❌

Normal view

Received — 4 November 2025 ⏭ Favourites

From Passive to Proactive: A Multi-Agent System with Dynamic Task Orchestration for Intelligent Medical Pre-Consultation

✇cs.AI, q-bio.NC updates on arXiv.org
By: ChengZhang Yu · YingRu He · Hongyan Cheng · nuo Cheng · Zhixing Liu · Dongxu Mu · Zhangrui Shen · Zhanpeng Jin
4 November 2025 at 13:00
arXiv:2511.01445v1 Announce Type: new Abstract: Global healthcare systems face critical challenges from increasing patient volumes and limited consultation times, with primary care visits averaging under 5 minutes in many countries. While pre-consultation processes encompassing triage and structured history-taking offer potential solutions, they remain limited by passive interaction paradigms and context management challenges in existing AI systems. This study introduces a hierarchical multi-agent framework that transforms passive medical AI systems into proactive inquiry agents through autonomous task orchestration. We developed an eight-agent architecture with centralized control mechanisms that decomposes pre-consultation into four primary tasks: Triage ($T_1$), History of Present Illness collection ($T_2$), Past History collection ($T_3$), and Chief Complaint generation ($T_4$), with $T_1$--$T_3$ further divided into 13 domain-specific subtasks. Evaluated on 1,372 validated electronic health records from a Chinese medical platform across multiple foundation models (GPT-OSS 20B, Qwen3-8B, Phi4-14B), the framework achieved 87.0% accuracy for primary department triage and 80.5% for secondary department classification, with task completion rates reaching 98.2% using agent-driven scheduling versus 93.1% with sequential processing. Clinical quality scores from 18 physicians averaged 4.56 for Chief Complaints, 4.48 for History of Present Illness, and 4.69 for Past History on a 5-point scale, with consultations completed within 12.7 rounds for $T_2$ and 16.9 rounds for $T_3$. The model-agnostic architecture maintained high performance across different foundation models while preserving data privacy through local deployment, demonstrating the potential for autonomous AI systems to enhance pre-consultation efficiency and quality in clinical settings.

Digital Twin based Automatic Reconfiguration of Robotic Systems in Smart Environments

✇cs.AI, q-bio.NC updates on arXiv.org
By: Angelos Alexopoulos · Agorakis Bompotas · Nikitas Rigas Kalogeropoulos · Panagiotis Kechagias · Athanasios P. Kalogeras · Christos Alexakos
4 November 2025 at 13:00
arXiv:2511.00094v1 Announce Type: cross Abstract: Robotic systems have become integral to smart environments, enabling applications ranging from urban surveillance and automated agriculture to industrial automation. However, their effective operation in dynamic settings - such as smart cities and precision farming - is challenged by continuously evolving topographies and environmental conditions. Traditional control systems often struggle to adapt quickly, leading to inefficiencies or operational failures. To address this limitation, we propose a novel framework for autonomous and dynamic reconfiguration of robotic controllers using Digital Twin technology. Our approach leverages a virtual replica of the robot's operational environment to simulate and optimize movement trajectories in response to real-world changes. By recalculating paths and control parameters in the Digital Twin and deploying the updated code to the physical robot, our method ensures rapid and reliable adaptation without manual intervention. This work advances the integration of Digital Twins in robotics, offering a scalable solution for enhancing autonomy in smart, dynamic environments.

Diffusion Models at the Drug Discovery Frontier: A Review on Generating Small Molecules versus Therapeutic Peptides

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yiquan Wang · Yahui Ma · Yuhan Chang · Jiayao Yan · Jialin Zhang · Minnuo Cai · Kai Wei
4 November 2025 at 13:00
arXiv:2511.00209v1 Announce Type: cross Abstract: Diffusion models have emerged as a leading framework in generative modeling, showing significant potential to accelerate and transform the traditionally slow and costly process of drug discovery. This review provides a systematic comparison of their application in designing two principal therapeutic modalities: small molecules and therapeutic peptides. We analyze how a unified framework of iterative denoising is adapted to the distinct molecular representations, chemical spaces, and design objectives of each modality. For small molecules, these models excel at structure-based design, generating novel, pocket-fitting ligands with desired physicochemical properties, yet face the critical hurdle of ensuring chemical synthesizability. Conversely, for therapeutic peptides, the focus shifts to generating functional sequences and designing de novo structures, where the primary challenges are achieving biological stability against proteolysis, ensuring proper folding, and minimizing immunogenicity. Despite these distinct challenges, both domains face shared hurdles: the need for more accurate scoring functions, the scarcity of high-quality experimental data, and the crucial requirement for experimental validation. We conclude that the full potential of diffusion models will be unlocked by bridging these modality-specific gaps and integrating them into automated, closed-loop Design-Build-Test-Learn (DBTL) platforms, thereby shifting the paradigm from chemical exploration to the targeted creation of novel therapeutics.

