❌

Normal view

Received — 28 January 2026 ⏭ Favourites

Agentic Digital Twins: A Taxonomy of Capabilities for Understanding Possible Futures

✇cs.AI, q-bio.NC updates on arXiv.org
By: Christopher Burr · Mark Enzer · Jason Shepherd · David Wagg
28 January 2026 at 13:00
arXiv:2601.18799v1 Announce Type: cross Abstract: As digital twins (DTs) evolve to become more agentic through the integration of artificial intelligence (AI), they acquire capabilities that extend beyond dynamic representation of their target systems. This paper presents a taxonomy of agentic DTs organised around three fundamental dimensions: the locus of agency (external, internal, distributed), the tightness of coupling (loose, tight, constitutive), and model evolution (static, adaptive, reconstructive). From the resulting 27-configuration space, we identify nine illustrative configurations grouped into three clusters: "The Present" (existing tools and emerging steering systems), "The Threshold" (where emergent properties appear and coupling becomes constitutive), and "The Frontier" (where systems gain reconstructive capabilities). Our analysis explores how agentic DTs exercise performative power--not merely representing physical systems but actively participating in constituting them. Using traffic navigation systems as examples, we show how even passive tools can exhibit emergent performativity, while advanced configurations risk performative lock-in. Drawing on performative prediction theory, we trace a progression from passive tools through active steering to ontological reconstruction, examining how constitutive coupling enables systems to create self-validating realities. Understanding these configurations is essential for navigating the transformation from DTs as mirror worlds to DTs as architects of new ontologies.

Is On-Policy Data always the Best Choice for Direct Preference Optimization-based LM Alignment?

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zetian Sun · Dongfang Li · Xuhui Chen · Baotian Hu · Min Zhang
28 January 2026 at 13:00
arXiv:2508.10530v2 Announce Type: replace Abstract: The alignment of language models~(LMs) with human preferences is critical for building reliable AI systems. The problem is typically framed as optimizing an LM policy to maximize the expected reward that reflects human preferences. Recently, Direct Preference Optimization~(DPO) was proposed as a LM alignment method that directly optimize the policy from static preference data, and further improved by incorporating on-policy sampling~(i.e., preference candidates generated during the training loop) for better LM alignment. However, we show on-policy data is not always optimal, with systematic effectiveness difference emerging between static and on-policy preference candidates. For example, on-policy data can result in a $3\times$ effectiveness compared with static data for Llama-3, and a $0.4\times$ effectiveness for Zephyr. To explain the phenomenon, we propose the alignment stage assumption, which divides the alignment process into two distinct stages: the preference injection stage, which benefits from diverse data, and the preference fine-tuning stage, which favors high-quality data. Through theoretical and empirical analysis, we characterize these stages and propose an effective algorithm to identify the boundaries between them. We perform experiments on $5$ models~(Llama, Zephyr, Phi-2, Qwen, Pythia) and $2$ alignment methods~(DPO, SLiC-HF) to show the generalizability of alignment stage assumption and the effectiveness of the boundary measurement algorithm.

Rethinking the AI Scientist: Interactive Multi-Agent Workflows for Scientific Discovery

✇cs.AI, q-bio.NC updates on arXiv.org
By: Lukas Weidener · Marko Brki\'c · Mihailo Jovanovi\'c · Ritvik Singh · Chiara Baccin · Emre Ulgac · Alex Dobrin · Aakaash Meduri
28 January 2026 at 13:00
arXiv:2601.12542v2 Announce Type: replace Abstract: Artificial intelligence systems for scientific discovery have demonstrated remarkable potential, yet existing approaches remain largely proprietary and operate in batch-processing modes requiring hours per research cycle, precluding real-time researcher guidance. This paper introduces Deep Research, a multi-agent system enabling interactive scientific investigation with turnaround times measured in minutes. The architecture comprises specialized agents for planning, data analysis, literature search, and novelty detection, unified through a persistent world state that maintains context across iterative research cycles. Two operational modes support different workflows: semi-autonomous mode with selective human checkpoints, and fully autonomous mode for extended investigations. Evaluation on the BixBench computational biology benchmark demonstrated state-of-the-art performance, achieving 48.8% accuracy on open response and 64.4% on multiple-choice evaluation, exceeding existing baselines by 14 to 26 percentage points. Analysis of architectural constraints, including open access literature limitations and challenges inherent to automated novelty assessment, informs practical deployment considerations for AI-assisted scientific workflows.

