Sat 31 Oct
LiveGet it free
ZOTPAPER

News without the noise

Research

New AI papers from arXiv, picked daily and explained with their claims, benchmark results and code.

2566 articles · page 45 of 54 · The Research front page

AllAgentsLanguage modelsReinforcement learningVisionGenerative mediaSpeech & audioRobotics & embodied AISafety & securityEvaluation & benchmarksEfficiency & systemsCode & softwareAI for science & healthTheory
Viewgridlistcompact
Reinforcement learning

Empathetic AI learns to adapt support priorities as conversations evolve

The authors identify a fundamental mismatch in empathetic reinforcement learning: reward specifications remain fixed while support needs change. They propose CARE (Context-Adaptive Rubric Evolution), which generates turn-level rubrics that adjust the weights and criteria of three empathy dimensions. On the EMPA benchmark, CARE improves the EPM-Idx metric from 28.11 to 83.54 under Gemini-2.5-Pro, a gain of over 55 points, and consistently outperforms baselines across three independent judges.

Agents

Terminal agents struggle when familiar environmental assumptions stop holding

Singh and colleagues introduce AGNI, a pipeline that identifies assumptions behind successful agent trajectories, invalidates them through controlled environmental changes, and verifies that the modified tasks remain solvable. Multiple LLM agents performed substantially worse on these novel variants, often observing evidence of a change without correctly diagnosing it. Post-training on environmental novelty improved performance on both held-out novel tasks and unchanged base tasks.

Agents

Draft-KV ensures the receiver actually uses the sender's latent message

The authors identify a fundamental flaw in prior latent communication methods: replacing the sent message with one from an unrelated question barely changes accuracy, meaning the receiver relies on the interface, not the message content. They propose Draft-KV, which transmits key-value states from the sharer's own answer draft via a gated attention branch with progressive training. A frozen 8B sharer lifts a frozen 0.5B receiver from 37.45% to 78.04% on MMLU-Redux, and message reassignment drops accuracy to 36.40%, confirming the gain stems from genuine latent information exchange.

Safety & security

Selective distillation improves safety alignment while halving rollout compute and preserving reasoning.

The authors identify two inefficiencies in on-policy self-distilled safety alignment: teacher supervision collapses on long student rollouts, and stylistic shifts from privileged prompts dilute genuine safety gradients. They propose EOPSA, which dynamically bounds rollout length via a Teacher Rescue Rate metric and selectively distills only safety-pivotal tokens, cutting rollout computation by roughly half and backpropagating through only about 2% of tokens while improving both safety and reasoning retention.

Language models

Structured reasoning framework improves LLM game-playing efficiency and payoff

The authors propose SAGE, a training-free inference-time framework that structures LLM strategic reasoning around three coordinated operations: anchor, adapt, and recalibrate. In repeated imperfect-information games (Leduc Hold'em, Liar's Dice, Goofspiel), SAGE outperforms reasoning-intensive LLM agents, achieving up to a 127.6% payoff improvement in Liar's Dice while reducing input and output token usage by up to 80% and 90% respectively.

Agents

Agent distillation improves when supervision follows outcomes, not model disagreement

Zhang et al. test a common assumption in on-policy distillation: that large token-level differences between teacher and student mark the most useful moments for supervision. They show that these local gaps often mispredict downstream benefit, then introduce OG-OPD, which weights teacher guidance using final outcomes from paired student continuations and improves success across three interactive benchmarks.

Safety & security

Jailbreak prompts improve uncertainty calibration in black-box reasoning models

Biechy et al. investigate uncertainty quantification (UQ) for large reasoning models (LRMs) in black-box settings, where logits are unavailable. They find that standard UQ methods such as paraphrase-based self-consistency and confidence verbalization offer little benefit over repeated sampling, and introduce J4U, a jailbreak-derived technique that empirically improves calibration across multiple models and datasets.

Efficiency & systems

Adaptive prefill-decode execution and elastic GPU reuse boost agentic RL throughput

The authors introduce PEARL, a system that coordinates external elastic GPU resources, temporary reuse of idle training GPUs, and adaptive prefill-decode (PD) configuration for asynchronous agentic reinforcement learning. In experiments with Qwen3-8B and Qwen3-30B-A3B models, PEARL achieves 2.17 to 2.79 times the throughput of a fixed-resource baseline (ROLL) and up to 26.9% and 36.3% improvement over RLBoost+ for the respective models.

Evaluation & benchmarks

LLMs often diagnose without correctly valuing available medical evidence

Feng et al. tested nine language models on MedEVM, where medical evidence arrives sequentially and models must choose whether to wait or diagnose. The authors find that models often submit diagnoses at the wrong evidential moment, remain sensitive to evidence order and misleading details, and gain 12.0 to 51.1 percentage points in accuracy when an evidence-verification harness controls submission.

Robotics & embodied AI

Robot-relative 3D features improve pretrained manipulation policies across diverse tasks

Liu et al. introduce Spatial Grafting, which converts features from frozen 3D reconstruction models into robot-relative spatial tokens and feeds them to flow-matching action generators through cross-attention. Across four simulation benchmarks and three physical robot platforms, the authors report consistent gains, with the largest improvements appearing in cluttered and long-horizon manipulation.

Reinforcement learning

Twin critics calibrate token-level advantages from a single rollout, boosting RL training efficiency

Qiao et al. introduce a twin-critic method for token-level credit assignment in reinforcement mid-training of language models. By training two critics and combining their advantage estimates with action-dependent weights learned via a conditional-moment objective, they reduce update drift and eliminate the need for repeated generation. Experiments show improved performance and substantially faster training.

Agents

Production-scale AutoResearch exposes five failure modes and a three-principle fix

The authors applied AutoResearch, an LLM that iteratively edits training scripts and keeps changes that improve a held-out metric, to two production embedding systems at Amazon. Over 220 experiments spanning 12 weeks, they identified five recurring failure modes and a three-principle multi-agent scaffolding design that maps each to a structural remedy. The framework improved Recall@6 by 1.82x, coherence by 2.1x, and expanded catalog coverage 5.8x.

Agents

A method that links failure localization to repair by identifying which editable system asset to change

The authors propose DAAF, a framework that moves beyond localizing where a failure occurs in an LLM agent to determining which persistent, editable system asset (e.g., a routing rule or knowledge segment) should be changed. By learning the effects of valid attribute replacements from controlled replays and amortizing this evidence at deployment time, DAAF achieves 80.72% attribute Hit@1 on held-out Telecom tasks while recovering 62.65% of failed executions and limiting clean-task regression to 3.23%.