Investigating Label Bias and Representational Sources of Age-Related Disparities in Medical Segmentation

✇cs.AI, q-bio.NC updates on arXiv.org
By: Aditya Parikh · Sneha Das · Aasa Feragen
4 November 2025 at 13:00
arXiv:2511.00477v1 Announce Type: cross Abstract: Algorithmic bias in medical imaging can perpetuate health disparities, yet its causes remain poorly understood in segmentation tasks. While fairness has been extensively studied in classification, segmentation remains underexplored despite its clinical importance. In breast cancer segmentation, models exhibit significant performance disparities against younger patients, commonly attributed to physiological differences in breast density. We audit the MAMA-MIA dataset, establishing a quantitative baseline of age-related bias in its automated labels, and reveal a critical Biased Ruler effect where systematically flawed labels for validation misrepresent a model's actual bias. However, whether this bias originates from lower-quality annotations (label bias) or from fundamentally more challenging image characteristics remains unclear. Through controlled experiments, we systematically refute hypotheses that the bias stems from label quality sensitivity or quantitative case difficulty imbalance. Balancing training data by difficulty fails to mitigate the disparity, revealing that younger patient cases are intrinsically harder to learn. We provide direct evidence that systemic bias is learned and amplified when training on biased, machine-generated labels, a critical finding for automated annotation pipelines. This work introduces a systematic framework for diagnosing algorithmic bias in medical segmentation and demonstrates that achieving fairness requires addressing qualitative distributional differences rather than merely balancing case counts.

Diagnosing Hallucination Risk in AI Surgical Decision-Support: A Sequential Framework for Sequential Validation

✇cs.AI, q-bio.NC updates on arXiv.org
By: Dong Chen · Yanzhe Wei · Zonglin He · Guan-Ming Kuang · Canhua Ye · Meiru An · Huili Peng · Yong Hu · Huiren Tao · Kenneth MC Cheung
4 November 2025 at 13:00
arXiv:2511.00588v1 Announce Type: cross Abstract: Large language models (LLMs) offer transformative potential for clinical decision support in spine surgery but pose significant risks through hallucinations, which are factually inconsistent or contextually misaligned outputs that may compromise patient safety. This study introduces a clinician-centered framework to quantify hallucination risks by evaluating diagnostic precision, recommendation quality, reasoning robustness, output coherence, and knowledge alignment. We assessed six leading LLMs across 30 expert-validated spinal cases. DeepSeek-R1 demonstrated superior overall performance (total score: 86.03 $\pm$ 2.08), particularly in high-stakes domains such as trauma and infection. A critical finding reveals that reasoning-enhanced model variants did not uniformly outperform standard counterparts: Claude-3.7-Sonnet's extended thinking mode underperformed relative to its standard version (80.79 $\pm$ 1.83 vs. 81.56 $\pm$ 1.92), indicating extended chain-of-thought reasoning alone is insufficient for clinical reliability. Multidimensional stress-testing exposed model-specific vulnerabilities, with recommendation quality degrading by 7.4% under amplified complexity. This decline contrasted with marginal improvements in rationality (+2.0%), readability (+1.7%) and diagnosis (+4.7%), highlighting a concerning divergence between perceived coherence and actionable guidance. Our findings advocate integrating interpretability mechanisms (e.g., reasoning chain visualization) into clinical workflows and establish a safety-aware validation framework for surgical LLM deployment.