General Binding Affinity Guidance for Diffusion Models in Structure-Based Drug Design

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yue Jian · Curtis Wu · Danny Reidenbach · Aditi S. Krishnapriyan
28 January 2026 at 13:00
arXiv:2406.16821v2 Announce Type: replace-cross Abstract: Structure-based drug design (SBDD) aims to generate ligands that bind strongly and specifically to target protein pockets. Recent diffusion models have advanced SBDD by capturing the distributions of atomic positions and types, yet they often underemphasize binding affinity control during generation. To address this limitation, we introduce \textbf{\textnormal{\textbf{BADGER}}}, a general \textbf{binding-affinity guidance framework for diffusion models in SBDD}. \textnormal{\textbf{BADGER} }incorporates binding affinity awareness through two complementary strategies: (1) \textit{classifier guidance}, which applies gradient-based affinity signals during sampling in a plug-and-play fashion, and (2) \textit{classifier-free guidance}, which integrates affinity conditioning directly into diffusion model training. Together, these approaches enable controllable ligand generation guided by binding affinity. \textnormal{\textbf{BADGER} } can be added to any diffusion model and achieves up to a \textbf{60\% improvement in ligand--protein binding affinity} of sampled molecules over prior methods. Furthermore, we extend the framework to \textbf{multi-constraint diffusion guidance}, jointly optimizing for binding affinity, drug-likeness (QED), and synthetic accessibility (SA) to design realistic and synthesizable drug candidates.

AI-generated data contamination erodes pathological variability and diagnostic reliability

✇cs.AI, q-bio.NC updates on arXiv.org
By: Hongyu He · Shaowen Xiang · Ye Zhang · Yingtao Zhu · Jin Zhang · Hao Deng · Emily Alsentzer · Qingyu Chen · Kun-Hsing Yu · Andrew Marshall · Tingting Chen · Srinivas Anumasa · Daniel Ebner · Dean Ho · Kee Yuan Ngiam · Ching-Yu Cheng · Dianbo Liu
28 January 2026 at 13:00
arXiv:2601.12946v3 Announce Type: replace-cross Abstract: Generative artificial intelligence (AI) is rapidly populating medical records with synthetic content, creating a feedback loop where future models are increasingly at risk of training on uncurated AI-generated data. However, the clinical consequences of this AI-generated data contamination remain unexplored. Here, we show that in the absence of mandatory human verification, this self-referential cycle drives a rapid erosion of pathological variability and diagnostic reliability. By analysing more than 800,000 synthetic data points across clinical text generation, vision-language reporting, and medical image synthesis, we find that models progressively converge toward generic phenotypes regardless of the model architecture. Specifically, rare but critical findings, including pneumothorax and effusions, vanish from the synthetic content generated by AI models, while demographic representations skew heavily toward middle-aged male phenotypes. Crucially, this degradation is masked by false diagnostic confidence; models continue to issue reassuring reports while failing to detect life-threatening pathology, with false reassurance rates tripling to 40%. Blinded physician evaluation confirms that this decoupling of confidence and accuracy renders AI-generated documentation clinically useless after just two generations. We systematically evaluate three mitigation strategies, finding that while synthetic volume scaling fails to prevent collapse, mixing real data with quality-aware filtering effectively preserves diversity. Ultimately, our results suggest that without policy-mandated human oversight, the deployment of generative AI threatens to degrade the very healthcare data ecosystems it relies upon.

