overfeed.news
Anuncio

NexTalk

Automatiza la primera respuesta. Mantén el toque humano.

Chat en vivo con agentes de IA en tu modelo y tu clave de API, con traspaso a tu equipo.

Empezar gratis

Tema

Seguridad y regulación

2441documentos

177últimos 7 días

arXiv AI Papers

Auditable Hierarchical Inference with Multimodal Evidence and Bayesian Active Inquiry

Most multimodal classifiers assume that all useful evidence is available upfront, whereas clinical inquiry is sequential. We present a synthetic evaluation framework separating objective multimodal inference from active inquiry. An objective model maps synthetic tongue, bilateral multipressure pulse, and acoustic observations to eight syndrome families. A provenance-bearing Knowledge Core and training-only answer model support Bayesian narrowing over 33 fine terms using 25 typed questions, while seven safety questions remain outside the diagnostic posterior. Weak shared symptom factors limit fine-term influence on objective phenotypes, and an objective-only probe tests leakage. In a preregistered 600-case synthetic evaluation, objective family accuracy was 87.0%. Cost-sensitive expected information gain (cost-EIG) raised family accuracy to 95.5% and exact fine-term accuracy to 42.5%, versus 92.7% and 33.0% for an unoptimized fixed-order baseline. Selective family accuracy reached 96.9% at 95.3% coverage. However, cost-EIG asked more questions on average than fixed order (9.795 versus 9.352). Missing-audio routing retained 94.0% family accuracy, and a workflow invariant enforced abstention when all objective modalities were absent. Permuting fine-term answer likelihoods reduced family accuracy to 78.7%. Five of six frozen checks passed, but the question-efficiency hypothesis failed. Comparisons are descriptive because only aggregate, not paired case-level, outputs were retained. These results establish a reproducible synthetic workflow, not clinical diagnostic or treatment validity.

en
arXiv AI Papers

Adversarial Self-Distillation

Even though current safety-aligned language models are able to refuse harmful requests, their guard can still be bypassed by adversarial perturbations. As in other domains, adversarial training is considered a standard defense. However, due to the autoregressive nature of language generation, supervised finetuning on template refusals does not necessarily teach the model to refuse. Evidence of this is that the adversarially trained model still complies when the attacker injects a compliant prefill. We propose Adversarial Self-Distillation (ASD), which extends a model's own safe behavior to the space of adversarial perturbations. The teacher (the model itself given the clean prompt) can refuse the harmful request by default, or with the help of an additional instruction in its context. ASD distills this knowledge to the student (which is the same model under attack) at the prefixes that the student visits (due to on-policy sampling). These potentially include compliant openings. ASD achieves lower attack success than state-of-the-art defenses with little loss in multiple-choice accuracy, and it becomes able to recover a refusal even after a compliant opening.

en
arXiv AI Papers

The Quine Test: How Much of a Model's Self Survives Its Own Description?

Frontier language models express coherent preferences, but behavioral evidence alone cannot say whether these belong to the model or to a character it portrays. Following Hofstadter's view of the self as a self-description, we ask whether a model's description of itself works as a portable specification of its preference profile, a quine test: can the text a model writes about itself regenerate its behavior elsewhere? Ten models from nine labs write self-descriptions; we place each one in the prompt of other models (receivers) and of fresh instances of the author, and measure how closely their choices on a 110-item preference battery match the author's. First, the tested native self-descriptions (native: the profile measured under a generic assistant prompt) show no average reconstruction gain across variations in length, framing, wording and prompt placement: on a 0 to 1 fidelity scale where a model's fidelity to a fresh instance of itself is 0.979 and to unrelated models about 0.89, the one-sided 95% upper bound on the pooled gain is below +0.01 in every tested system-prompt condition, and descriptions based on the author's observed choices do no better. Second, the channel itself is strong: explicitly listing the author's answers lets receivers reproduce its profile at fidelity 0.955 to 0.960 across three model grids, within 0.03 of same-model repeatability. Third, self-descriptions written under an explicit preference policy transfer the assigned preferences almost as well as the policy text itself. Fourth, reading its own self-description shifts even the author away from its baseline, modestly more than unrelated prose of the same length. For the preferences measured here, native self-descriptions influence behavior without reliably specifying how to reproduce it in another model. We release the pipeline, battery, raw call logs, and 198 model-written descriptions.