How Far Are Surgeons from Surgical World Models? A Pilot Study on Zero-shot Surgical Video Generation with Expert Assessment

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhen Chen · Qing Xu · Jinlin Wu · Biao Yang · Yuhao Zhai · Geng Guo · Jing Zhang · Yinlu Ding · Nassir Navab · Jiebo Luo
4 November 2025 at 13:00
arXiv:2511.01775v1 Announce Type: cross Abstract: Foundation models in video generation are demonstrating remarkable capabilities as potential world models for simulating the physical world. However, their application in high-stakes domains like surgery, which demand deep, specialized causal knowledge rather than general physical rules, remains a critical unexplored gap. To systematically address this challenge, we present SurgVeo, the first expert-curated benchmark for video generation model evaluation in surgery, and the Surgical Plausibility Pyramid (SPP), a novel, four-tiered framework tailored to assess model outputs from basic appearance to complex surgical strategy. On the basis of the SurgVeo benchmark, we task the advanced Veo-3 model with a zero-shot prediction task on surgical clips from laparoscopic and neurosurgical procedures. A panel of four board-certified surgeons evaluates the generated videos according to the SPP. Our results reveal a distinct "plausibility gap": while Veo-3 achieves exceptional Visual Perceptual Plausibility, it fails critically at higher levels of the SPP, including Instrument Operation Plausibility, Environment Feedback Plausibility, and Surgical Intent Plausibility. This work provides the first quantitative evidence of the chasm between visually convincing mimicry and causal understanding in surgical AI. Our findings from SurgVeo and the SPP establish a crucial foundation and roadmap for developing future models capable of navigating the complexities of specialized, real-world healthcare domains.
  • ✇cs.AI, q-bio.NC updates on arXiv.org
  • The Limits of AI Explainability: An Algorithmic Information Theory Approach Shrisha Rao
    arXiv:2504.20676v2 Announce Type: replace Abstract: This paper establishes a theoretical foundation for understanding the fundamental limits of AI explainability through algorithmic information theory. We formalize explainability as the approximation of complex models by simpler ones, quantifying both approximation error and explanation complexity using Kolmogorov complexity. Our key theoretical contributions include: (1) a complexity gap theorem proving that any explanation significantly simpl
     
  • 🏷️
  • ↗️

The Limits of AI Explainability: An Algorithmic Information Theory Approach

✇cs.AI, q-bio.NC updates on arXiv.org
By: Shrisha Rao
4 November 2025 at 13:00
arXiv:2504.20676v2 Announce Type: replace Abstract: This paper establishes a theoretical foundation for understanding the fundamental limits of AI explainability through algorithmic information theory. We formalize explainability as the approximation of complex models by simpler ones, quantifying both approximation error and explanation complexity using Kolmogorov complexity. Our key theoretical contributions include: (1) a complexity gap theorem proving that any explanation significantly simpler than the original model must differ from it on some inputs; (2) precise bounds showing that explanation complexity grows exponentially with input dimension but polynomially with error tolerance for Lipschitz functions; and (3) a characterization of the gap between local and global explainability, demonstrating that local explanations can be significantly simpler while maintaining accuracy in relevant regions. We further establish a regulatory impossibility theorem proving that no governance framework can simultaneously pursue unrestricted AI capabilities, human-interpretable explanations, and negligible error. These results highlight considerations likely to be relevant to the design, evaluation, and oversight of explainable AI systems.
  • ✇cs.AI, q-bio.NC updates on arXiv.org
  • Will Humanity Be Rendered Obsolete by AI? Mohamed El Louadi · Emna Ben Romdhane
    arXiv:2510.22814v2 Announce Type: replace Abstract: This article analyzes the existential risks artificial intelligence (AI) poses to humanity, tracing the trajectory from current AI to ultraintelligence. Drawing on Irving J. Good and Nick Bostrom's theoretical work, plus recent publications (AI 2027; If Anyone Builds It, Everyone Dies), it explores AGI and superintelligence. Considering machines' exponentially growing cognitive power and hypothetical IQs, it addresses the ethical and existenti
     
  • 🏷️
  • ↗️

Will Humanity Be Rendered Obsolete by AI?

✇cs.AI, q-bio.NC updates on arXiv.org
By: Mohamed El Louadi · Emna Ben Romdhane
4 November 2025 at 13:00
arXiv:2510.22814v2 Announce Type: replace Abstract: This article analyzes the existential risks artificial intelligence (AI) poses to humanity, tracing the trajectory from current AI to ultraintelligence. Drawing on Irving J. Good and Nick Bostrom's theoretical work, plus recent publications (AI 2027; If Anyone Builds It, Everyone Dies), it explores AGI and superintelligence. Considering machines' exponentially growing cognitive power and hypothetical IQs, it addresses the ethical and existential implications of an intelligence vastly exceeding humanity's, fundamentally alien. Human extinction may result not from malice, but from uncontrollable, indifferent cognitive superiority.

