cs.AI papers | Gist.Science

A Geometric Taxonomy of Hallucinations in LLMs

This paper proposes a geometric taxonomy of LLM hallucinations into three distinct types (unfaithfulness, confabulation, and factual error) and introduces corresponding detection metrics, the Semantic Grounding Index and Directional Grounding Index, which effectively identify unfaithful and confabulated outputs while revealing that apparent signals for factual errors in existing benchmarks often stem from stylistic annotation confounds rather than genuine geometric distinctions.

Javier Marín2026-03-10💬 cs.CL

TrasMuon: Trust-Region Adaptive Scaling for Orthogonalized Momentum Optimizers

TrasMuon is a novel optimizer that enhances the stability and convergence of Muon-style methods by integrating global RMS calibration and energy-based trust-region clipping to preserve near-isometric geometry while mitigating sensitivity to step-size hyperparameters and high-energy outliers.

Peng Cheng, Jiucheng Zang, Qingnan Li, Liheng Ma, Yufei Cui, Yingxue Zhang, Boxing Chen, Ming Jian, Wen Tong2026-03-10🤖 cs.LG

Can a Lightweight Automated AI Pipeline Solve Research-Level Mathematical Problems?

This paper demonstrates that a lightweight, automated AI pipeline integrating next-generation large language models with citation-based verification can successfully generate and solve sophisticated, research-grade mathematical problems, including previously unpublished questions, with verified results and open-sourced tools.

Lve Meng (University of Science,Technology of China, Zhongguancun Academy), Weilong Zhao (Université Paris Cité), Yanzhi Zhang (Zhongguancun Academy), Haoxiang Guan (Zhongguancun Academy), Jiyan He (Zhongguancun Academy)2026-03-10🔢 math

Mean Flow Policy with Instantaneous Velocity Constraint for One-step Action Generation

This paper introduces the Mean Velocity Policy (MVP), a novel one-step generative policy that employs an Instantaneous Velocity Constraint (IVC) to theoretically guarantee high expressiveness while achieving state-of-the-art performance and significantly faster training and inference speeds on challenging robotic manipulation tasks compared to existing flow-based baselines.

Guojian Zhan, Letian Tao, Pengcheng Wang, Yixiao Wang, Yiheng Li, Yuxin Chen, Hongyang Li, Masayoshi Tomizuka, Shengbo Eben Li2026-03-10🤖 cs.LG

Pawsterior: Variational Flow Matching for Structured Simulation-Based Inference

The paper introduces Pawsterior, a variational flow-matching framework that enhances simulation-based inference by incorporating geometric confinement for structured domains and enabling the handling of discrete latent structures, thereby improving posterior fidelity and expanding applicability to complex physical systems.

Jorge Carrasco-Pollo, Floor Eijkelboom, Jan-Willem van de Meent2026-03-10🤖 cs.LG

Explainable Token-level Noise Filtering for LLM Fine-tuning Datasets

This paper introduces XTF, an explainable framework that improves LLM fine-tuning performance by decomposing token contributions into reasoning importance, knowledge novelty, and task relevance to identify and mask noisy tokens, achieving up to a 13.7% improvement across math, code, and medical tasks.

Yuchen Yang, Wenze Lin, Enhao Huang, Zhixuan Chu, Hongbin Zhou, Lan Tao, Yiming Li, Zhan Qin, Kui Ren2026-03-10💬 cs.CL

LongAudio-RAG: Event-Grounded Question Answering over Multi-Hour Long Audio

LongAudio-RAG is a hybrid edge-cloud framework that enables precise, low-hallucination question answering over multi-hour audio streams by converting recordings into timestamped event records for SQL-based retrieval, which then grounds Large Language Model responses in structured evidence rather than raw audio.

Naveen Vakada, Kartik Hegde, Arvind Krishna Sridhar, Yinyi Guo, Erik Visser2026-03-10🤖 cs.LG

Understand Then Memory: A Cognitive Gist-Driven RAG Framework with Global Semantic Diffusion

CogitoRAG is a novel Retrieval-Augmented Generation framework inspired by human episodic memory that enhances complex reasoning and reduces hallucinations by extracting semantic gists into a multi-dimensional knowledge graph, utilizing query decomposition and entity diffusion for associative retrieval, and employing a fusion-based reranking algorithm to deliver high-density evidence.

Pengcheng Zhou, Haochen Li, Zhiqiang Nie, JiaLe Chen, Qing Gong, Weizhen Zhang, Chun Yu2026-03-10💬 cs.CL

Condition-Gated Reasoning for Context-Dependent Biomedical Question Answering

This paper introduces CondMedQA, the first benchmark for conditional biomedical question answering, and proposes Condition-Gated Reasoning (CGR), a framework that constructs condition-aware knowledge graphs to dynamically prune reasoning paths based on patient-specific factors, thereby improving the reliability of medical decision-making.

Jash Rajesh Parekh, Wonbin Kweon, Joey Chan, Rezarta Islamaj, Robert Leaman, Pengcheng Jiang, Chih-Hsuan Wei, Zhizheng Wang, Zhiyong Lu, Jiawei Han2026-03-10💬 cs.CL

Conformal Tradeoffs: Guarantees Beyond Coverage

This paper introduces a framework for operational certification of split conformal predictors that moves beyond marginal coverage by providing finite-sample guarantees for critical deployment metrics like commitment frequency and error exposure through Small-Sample Beta Correction, an independent audit-based auditing protocol, and a geometric analysis of Pareto trade-offs.