Products, Performance, and Technological Development of Ambulatory Oxygen Therapy Devices: Scoping Review

✇Journal of Medical Internet Research
By: Shohei Kawachi · Mariana Hoffman · Lorena Romero · Magnus Ekström · Jerry A Krishnan · Anne E Holland
28 January 2026 at 03:00
Background: Ambulatory oxygen therapy is prescribed for patients with chronic lung diseases who experience exertional hypoxemia. However, available devices may not adequately meet user requirements, and their performance characteristics are heterogeneous. Objective: This study aims to identify devices available for delivery of ambulatory oxygen therapy, the technologies that they use to generate oxygen, the performance characteristics of each device, and the development status. Methods: We used medical and engineering databases to identify peer-reviewed papers (eg, MEDLINE, IEEE). Gray literature was used to identify additional descriptions of ambulatory oxygen devices in military medicine, space exploration, or patents. The last search was conducted in September 2025. Documents that described a device that can deliver oxygen in an ambulatory context (defined as weighing less than 10 kg) and were written in English were included. Search results were screened for inclusion by 2 independent reviewers. Data were synthesized by descriptively mapping the performance of each product, the technology used, and the development status of emerging technologies. Results: From 9702 records identified, a total of 166 met eligibility criteria (106 scientific publications and 60 gray literature). We identified 33 portable oxygen concentrators (POCs; 29 commercially available), 10 oxygen cylinders, and 6 portable liquid oxygen (LOX) devices. The POC products showed a trade-off between portability and oxygen delivery capacity (maximum flow rate ranging from 2.0 to 6.0 L/min; device weight ranging from 1.0 to 9.1 kg). Pressure swing adsorption with zeolite was the most common oxygen generation technology in POCs on the market. The mean maximum continuous operating time of POCs was 3.8 hours. Two prototype POCs (maximum flow rate of 4-6 L/min and device weight of 8-9 kg) were developed for space exploration using modified adsorbents. LOX devices were the lightest and had the longest continuous operating time. Innovations in delivery included the downsizing of a POC by using nanozeolite as an adsorbent and pulse oximeter oxygen saturation (SpO2)–targeted automatic titration of oxygen delivery based on the user’s SpO2. Conclusions: This scoping review is the first study to integrate medical, engineering, and gray literature on ambulatory oxygen devices and their development. Although prior literature has narratively explained the products and technologies, no previous research has systematically investigated them. This review showed that POCs available to consumers may not meet the needs of patients in terms of flow rate, portability, and operating time. LOX devices offered superior performance but are limited by high costs. Limitations of this review include the difficulty of comparing product performance across oxygen delivery settings and that the records were largely obtained from English-language sources. Innovation in ambulatory oxygen technology has been limited over the past decade, highlighting urgent need for research and development of new lightweight devices with higher oxygen delivery. Clinical Trial: OSF Registries 10.17605/OSF.IO/QS7FX; https://osf.io/qs7fx

Exploring the key molecular mechanisms and immune microenvironment of oxidative stress-related pathways in pancreatic neuroendocrine tumor combining scRNA-seq and bulk RNA

✇(Multiomics OR Omics) AND (Pancreatic)
By: Zhong Cai · Jiangtao Li · Ting Li · Qiuxia Peng
27 January 2026 at 19:00

Discov Oncol. 2026 Jan 27. doi: 10.1007/s12672-026-04515-1. Online ahead of print.

ABSTRACT

BACKGROUND: Pancreatic neuroendocrine tumor (pNET) is a heterogeneous tumor originating from pancreatic endocrine cells. Emerging evidence suggests that oxidative stress plays a crucial role in pNET pathogenesis, yet the precise molecular mechanisms and their interplay with the tumor microenvironment remain unclear. This study aims to systematically elucidate how oxidative stress-related pathways drive pNET progression through an integrated multi-omics approach.

METHODS: We designed a three-tier analytical strategy to address interconnected scientific questions. First, to identify which oxidative stress-related genes are dysregulated in pNET, we performed differential expression analysis and weighted gene co-expression network analysis (WGCNA) on the GSE73338 dataset (63 pNET samples, 5 controls), intersecting the. results with oxidative stress gene sets to obtain 71 candidate genes. Second, to understand the functional implications of these genes, we conducted GO/KEGG enrichment analysis and constructed protein-protein interaction (PPI) networks, from which we identified BCL2L1 and PHGDH as key hub genes using three independent algorithms. We then assessed their diagnostic value through ROC analysis and built a prognostic nomogram model. Third, to explore how these key genes influence the tumor microenvironment, we performed immune infiltration analysis using CIBERSORTx. Fourth, to reveal upstream regulatory mechanisms, we constructed ceRNA networks and predicted transcription factors. Fifth, to identify potential therapeutic interventions, we conducted drug prediction and molecular docking analyses. Finally, to validate our findings at cellular resolution and understand cellular heterogeneity, we analyzed single-cell RNA sequencing data from GSE256136 (20 samples), identifying cell types, quantifying cell-cell communications, and confirming key gene expression patterns across different cell populations.