A Survey on Cache Methods in Diffusion Models: Toward Efficient Multi-Modal Generation

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jiacheng Liu · Xinyu Wang · Yuqi Lin · Zhikai Wang · Peiru Wang · Peiliang Cai · Qinming Zhou · Zhengan Yan · Zexuan Yan · Zhengyi Shi · Chang Zou · Yue Ma · Linfeng Zhang
4 November 2025 at 13:00
arXiv:2510.19755v3 Announce Type: replace-cross Abstract: Diffusion Models have become a cornerstone of modern generative AI for their exceptional generation quality and controllability. However, their inherent \textit{multi-step iterations} and \textit{complex backbone networks} lead to prohibitive computational overhead and generation latency, forming a major bottleneck for real-time applications. Although existing acceleration techniques have made progress, they still face challenges such as limited applicability, high training costs, or quality degradation. Against this backdrop, \textbf{Diffusion Caching} offers a promising training-free, architecture-agnostic, and efficient inference paradigm. Its core mechanism identifies and reuses intrinsic computational redundancies in the diffusion process. By enabling feature-level cross-step reuse and inter-layer scheduling, it reduces computation without modifying model parameters. This paper systematically reviews the theoretical foundations and evolution of Diffusion Caching and proposes a unified framework for its classification and analysis. Through comparative analysis of representative methods, we show that Diffusion Caching evolves from \textit{static reuse} to \textit{dynamic prediction}. This trend enhances caching flexibility across diverse tasks and enables integration with other acceleration techniques such as sampling optimization and model distillation, paving the way for a unified, efficient inference framework for future multimodal and interactive applications. We argue that this paradigm will become a key enabler of real-time and efficient generative AI, injecting new vitality into both theory and practice of \textit{Efficient Generative Intelligence}.

Generative Artificial Intelligence in Medical Education: Enhancing Critical Thinking or Undermining Cognitive Autonomy?

✇Journal of Medical Internet Research
By: Juan S Izquierdo-Condoy · Marlon Arias-Intriago · Andrea Tello-De-la-Torre · Felix Busch · Esteban Ortiz-Prado
4 November 2025 at 05:00
Generative artificial intelligence (GenAI) enables the production of coherent and contextually relevant text by processing large-scale linguistic datasets. Tools such as ChatGPT, Gemini, Claude, and LLaMA are increasingly integrated into medical education, assisting students with a range of tasks, including clinical reasoning, literature review, scientific writing, and formative assessment. Although these tools offer significant advantages in terms of productivity, personalization, and cognitive support, their impact on critical thinking—a cornerstone of medical education—remains uncertain. The aim of this viewpoint paper is to critically assess the influence of GenAI on critical thinking within medical training, examining both its potential to enhance cognitive skills and the risks it poses to cognitive autonomy. Users have reported increased efficiency and improved linguistic output; however, concerns have also been raised regarding the risk of cognitive overreliance. Current evidence presents a mixed picture, indicating both improvements in learner engagement and potential drawbacks such as passivity or susceptibility to misinformation. Without curricular integration that prioritizes ethical use, prompt engineering, and critical evaluation, GenAI may compromise the cognitive autonomy of medical students. Conversely, when thoughtfully embedded into pedagogical frameworks, these tools can act as cognitive enhancers—supporting, rather than replacing, clinical reasoning. Medical education must adapt to ensure that future physicians engage with GenAI in a critical, ethical, and context-aware manner, especially in complex decision-making scenarios. This transformation demands not only technological fluency but also reflective practice and sustained oversight by faculty and academic institutions.

Can human connection amplify digital health outcomes? Familial involvement in a mobile health app

✇npj Digital Medicine
By: Elizabeth J. Enichen · Kimia Heydari · Ben Li · Joseph C. Kvedar
3 November 2025 at 08:00

npj Digital Medicine, Published online: 03 November 2025; doi:10.1038/s41746-025-02037-8

In “A Randomized Controlled Trial of Mobile Intervention Using Health Support Bubbles to Prevent Social Frailty”, Hayashi et al. investigated the effects of using a mobile health app with family or individually. Greater improvements in social behavior and frailty were noted in participants who used the app with family. In an era of remote healthcare and app-based health interventions, Hayashi et al.’s study reminds of the importance of human connection.

