LIVE · refreshes every 20 min

updated Jul 19, 17:25 UTC

110010.art / archive

Research

Jul 16

3d ago arXiv paper introduces ADR task for LLMs; creates ADR-bench benchmark for foresight analysis arXiv cs.CL 3d ago arXiv paper introduces cross-device agent benchmark; addresses limitations of single-environment evaluations arXiv cs.CL 3d ago arXiv paper introduces GFlowRL, a distribution-matching RL approach for large language models; scaling remains challenging due to model size, rollout horizon, reward noise, and distributed systems. arXiv cs.CL 3d ago arXiv paper introduces handbook for making evolving AI agent harnesses readable, navigable, and editable arXiv cs.AI 3d ago arXiv paper introduces HEDGEHOG for hierarchical evaluation of drug generators; existing metrics poorly reflect medicinal plausibility arXiv cs.LG 3d ago arXiv paper introduces multi-agent collaborative reasoning with tool-augmented evidence for urban region profiling; existing methods rely on multimodal representation learning arXiv cs.AI 3d ago arXiv paper introduces risk-based method for event-triggered LLM invocation in streaming systems arXiv cs.LG 3d ago arXiv paper introduces SPINE framework to address deployment challenges in Embodied AI by debugging bimanual robots. arXiv cs.AI 3d ago arXiv paper introduces SteinGate, a safe reinforcement learning method using Stein Discrepancy to address tail events arXiv cs.LG 3d ago arXiv paper introduces theory-level autoformalization, shifting from isolated statements to unified formal knowledge bases arXiv cs.AI 3d ago arXiv paper presents prefix-to-prefix data-driven method for SimulST without architectural changes arXiv cs.CL 3d ago arXiv paper proposes accuracy-preserving stability regularization for large-scale retail demand forecasting; evaluates training-time penalty on consecutive movements to improve forecast stability without sacrificing point accuracy. arXiv cs.LG 3d ago arXiv paper proposes AI-native framework for underwriting, pricing, and contract design of agentic AI deployments arXiv cs.AI 3d ago arXiv paper proposes EXECUTE-ASK-REFUSE routing for context-aware human intervention in LLM agents; addresses limitations of binary action safety labeling arXiv cs.AI 3d ago arXiv paper proposes graph-based framework to enhance molecular property prediction in small language models arXiv cs.AI 3d ago arXiv paper proposes inductive logic programming for explainable reinforcement learning arXiv cs.AI 3d ago arXiv paper proposes learned adaptive memory management for LLM agents to address static heuristic limitations arXiv cs.CL 3d ago arXiv paper proposes probabilistic extension of neuro-symbolic AI using Belnap's IFOL_B with Nilsson's probability structure. arXiv cs.AI 3d ago arXiv study examines LLM agents' adaptation to hidden tool reliability shifts; benchmark uses redundancies in tool-skill libraries arXiv cs.AI 3d ago arXiv study introduces Fair Oracle to measure perception-action gap in LLM agents arXiv cs.AI 3d ago arXiv:2607.13188v1 paper proposes self-correcting coupled Markov jump processes for concurrent image understanding and generation; MDMs used arXiv cs.LG 3d ago arXiv:2607.13395v1: Enlightenment-style post-tuning without training for large models inspired by human 'enlightenment' concept arXiv cs.LG 3d ago CayleyR solves TopSpin puzzle by detecting cycle intersections in Cayley graphs arXiv cs.AI 3d ago Cost-optimal foundation model deployment portfolio for transportation management arXiv cs.AI 3d ago Deconstructing Actor-Critic: a large-scale empirical study of design components for practitioners arXiv cs.LG 3d ago Discourse-aware policy analysis with argumentation: a hybrid LLM-symbolic framework for disaster governance arXiv cs.CL 3d ago Disentangling knowledge states with ability and proficiency modeling for knowledge tracing arXiv cs.LG 3d ago EMAGN uses learned clustering for scalable traffic forecasting. arXiv cs.LG 3d ago Evaluation ability does not imply optimization utility; LLM-as-a-judge signals in closed-loop table recognition show weak judge signals on FinTabNet and OmniDocBench arXiv cs.CL 3d ago FixItFlow automates troubleshooting guide generation from historical cloud incidents using large language models arXiv cs.CL 3d ago GIFT teaches vision-language generative AI models to produce CAD programs for simulating and testing 3D objects. MIT News (AI) 3d ago Gsm-Plus-Bn Introduces Perturbation-Based Benchmark for Bangla Mathematical Reasoning in LLMs; Addresses Lack of Evaluation in Bengali arXiv cs.CL 3d ago LAPO uses leave-one-turn attribution for multi-turn search reasoning as self-generated process supervision. arXiv cs.AI 3d ago LLM-powered agentic system enables automatic discovery of ordinary differential equations in biological systems arXiv cs.AI 3d ago Muon emerges as strong optimizer for deep learning, outperforming Adam and AdamW; theoretical work interprets it as steepest descent under spectral norm. arXiv cs.LG 3d ago MyAG is a graph-based framework for designing and analyzing composable LLM agent systems with three graph abstractions. arXiv cs.CL 3d ago Networked intelligence: active shared context graphs for human-AI team science arXiv cs.AI 3d ago New framework identifies discarded information in ML models with Lie group inputs; arXiv paper details symmetry and null fibers analysis. arXiv cs.LG 3d ago New paper explores federated learning and explainable AI integration for privacy-preserving transparent machine learning arXiv cs.LG 3d ago New paper introduces benchmark and corpus-adaptive embeddings for SQL schema retrieval; arXiv:2607.13311v1 arXiv cs.CL

← 1 2 3 4 5 →