RESULTS: Our systematic analysis revealed that oxidative stress-related genes in pNET were significantly enriched in the PI3K-Akt signaling pathway, cysteine and methionine metabolism, and HIF-1 signaling pathway. BCL2L1 and PHGDH emerged as central regulators with excellent diagnostic performance (AUC > 0.9). Immune infiltration analysis demonstrated significant alterations in activated dendritic cells, memory B cells, and resting NK cells, which correlated strongly with BCL2L1 and PHGDH expression, suggesting these genes link oxidative stress to immune dysfunction. The ceRNA network centered on KCNQ1OT1 and hsa-miR-15a-5p revealed multi-layered transcriptional and post-transcriptional regulation. Drug prediction identified sertindole and cabozantinib as promising therapeutic candidates. Single-cell analysis identified 11 cell types and confirmed that endocrine cells are the primary site of BCL2L1 and PHGDH dysregulation, with extensive crosstalk between endocrine cells and T cells potentially mediating immune evasion.

CONCLUSION: Through integrated multi-omics analysis, we established that oxidative stress pathways may drive pNET progression through a coordinated mechanism involving metabolic reprogramming (via BCL2L1 and PHGDH downregulation), immune microenvironment remodeling (through altered dendritic cell and NK cell function), and complex regulatory networks. BCL2L1 and PHGDH represent potential diagnostic biomarkers and candidate therapeutic targets that require experimental validation, providing new directions for precision medicine in pNET.

PMID:41591671 | DOI:10.1007/s12672-026-04515-1

  • ✇InfoQ
  • OpenAI and Anthropic Introduce Healthcare-Focused AI Platforms Robert Krzaczyński
    OpenAI and Anthropic have announced new healthcare-oriented AI offerings that extend their models beyond general conversational use and into regulated clinical and life sciences environments. Both releases emphasize technical integration, interoperability, and governance, reflecting a shift toward AI systems designed to operate directly within existing healthcare infrastructure. By Robert Krzaczyński
     
  • 🏷️
  • ↗️

OpenAI and Anthropic Introduce Healthcare-Focused AI Platforms

✇InfoQ
By: Robert Krzaczyński
27 January 2026 at 18:10

OpenAI and Anthropic have announced new healthcare-oriented AI offerings that extend their models beyond general conversational use and into regulated clinical and life sciences environments. Both releases emphasize technical integration, interoperability, and governance, reflecting a shift toward AI systems designed to operate directly within existing healthcare infrastructure.

By Robert Krzaczyński

Publisher Correction: Best practice recommendations and considerations for designing and electronically implementing event-driven diaries in clinical trials

✇npj Digital Medicine
By: Florence D. Mowlem · Jeremiah J. Trudeau · Jill V. Platko · Jos Bloemers · Emuella Flood · Jessica L. Abel · Kelly Dumais · Paul O’Donohoe · Sabrina Grant · Ryan Naville-Cook · Onyekachukwu Illoh · Shannon D. Keith · Jing Ju · Megan Fitter · Dorothee Oberdhan · Randall Winnette · Manuela Bossi · Naomi Suminski · Sonya Eremenco · Lindsay Hughes · Amy Fasiczka · Luisana Rojas · Michelle Campbell · Scottie Kern
27 January 2026 at 08:00

npj Digital Medicine, Published online: 27 January 2026; doi:10.1038/s41746-026-02396-w

Publisher Correction: Best practice recommendations and considerations for designing and electronically implementing event-driven diaries in clinical trials
Received — 27 January 2026 ⏭ Favourites