Leveraging artificial intelligence to validate traditional biomarkers and drug targets in liver cancer recovery: a mini review

✇Omics in Hepatocellular
By: Shengjian Wu · Xiaoqiao Chen · Yuxiu Ji · Chi Zhang · Yujie Xie · Bin Liang
3 November 2025 at 19:00

Front Pharmacol. 2025 Oct 17;16:1697608. doi: 10.3389/fphar.2025.1697608. eCollection 2025.

ABSTRACT

Hepatocellular carcinoma (HCC) remains a leading cause of cancer death, and recovery after therapy is shaped by heterogeneous etiologies, genomes and microenvironments. Targeted and immunotherapy combinations have broadened first-line options; yet durable benefit is uneven, and serum/imaging anchors (AFP, AFP-L3%, PIVKA-II, LI-RADS/mRECIST) incompletely resolve residual disease or functional restoration. In this review we summarise AI-enabled radiology, digital pathology and multi-omic/liquid-biopsy analytics that test and refine traditional biomarkers and drug-target readouts, and appraise translational opportunities in composite surveillance and recovery forecasting. We also discuss enduring challenges-including assay standardisation, spectrum bias, data leakage, domain shift and limited prospective external validation-that temper implementation. By integrating established anchors (AFP/AFP-L3%, PIVKA-II, ALBI, contrast-enhanced hallmarks) with AI-derived signals (radiomics/pathomics, cfDNA methylation) and pathway contexts (VEGF-VEGFR, WNT/β-catenin), emerging strategies align predictions with clinical endpoints, individualise therapy and chart hepatic function. Our synthesis provides an appraisal of AI-traditional integration in liver cancer recovery and outlines pragmatic standards-analytical robustness, transparent reporting and prospective, guideline-conformant evaluation-required for clinical adoption. We hope these insights will aid researchers and clinicians as they implement more effective, individualised monitoring and treatment pathways.

PMID:41181592 | PMC:PMC12575369 | DOI:10.3389/fphar.2025.1697608

Artificial intelligence in pancreatitis: A narrative review on advancing precision diagnosis, prognosis, and therapeutic strategies

✇(Multiomics OR Omics) AND (Pancreatic)
By: Xi-Yue Zhang · Meng-Di Hu · Diliare Maimaitijiang · Tao Wang · Lin Wang
3 November 2025 at 19:00

World J Gastroenterol. 2025 Oct 21;31(39):110971. doi: 10.3748/wjg.v31.i39.110971.

ABSTRACT

Pancreatitis poses persistent diagnostic and therapeutic challenges due to its heterogeneous clinical presentation, variable disease course, and lack of targeted interventions. Conventional tools, such as serum enzymes, cross-sectional imaging and clinical scoring systems, often exhibit limited sensitivity and prognostic value, especially during early or atypical stages. Moreover, therapeutic development remains slow, with limited progress toward personalized or mechanism-based strategies. These limitations highlight a critical need for integrative data-driven approaches. Artificial intelligence (AI) has emerged as a promising tool to enhance clinical decision-making in pancreatitis. This narrative review synthesizes recent progress in AI applications across three domains. First, AI-enabled diagnostic platforms incorporating radiomics, deep learning-based imaging analysis, and biomarker optimization have improved early detection and differentiation of pancreatic diseases. Second, AI-driven prognostic models now allow real-time severity prediction, complication forecasting, and recurrence risk assessment, some of which have been deployed in hospital information systems for intensive care units and mortality risk triage. Third, AI-assisted drug discovery and network pharmacology, particularly in combination with traditional Chinese medicine, have revealed novel therapeutic opportunities. Despite encouraging developments, challenges remain in data standardization, model transparency and clinical validation. A multidisciplinary strategy integrating omics data, longitudinal monitoring and pharmacological modeling may help bridge current gaps and advance precision medicine in pancreatitis care.