Petrus H. Zwart2026-03-10🤖 cs.LG

ABD: Default Exception Abduction in Finite First Order Worlds

This paper introduces ABD, a benchmark for default-exception abduction in finite first-order worlds that evaluates ten frontier LLMs on their ability to generate sparse, satisfiability-restoring formulas across three observation regimes, revealing that while models achieve high validity, they struggle with parsimony and exhibit distinct generalization failures.

Serafim Batzoglou2026-03-10✓ Author reviewed ⓘ💻 cs

INDUCTION: Finite-Structure Concept Synthesis in First-Order Logic

This paper introduces INDUCTION, a benchmark designed to evaluate the ability of AI models to synthesize compact, generalizable first-order logic formulas that explain target predicates across small finite relational worlds, revealing distinct performance patterns and generalization strategies among recent elite models.

Serafim Batzoglou2026-03-10💻 cs

Characterizing MARL for Energy Control: A Multi-KPI Benchmark on the CityLearn Environment

This paper establishes a comprehensive multi-KPI benchmark for Multi-Agent Reinforcement Learning in urban energy management using the CityLearn environment, demonstrating that Decentralized Training with Decentralized Execution (DTDE) consistently outperforms Centralized Training with Decentralized Execution (CTDE) in both average and worst-case performance while offering greater resilience and sustainability.

Aymen Khouja, Imen Jendoubi, Oumayma Mahjoub, Oussama Mahfoudhi, Ruan De Kock, Siddarth Singh, Claude Formanek2026-03-10🤖 cs.LG

MrBERT: Modern Multilingual Encoders via Vocabulary, Domain, and Dimensional Adaptation

The paper introduces MrBERT, a family of efficient, open-source multilingual encoders built on the ModernBERT architecture that achieves state-of-the-art performance in specific languages and specialized domains while leveraging Matryoshka Representation Learning to reduce inference and storage costs.

Daniel Tamayo, Iñaki Lacunza, Paula Rivera-Hidalgo, Severino Da Dalt, Javier Aula-Blasco, Aitor Gonzalez-Agirre, Marta Villegas2026-03-10🤖 cs.LG

ARLArena: A Unified Framework for Stable Agentic Reinforcement Learning

This paper introduces ARLArena, a unified framework that systematically analyzes training instability in agentic reinforcement learning to derive SAMPO, a stable optimization method that ensures consistent performance across diverse agentic tasks.

Xiaoxuan Wang, Han Zhang, Haixin Wang, Yidan Shi, Ruoyan Li, Kaiqiao Han, Chenyi Tong, Haoran Deng, Renliang Sun, Alexander Taylor, Yanqiao Zhu, Jason Cong, Yizhou Sun, Wei Wang2026-03-10💻 cs

CryoNet.Refine: A One-step Diffusion Model for Rapid Refinement of Structural Models with Cryo-EM Density Map Restraints

CryoNet.Refine is a novel one-step diffusion model that automates and accelerates the refinement of atomic structures against cryo-EM density maps, outperforming traditional tools like Phenix in both model-map correlation and geometric quality while supporting diverse protein and nucleic acid complexes.

Fuyao Huang, Xiaozhu Yu, Kui Xu, Qiangfeng Cliff Zhang2026-03-10💻 cs

Vibe Researching as Wolf Coming: Can AI Agents with Skills Replace or Augment Social Scientists?

This paper argues that AI agents equipped with specialized skills can augment, but not fully replace, social scientists by executing codifiable research tasks autonomously through "vibe researching," while highlighting the enduring necessity of human theoretical originality and tacit knowledge alongside the profession's emerging risks of stratification and pedagogical crisis.

Yongjun Zhang2026-03-10💻 cs

A Mathematical Theory of Agency and Intelligence

This paper introduces "bipredictability" (P) as a fundamental, bounded measure of shared information between observations, actions, and outcomes to distinguish mere agency from true intelligence, demonstrating that current AI systems lack the self-monitoring feedback loops necessary for adaptive learning and proposing a thalamocortical-inspired architecture to restore it.

Wael Hafez, Chenan Wei, Rodrigo Pena, Amir Nazeri, Cameron Reid2026-03-10🔢 math

Autoregressive Visual Decoding from EEG Signals

The paper introduces AVDE, a lightweight and efficient autoregressive framework that leverages contrastive learning and multi-scale token prediction to decode EEG signals into coherent images, outperforming state-of-the-art methods with significantly fewer parameters while mimicking the hierarchical nature of human visual perception.

Sicheng Dai, Hongwang Xiao, Shan Yu, Qiwei Ye2026-03-10🤖 cs.LG

Decomposing Physician Disagreement in HealthBench

This paper analyzes physician disagreement in the HealthBench dataset, revealing that while the majority of variance is structural and irreducible, a small but actionable portion stems from reducible uncertainties like missing context, suggesting that improving evaluation design to close information gaps could meaningfully reduce disagreement on borderline medical AI cases.

Satya Borgohain, Roy Mariathas2026-03-10💻 cs

← Previous Next →