High-Fidelity Longitudinal Patient Simulation Using Real-World Data

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yu Akagi · Tomohisa Seki · Hiromasa Ito · Toru Takiguchi · Kazuhiko Ohe · Yoshimasa Kawazoe
27 January 2026 at 13:00
arXiv:2601.17310v1 Announce Type: new Abstract: Simulation is a powerful tool for exploring uncertainty. Its potential in clinical medicine is transformative and includes personalized treatment planning and virtual clinical trials. However, simulating patient trajectories is challenging because of complex biological and sociocultural influences. Here, we show that real-world clinical records can be leveraged to empirically model patient timelines. We developed a generative simulator model that takes a patient's history as input and synthesizes fine-grained, realistic future trajectories. The model was pretrained on more than 200 million clinical records. It produced high-fidelity future timelines, closely matching event occurrence rates, laboratory test results, and temporal dynamics in real patient future data. It also accurately estimated future event probabilities, with observed-to-expected ratios consistently near 1.0 across diverse outcomes and time horizons. Our results reveal the untapped value of real-world data in electronic health records and introduce a scalable framework for in silico modeling of clinical care.

"Rebuilding" Statistics in the Age of AI: A Town Hall Discussion on Culture, Infrastructure, and Training

✇cs.AI, q-bio.NC updates on arXiv.org
By: David L. Donoho · Jian Kang · Xihong Lin · Bhramar Mukherjee · Dan Nettleton · Rebecca Nugent · Abel Rodriguez · Eric P. Xing · Tian Zheng · Hongtu Zhu
27 January 2026 at 13:00
arXiv:2601.17510v1 Announce Type: cross Abstract: This article presents the full, original record of the 2024 Joint Statistical Meetings (JSM) town hall, "Statistics in the Age of AI," which convened leading statisticians to discuss how the field is evolving in response to advances in artificial intelligence, foundation models, large-scale empirical modeling, and data-intensive infrastructures. The town hall was structured around open panel discussion and extensive audience Q&A, with the aim of eliciting candid, experience-driven perspectives rather than formal presentations or prepared statements. This document preserves the extended exchanges among panelists and audience members, with minimal editorial intervention, and organizes the conversation around five recurring questions concerning disciplinary culture and practices, data curation and "data work," engagement with modern empirical modeling, training for large-scale AI applications, and partnerships with key AI stakeholders. By providing an archival record of this discussion, the preprint aims to support transparency, community reflection, and ongoing dialogue about the evolving role of statistics in the data- and AI-centric future.

GenAI-Net: A Generative AI Framework for Automated Biomolecular Network Design

✇cs.AI, q-bio.NC updates on arXiv.org
By: Maurice Filo · Nicol\`o Rossi · Zhou Fang · Mustafa Khammash
27 January 2026 at 13:00
arXiv:2601.17582v1 Announce Type: cross Abstract: Biomolecular networks underpin emerging technologies in synthetic biology-from robust biomanufacturing and metabolic engineering to smart therapeutics and cell-based diagnostics-and also provide a mechanistic language for understanding complex dynamics in natural and ecological systems. Yet designing chemical reaction networks (CRNs) that implement a desired dynamical function remains largely manual: while a proposed network can be checked by simulation, the reverse problem of discovering a network from a behavioral specification is difficult, requiring substantial human insight to navigate a vast space of topologies and kinetic parameters with nonlinear and possibly stochastic dynamics. Here we introduce GenAI-Net, a generative AI framework that automates CRN design by coupling an agent that proposes reactions to simulation-based evaluation defined by a user-specified objective. GenAI-Net efficiently produces novel, topologically diverse solutions across multiple design tasks, including dose responses, complex logic gates, classifiers, oscillators, and robust perfect adaptation in deterministic and stochastic settings (including noise reduction). By turning specifications into families of circuit candidates and reusable motifs, GenAI-Net provides a general route to programmable biomolecular circuit design and accelerates the translation from desired function to implementable mechanisms.