PMID:41180795 | PMC:PMC12576603 | DOI:10.3748/wjg.v31.i39.110971

Received — 3 November 2025 ⏭ Favourites

The Denario project: Deep knowledge AI agents for scientific discovery

✇cs.AI, q-bio.NC updates on arXiv.org
By: Francisco Villaescusa-Navarro · Boris Bolliet · Pablo Villanueva-Domingo · Adrian E. Bayer · Aidan Acquah · Chetana Amancharla · Almog Barzilay-Siegal · Pablo Bermejo · Camille Bilodeau · Pablo C\'ardenas Ram\'irez · Miles Cranmer · Urbano L. Fran\c{c}a · ChangHoon Hahn · Yan-Fei Jiang · Raul Jimenez · Jun-Young Lee · Antonio Lerario · Osman Mamun · Thomas Meier · Anupam A. Ojha · Pavlos Protopapas · Shimanto Roy · David N. Spergel · Pedro Taranc\'on-\'Alvarez · Ujjwal Tiwari · Matteo Viel · Digvijay Wadekar · Chi Wang · Bonny Y. Wang · Licong Xu · Yossi Yovel · Shuwen Yue · Wen-Han Zhou · Qiyao Zhu · Jiajun Zou · \'I\~nigo Zubeldia
3 November 2025 at 13:00
arXiv:2510.26887v1 Announce Type: new Abstract: We present Denario, an AI multi-agent system designed to serve as a scientific research assistant. Denario can perform many different tasks, such as generating ideas, checking the literature, developing research plans, writing and executing code, making plots, and drafting and reviewing a scientific paper. The system has a modular architecture, allowing it to handle specific tasks, such as generating an idea, or carrying out end-to-end scientific analysis using Cmbagent as a deep-research backend. In this work, we describe in detail Denario and its modules, and illustrate its capabilities by presenting multiple AI-generated papers generated by it in many different scientific disciplines such as astrophysics, biology, biophysics, biomedical informatics, chemistry, material science, mathematical physics, medicine, neuroscience and planetary science. Denario also excels at combining ideas from different disciplines, and we illustrate this by showing a paper that applies methods from quantum physics and machine learning to astrophysical data. We report the evaluations performed on these papers by domain experts, who provided both numerical scores and review-like feedback. We then highlight the strengths, weaknesses, and limitations of the current system. Finally, we discuss the ethical implications of AI-driven research and reflect on how such technology relates to the philosophy of science. We publicly release the code at https://github.com/AstroPilot-AI/Denario. A Denario demo can also be run directly on the web at https://huggingface.co/spaces/astropilot-ai/Denario, and the full app will be deployed on the cloud.

Causal Masking on Spatial Data: An Information-Theoretic Case for Learning Spatial Datasets with Unimodal Language Models

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jared Junkin · Samuel Nathanson
3 November 2025 at 13:00
arXiv:2510.27009v1 Announce Type: new Abstract: Language models are traditionally designed around causal masking. In domains with spatial or relational structure, causal masking is often viewed as inappropriate, and sequential linearizations are instead used. Yet the question of whether it is viable to accept the information loss introduced by causal masking on nonsequential data has received little direct study, in part because few domains offer both spatial and sequential representations of the same dataset. In this work, we investigate this issue in the domain of chess, which naturally supports both representations. We train language models with bidirectional and causal self-attention mechanisms on both spatial (board-based) and sequential (move-based) data. Our results show that models trained on spatial board states - \textit{even with causal masking} - consistently achieve stronger playing strength than models trained on sequential data. While our experiments are conducted on chess, our results are methodological and may have broader implications: applying causal masking to spatial data is a viable procedure for training unimodal LLMs on spatial data, and in some domains is even preferable to sequentialization.

Adaptive Data Flywheel: Applying MAPE Control Loops to AI Agent Improvement

✇cs.AI, q-bio.NC updates on arXiv.org
By: Aaditya Shukla · Sidney Knowles · Meenakshi Madugula · Dave Farris · Ryan Angilly · Santiago Pombo · Anbang Xu · Lu An · Abhinav Balasubramanian · Tan Yu · Jiaxiang Ren · Rama Akkiraju
3 November 2025 at 13:00
arXiv:2510.27051v1 Announce Type: new Abstract: Enterprise AI agents must continuously adapt to maintain accuracy, reduce latency, and remain aligned with user needs. We present a practical implementation of a data flywheel in NVInfo AI, NVIDIA's Mixture-of-Experts (MoE) Knowledge Assistant serving over 30,000 employees. By operationalizing a MAPE-driven data flywheel, we built a closed-loop system that systematically addresses failures in retrieval-augmented generation (RAG) pipelines and enables continuous learning. Over a 3-month post-deployment period, we monitored feedback and collected 495 negative samples. Analysis revealed two major failure modes: routing errors (5.25\%) and query rephrasal errors (3.2\%). Using NVIDIA NeMo microservices, we implemented targeted improvements through fine-tuning. For routing, we replaced a Llama 3.1 70B model with a fine-tuned 8B variant, achieving 96\% accuracy, a 10x reduction in model size, and 70\% latency improvement. For query rephrasal, fine-tuning yielded a 3.7\% gain in accuracy and a 40\% latency reduction. Our approach demonstrates how human-in-the-loop (HITL) feedback, when structured within a data flywheel, transforms enterprise AI agents into self-improving systems. Key learnings include approaches to ensure agent robustness despite limited user feedback, navigating privacy constraints, and executing staged rollouts in production. This work offers a repeatable blueprint for building robust, adaptive enterprise AI agents capable of learning from real-world usage at scale.

