measures, stimuli, settings, etc.) are tested and selected. Features that lead to large/novel effects may be overselected for, which can harm the generalizability of final results. Reporting pilots can reveal such selection biases.
November 4, 2026 at 1:11 PM
3/ (b) during piloting, study design features (measures, stimuli, settings, etc.) are tested and selected. Features that lead to large/novel effects may be overselected for, which can harm the generalizability of final results. Reporting pilots can reveal such selection biases.
The idea was that by excluding untested people altogether, (only PCR counted), they would reduce confounders. In that respect, the logic is sound. It also reduces generalizability outside of those likely to be tested, of course.
But it provides extra support for COVID attribution in some cases.
But it provides extra support for COVID attribution in some cases.
November 10, 2026 at 12:08 AM
The idea was that by excluding untested people altogether, (only PCR counted), they would reduce confounders. In that respect, the logic is sound. It also reduces generalizability outside of those likely to be tested, of course.
But it provides extra support for COVID attribution in some cases.
But it provides extra support for COVID attribution in some cases.
Autonomous generalizability auditing of scientific foundation models https://www.biorxiv.org/content/10.64898/2026.11.03.763979v1
November 4, 2026 at 10:46 PM
Autonomous generalizability auditing of scientific foundation models https://www.biorxiv.org/content/10.64898/2026.11.03.763979v1
"Achieving consensus on the definition of [acute-on-chronic liver failure] improves the generalizability of research across the globe"
Global Liver Societies Agree on One Definition of ACLF
Five international liver societies have reached a global consensus on the definition of acute-on-chronic liver failure, which will affect research and, eventually, patients.
www.medscape.com
November 10, 2026 at 7:33 PM
"Achieving consensus on the definition of [acute-on-chronic liver failure] improves the generalizability of research across the globe"
Autonomous generalizability auditing of scientific foundation models https://www.biorxiv.org/content/10.64898/2026.11.03.763979v1
November 4, 2026 at 10:46 PM
Autonomous generalizability auditing of scientific foundation models https://www.biorxiv.org/content/10.64898/2026.11.03.763979v1
💡 Optimising reconstruction protocols is key to more reliable CT radiomics.
link.springer.com/article/10.1... (Huiyuan Zhu et al.)
link.springer.com/article/10.1... (Huiyuan Zhu et al.)
Impact of deep learning image reconstruction on CT radiomics reproducibility in patients with pulmonary nodules - Insights into Imaging
Objective Ensuring CT feature reproducibility across imaging protocols and reconstruction algorithms is essential for the generalizability of radiomics applications. This study aimed to evaluate the…
link.springer.com
November 11, 2026 at 10:02 AM
💡 Optimising reconstruction protocols is key to more reliable CT radiomics.
link.springer.com/article/10.1... (Huiyuan Zhu et al.)
link.springer.com/article/10.1... (Huiyuan Zhu et al.)
Q&A - someone asked about transportability of models developed in the US and working outside the US
We need to talk about transportability (different from generalizability) more
#ACR26 #GlobalSummit
We need to talk about transportability (different from generalizability) more
#ACR26 #GlobalSummit
November 6, 2026 at 4:10 PM
Q&A - someone asked about transportability of models developed in the US and working outside the US
We need to talk about transportability (different from generalizability) more
#ACR26 #GlobalSummit
We need to talk about transportability (different from generalizability) more
#ACR26 #GlobalSummit
Rethinking Fitness–Cingulate Cortex Associations: Interpretive and Generalizability Considerations from the YoungFit Study Phys&Beh
Rethinking Fitness–Cingulate Cortex Associations: Interpretive and Generalizability Considerations from the YoungFit Study
Publication date: Available online 3 November 2026
Source: Physiology & Behavior
Author(s): Riza Amalia, Fendahapsari Singgih Sendayu, Muhammad Sandy Al Fath, Ganny Baihaqi
dlvr.it
November 4, 2026 at 8:53 PM
Rethinking Fitness–Cingulate Cortex Associations: Interpretive and Generalizability Considerations from the YoungFit Study Phys&Beh
Yingchao Huang, Yuhan Su, Xin Wang, Shanshan Yao: Improving the Clinical Generalizability of Machine Learning-Based Alzheimer's Disease Detection Across Domains and Modalities: A Comprehensiv... https://arxiv.org/abs/2611.08188 https://arxiv.org/pdf/2611.08188 https://arxiv.org/html/2611.08188
November 6, 2026 at 6:45 AM
Yingchao Huang, Yuhan Su, Xin Wang, Shanshan Yao: Improving the Clinical Generalizability of Machine Learning-Based Alzheimer's Disease Detection Across Domains and Modalities: A Comprehensiv... https://arxiv.org/abs/2611.08188 https://arxiv.org/pdf/2611.08188 https://arxiv.org/html/2611.08188
Generative Active Learning for Label-Efficient Pan-Organ Ultrasound Diagnosis
Yaofei Duan et al.
#arXiv #cs.AI #cs.HC
Yaofei Duan et al.
#arXiv #cs.AI #cs.HC
Generative Active Learning for Label-Efficient Pan-Organ Ultrasound Diagnosis
Medical ultrasound (US) enables real-time, non-invasive diagnosis, yet building US-based AI systems is limited by data scarcity, costly expert annotation, and limited cross-center generalizability. Privacy constraints and instance-level reporting hinder the collection of diverse abnormal cases and …
arxiv.org
November 10, 2026 at 6:37 PM
11/ Note that the concern here is not false positives. These effects may be real and hold up in direct replications. The concern is generalizability.
November 4, 2026 at 5:42 PM
11/ Note that the concern here is not false positives. These effects may be real and hold up in direct replications. The concern is generalizability.
Certainty of evidence was graded very low. Probiotics may reduce migraine severity and frequency; however, the heterogeneity of studies limits generalizability. (🧵 5/6)
November 4, 2026 at 10:41 PM
Certainty of evidence was graded very low. Probiotics may reduce migraine severity and frequency; however, the heterogeneity of studies limits generalizability. (🧵 5/6)
Looks like Huckabee. Gives some credence to generalizability of physiognomy.
November 1, 2026 at 10:11 PM
Looks like Huckabee. Gives some credence to generalizability of physiognomy.
Reply to “Rethinking Fitness–Cingulate Cortex Associations: Interpretive and Generalizability Considerations from the YoungFit Study” Phys&Beh
Reply to “Rethinking Fitness–Cingulate Cortex Associations: Interpretive and Generalizability Considerations from the YoungFit Study”
Publication date: Available online 3 November 2026
Source: Physiology & Behavior
Author(s): Adrià Bermudo-Gallaguet, Neus Camins-Vila, Samira Rostami, Rosalia Dacosta-Aguayo, Judit Escarré-Grifell, Blai Ferrer-Uris, Albert Busquets, Louis Bherer, Maria Mataró
dlvr.it
November 4, 2026 at 8:56 AM
Reply to “Rethinking Fitness–Cingulate Cortex Associations: Interpretive and Generalizability Considerations from the YoungFit Study” Phys&Beh
cautiously. However, the generalizability of these results was limited because the study was clinic-based rather than population-based. (🧵 10/10)
November 3, 2026 at 5:22 PM
cautiously. However, the generalizability of these results was limited because the study was clinic-based rather than population-based. (🧵 10/10)