The Limits of AI Data Transparency Policy: Three Disclosure Fallacies

✇cs.AI, q-bio.NC updates on arXiv.org
By: Judy Hanwen Shen · Ken Liu · Angelina Wang · Sarah H. Cen · Andy K. Zhang · Caroline Meinhardt · Daniel Zhang · Kevin Klyman · Rishi Bommasani · Daniel E. Ho
27 January 2026 at 13:00
arXiv:2601.18127v1 Announce Type: cross Abstract: Data transparency has emerged as a rallying cry for addressing concerns about AI: data quality, privacy, and copyright chief among them. Yet while these calls are crucial for accountability, current transparency policies often fall short of their intended aims. Similar to nutrition facts for food, policies aimed at nutrition facts for AI currently suffer from a limited consideration of research on effective disclosures. We offer an institutional perspective and identify three common fallacies in policy implementations of data disclosures for AI. First, many data transparency proposals exhibit a specification gap between the stated goals of data transparency and the actual disclosures necessary to achieve such goals. Second, reform attempts exhibit an enforcement gap between required disclosures on paper and enforcement to ensure compliance in fact. Third, policy proposals manifest an impact gap between disclosed information and meaningful changes in developer practices and public understanding. Informed by the social science on transparency, our analysis identifies affirmative paths for transparency that are effective rather than merely symbolic.

Unheard in the Digital Age: Rethinking AI Bias and Speech Diversity

✇cs.AI, q-bio.NC updates on arXiv.org
By: Onyedikachi Hope Amaechi-Okorie · Branislav Radeljic
27 January 2026 at 13:00
arXiv:2601.18641v1 Announce Type: cross Abstract: Speech remains one of the most visible yet overlooked vectors of inclusion and exclusion in contemporary society. While fluency is often equated with credibility and competence, individuals with atypical speech patterns are routinely marginalized. Given the current state of the debate, this article focuses on the structural biases that shape perceptions of atypical speech and are now being encoded into artificial intelligence. Automated speech recognition (ASR) systems and voice interfaces, trained predominantly on standardized speech, routinely fail to recognize or respond to diverse voices, compounding digital exclusion. As AI technologies increasingly mediate access to opportunity, the study calls for inclusive technological design, anti-bias training to minimize the impact of discriminatory algorithmic decisions, and enforceable policy reform that explicitly recognize speech diversity as a matter of equity, not merely accessibility. Drawing on interdisciplinary research, the article advocates for a cultural and institutional shift in how we value voice, urging co-created solutions that elevate the rights, representation, and realities of atypical speakers in the digital age. Ultimately, the article reframes speech inclusion as a matter of equity (not accommodation) and advocates for co-created AI systems that reflect the full spectrum of human voices.

Learning temporal embeddings from electronic health records of chronic kidney disease patients

✇cs.AI, q-bio.NC updates on arXiv.org
By: Aditya Kumar · Mario A. Cypko · Oliver Amft
27 January 2026 at 13:00
arXiv:2601.18675v1 Announce Type: cross Abstract: We investigate whether temporal embedding models trained on longitudinal electronic health records can learn clinically meaningful representations without compromising predictive performance, and how architectural choices affect embedding quality. Model-guided medicine requires representations that capture disease dynamics while remaining transparent and task agnostic, whereas most clinical prediction models are optimised for a single task. Representation learning facilitates learning embeddings that generalise across downstream tasks, and recurrent architectures are well-suited for modelling temporal structure in observational clinical data. Using the MIMIC-IV dataset, we study patients with chronic kidney disease (CKD) and compare three recurrent architectures: a vanilla LSTM, an attention-augmented LSTM, and a time-aware LSTM (T-LSTM). All models are trained both as embedding models and as direct end-to-end predictors. Embedding quality is evaluated via CKD stage clustering and in-ICU mortality prediction. The T-LSTM produces more structured embeddings, achieving a lower Davies-Bouldin Index (DBI = 9.91) and higher CKD stage classification accuracy (0.74) than the vanilla LSTM (DBI = 15.85, accuracy = 0.63) and attention-augmented LSTM (DBI = 20.72, accuracy = 0.67). For in-ICU mortality prediction, embedding models consistently outperform end-to-end predictors, improving accuracy from 0.72-0.75 to 0.82-0.83, which indicates that learning embeddings as an intermediate step is more effective than direct end-to-end learning.