Glia: A Human-Inspired AI for Automated Systems Design and Optimization

✇cs.AI, q-bio.NC updates on arXiv.org
By: Pouya Hamadanian · Pantea Karimi · Arash Nasr-Esfahany · Kimia Noorbakhsh · Joseph Chandler · Ali ParandehGheibi · Mohammad Alizadeh · Hari Balakrishnan
3 November 2025 at 13:00
arXiv:2510.27176v1 Announce Type: new Abstract: Can an AI autonomously design mechanisms for computer systems on par with the creativity and reasoning of human experts? We present Glia, an AI architecture for networked systems design that uses large language models (LLMs) in a human-inspired, multi-agent workflow. Each agent specializes in reasoning, experimentation, and analysis, collaborating through an evaluation framework that grounds abstract reasoning in empirical feedback. Unlike prior ML-for-systems methods that optimize black-box policies, Glia generates interpretable designs and exposes its reasoning process. When applied to a distributed GPU cluster for LLM inference, it produces new algorithms for request routing, scheduling, and auto-scaling that perform at human-expert levels in significantly less time, while yielding novel insights into workload behavior. Our results suggest that by combining reasoning LLMs with structured experimentation, an AI can produce creative and understandable designs for complex systems problems.

An In-depth Study of LLM Contributions to the Bin Packing Problem

✇cs.AI, q-bio.NC updates on arXiv.org
By: Julien Herrmann · Guillaume Pallez
3 November 2025 at 13:00
arXiv:2510.27353v1 Announce Type: new Abstract: Recent studies have suggested that Large Language Models (LLMs) could provide interesting ideas contributing to mathematical discovery. This claim was motivated by reports that LLM-based genetic algorithms produced heuristics offering new insights into the online bin packing problem under uniform and Weibull distributions. In this work, we reassess this claim through a detailed analysis of the heuristics produced by LLMs, examining both their behavior and interpretability. Despite being human-readable, these heuristics remain largely opaque even to domain experts. Building on this analysis, we propose a new class of algorithms tailored to these specific bin packing instances. The derived algorithms are significantly simpler, more efficient, more interpretable, and more generalizable, suggesting that the considered instances are themselves relatively simple. We then discuss the limitations of the claim regarding LLMs' contribution to this problem, which appears to rest on the mistaken assumption that the instances had previously been studied. Our findings instead emphasize the need for rigorous validation and contextualization when assessing the scientific value of LLM-generated outputs.

ToolScope: An Agentic Framework for Vision-Guided and Long-Horizon Tool Use

✇cs.AI, q-bio.NC updates on arXiv.org
By: Mengjie Deng · Guanting Dong · Zhicheng Dou
3 November 2025 at 13:00
arXiv:2510.27363v1 Announce Type: new Abstract: Recently, large language models (LLMs) have demonstrated remarkable problem-solving capabilities by autonomously integrating with external tools for collaborative reasoning. However, due to the inherently complex and diverse nature of multimodal information, enabling multimodal large language models (MLLMs) to flexibly and efficiently utilize external tools during reasoning remains an underexplored challenge. In this work, we introduce ToolScope, an agentic framework designed to unify global planning with local multimodal perception, adopting a specialized Perceive tool to mitigates visual context degradation in long-horizon VQA task. ToolScope comprises three primary components: the Global Navigator, the Agentic Executor, and the Response Synthesizer. The Global Navigator functions as a "telescope", offering high-level strategic guidance. The Agentic Executor operates iteratively to augment MLLM with local perception through the integration of external tools-Search, Code, and Perceive. Finally, the Response Synthesizer consolidates and organizes the reasoning process into a coherent, user-friendly output. We evaluate ToolScope on four VQA benchmarks across diverse domains, including VQA 2.0, ScienceQA, MAT-Search and MathVista. It demonstrates strong generalization capabilities, achieving an average performance improvement of up to +6.69% across all datasets.
❌