Computational Phenomenology of Borderline Personality Disorder: A Comparative Evaluation of LLM-Simulated Expert Personas and Human Clinical Experts

✇cs.AI, q-bio.NC updates on arXiv.org
By: Marcin Moskalewicz · Anna Sterna · Karolina Dro\.zd\.z · Kacper Dudzic · Marek Pokropski · Paula Flores
27 January 2026 at 13:00
arXiv:2508.19008v2 Announce Type: replace Abstract: Building on a human-led thematic analysis of life-story interviews with inpatients with Borderline Personality Disorder, this study examines the capacity of large language models (OpenAI's GPT, Google's Gemini, and Anthropic's Claude) to support qualitative clinical analysis. The models were evaluated through a mixed procedure. Study A involved blinded and non-blinded expert judges in phenomenology and clinical psychology. Assessments included semantic congruence, Jaccard coefficients for overlap of outputs, multidimensional validity ratings of credibility, coherence, and the substantiveness of results, and their grounding in qualitative data. In Study B, neural methods were used to embed the theme descriptions created by humans and the models in a two-dimensional vector space to provide a computational measure of the difference between human and model semantics and linguistic style. In Study C, complementary non-expert evaluations were conducted to examine the influence of thematic verbosity on the perception of human authorship and content validity. Results of all three studies revealed variable overlap with the human analysis, with models being partly indistinguishable from, and also identifying themes originally omitted by, human researchers. The findings highlight both the variability and potential of AI-augmented thematic qualitative analysis to mitigate human interpretative bias and enhance sensitivity.

MEDIC: Comprehensive Evaluation of Leading Indicators for LLM Safety and Utility in Clinical Applications

✇cs.AI, q-bio.NC updates on arXiv.org
By: Praveenkumar Kanithi · Cl\'ement Christophe · Marco AF Pimentel · Tathagata Raha · Prateek Munjal · Nada Saadi · Hamza A Javed · Svetlana Maslenkova · Nasir Hayat · Ronnie Rajan · Shadab Khan
27 January 2026 at 13:00
arXiv:2409.07314v2 Announce Type: replace-cross Abstract: While Large Language Models (LLMs) achieve superhuman performance on standardized medical licensing exams, these static benchmarks have become saturated and increasingly disconnected from the functional requirements of clinical workflows. To bridge the gap between theoretical capability and verified utility, we introduce MEDIC, a comprehensive evaluation framework establishing leading indicators across various clinical dimensions. Beyond standard question-answering, we assess operational capabilities using deterministic execution protocols and a novel Cross-Examination Framework (CEF), which quantifies information fidelity and hallucination rates without reliance on reference texts. Our evaluation across a heterogeneous task suite exposes critical performance trade-offs: we identify a significant knowledge-execution gap, where proficiency in static retrieval does not predict success in operational tasks such as clinical calculation or SQL generation. Furthermore, we observe a divergence between passive safety (refusal) and active safety (error detection), revealing that models fine-tuned for high refusal rates often fail to reliably audit clinical documentation for factual accuracy. These findings demonstrate that no single architecture dominates across all dimensions, highlighting the necessity of a portfolio approach to clinical model deployment. As part of this investigation, we released a public leaderboard on Hugging Face.\footnote{https://huggingface.co/spaces/m42-health/MEDIC-Benchmark}

Pretrain Value, Not Reward: Decoupled Value Policy Optimization

✇cs.AI, q-bio.NC updates on arXiv.org
By: Chenghua Huang · Lu Wang · Fangkai Yang · Pu Zhao · Zhixu Li · Qingwei Lin · Dongmei Zhang · Saravan Rajmohan · Qi Zhang
27 January 2026 at 13:00
arXiv:2502.16944v2 Announce Type: replace-cross Abstract: In this paper, we explore how directly pretraining a value model simplifies and stabilizes reinforcement learning from human feedback (RLHF). In reinforcement learning, value estimation is the key to policy optimization, distinct from reward supervision. The value function predicts the \emph{return-to-go} of a partial answer, that is, how promising the partial answer is if it were continued to completion. In RLHF, however, the standard pipeline first pretrains a reward model and then learns a value function online, even though no new reward signals are available once preference data is collected. This makes critic learning redundant, as the process of training a reward model and then deriving a value model is informationally equivalent to directly pretraining a value model. Importantly, this requires no additional supervision, and our value model is trained on exactly the same data used for reward modeling. Building on this insight, we introduce \emph{Decoupled Value Policy Optimization} (DVPO), a framework that pretrains a \emph{Global Value Model} (GVM) offline and freezes it as a universal critic for policy learning. The GVM provides stable, fine-grained credit assignment without critic drift or trajectory sampling. Experiments across MT-Bench, Alpaca-Eval, and Arena-Hard demonstrate that DVPO matches or surpasses state-of-the-art RLHF methods. These results highlight RLHF can be reframed as policy-only optimization guided by a single pretrained value model.

uPVC-Net: A Universal Premature Ventricular Contraction Detection Deep Learning Algorithm

✇cs.AI, q-bio.NC updates on arXiv.org
By: Hagai Hamami · Yosef Solewicz · Daniel Zur · Yonatan Kleerekoper · Joachim A. Behar
27 January 2026 at 13:00
arXiv:2506.11238v2 Announce Type: replace-cross Abstract: Introduction: Premature Ventricular Contractions (PVCs) are common cardiac arrhythmias originating from the ventricles. Accurate detection remains challenging due to variability in electrocardiogram (ECG) waveforms caused by differences in lead placement, recording conditions, and population demographics. Methods: We developed uPVC-Net, a universal deep learning model to detect PVCs from any single-lead ECG recordings. The model is developed on four independent ECG datasets comprising a total of 8.3 million beats collected from Holter monitors and a modern wearable ECG patch. uPVC-Net employs a custom architecture and a multi-source, multi-lead training strategy. For each experiment, one dataset is held out to evaluate out-of-distribution (OOD) generalization. Results: uPVC-Net achieved an AUC between 97.8% and 99.1% on the held-out datasets. Notably, performance on wearable single-lead ECG data reached an AUC of 99.1%. Conclusion: uPVC-Net exhibits strong generalization across diverse lead configurations and populations, highlighting its potential for robust, real-world clinical deployment.

On the Fundamental Limits of LLMs at Scale

✇cs.AI, q-bio.NC updates on arXiv.org
By: Muhammad Ahmed Mohsin · Muhammad Umer · Ahsan Bilal · Zeeshan Memon · Muhammad Ibtsaam Qadir · Sagnik Bhattacharya · Hassan Rizwan · Abhiram R. Gorle · Maahe Zehra Kazmi · Nukhba Amir · Ali Subhan · Muhammad Usman Rafique · Zihao He · Pulkit Mehta · Muhammad Ali Jamshed · John M. Cioffi
27 January 2026 at 13:00
arXiv:2511.12869v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have benefited enormously from scaling, yet these gains are bounded by five fundamental limitations: (1) hallucination, (2) context compression, (3) reasoning degradation, (4) retrieval fragility, and (5) multimodal misalignment. While existing surveys describe these phenomena empirically, they lack a rigorous theoretical synthesis connecting them to the foundational limits of computation, information, and learning. This work closes that gap by presenting a unified, proof-informed framework that formalizes the innate theoretical ceilings of LLM scaling. First, computability and uncomputability imply an irreducible residue of error: for any computably enumerable model family, diagonalization guarantees inputs on which some model must fail, and undecidable queries (e.g., halting-style tasks) induce infinite failure sets for all computable predictors. Second, information-theoretic and statistical constraints bound attainable accuracy even on decidable tasks, finite description length enforces compression error, and long-tail factual knowledge requires prohibitive sample complexity. Third, geometric and computational effects compress long contexts far below their nominal size due to positional under-training, encoding attenuation, and softmax crowding. We further show how likelihood-based training favors pattern completion over inference, how retrieval under token limits suffers from semantic drift and coupling noise, and how multimodal scaling inherits shallow cross-modal alignment. Across sections, we pair theorems and empirical evidence to outline where scaling helps, where it saturates, and where it cannot progress, providing both theoretical foundations and practical mitigation paths like bounded-oracle retrieval, positional curricula, and sparse or hierarchical attention.
❌