en
arXiv AI Papers

Latent Decision Flow Rewards for Clinical Guideline Reasoning

Clinical practice guidelines encode evidence-based recommendations as conditional decision flows, which makes them a natural source of process supervision for training reasoning models. Two properties of clinical reasoning make this supervision difficult to use. A decision flow consistent with a guideline does not by itself determine the correct answer for a given patient, which we call the path--answer gap, and several valid decision flows can support the same clinical decision, which we call path multiplicity. Current training approaches address parts of this problem. Answer rewards supervise the final decision, and process rewards computed against a reference flow supervise one valid decision flow per case. Supervised fine-tuning on annotated flows teaches guideline-consistent reasoning but constrains the model to one trajectory per case and can reduce answer accuracy. We introduce GUIDER, a method that jointly optimizes a decision flow controller and a reasoning policy. The controller learns to generate plausible decision flows from guideline-derived supervision, and the policy learns to generate a decision flow and a final answer for each patient case. GUIDER computes the process reward for the policy over several controller-generated flows, which accounts for path multiplicity, and gates this reward on answer correctness, which addresses the path--answer gap. Joint training allows the process supervision to evolve with the policy. On MedGUIDE, GUIDER with Qwen3-14B achieves 82.5 weighted accuracy and 73.5 guideline adherence, 3.3 and 3.1 points above the strongest clinical reinforcement learning baseline. With Qwen3-14B and Gemma4-12B backbones, GUIDER also improves over CRPO, MRPO, and FaithMed on the out-of-domain benchmarks AMEGA-LLM, RealMedQA, and CancerGUIDE.

en
arXiv AI Papers

Beyond fabrication: Misattributed numbers in papers written by AI research agents

Verifying numerical claims in research written by AI agents is an important but challenging problem. Recent benchmarks show that many agent-written papers fail integrity checks, but such verdicts, often produced by a single LLM judge, identify failure without explaining its nature, and existing checks either test a paper for internal consistency or test whether a reported value appears somewhere in an agent's output. We re-analyze all 540 released judgments from a recent public benchmark of AI research agents (nine agents, fifteen open research problems) and manually label 256 numerical claims from 18 submissions. Failures concentrate on alignment between the paper and the underlying work rather than overt cheating, and about four in five unsupported claims correspond to values in the agent's own files that come from the wrong experiment, metric, or run. We introduce a deterministic claim-to-evidence checker that scores whether the surrounding context identifies the claimed result and calibrates each claim against numerical decoys. With its rules fixed before labeling, it certifies claims with over 90% precision against blinded annotators (78 of 85; Cohen's kappa = 0.89), while verifying only a subset of claims; most claims it cannot verify are nevertheless traceable, so failure to verify is not evidence of falsehood. An LLM verifier given the same evidence recovers more traceable claims but accepts about three times as many misattributed ones, so the two are best used together. Applied to 99 submissions, the checker reveals large differences in traceability across agents, and a second judge from a different vendor preserves the official agent ranking while passing twice as many papers. These findings motivate provenance-first auditing of AI-produced research.

en
arXiv AI Papers

Deogen: Automatic Benchmark Generation for Reasoning over Policies

Language models are increasingly being deployed in regulated settings, where a model must apply given rules to the facts of a situation to determine a verdict. Benchmarking this capability, known as deontic reasoning, remains a costly and manual process, with existing datasets largely hand-crafted. In this paper, we present DeoGen, a scalable, automatic, and parametrizable benchmark generator for deontic reasoning. By autoformalizing policy rules to first-order logic, we can construct decision trees that capture possible outcome trajectories. Each path down a tree becomes a benchmark entry, with each tree producing combinatorially many entries relative to tree depth. Paths ending at a leaf entail a verdict, while non-leaf paths measure a model's ability to abstain when presented with insufficient information. Evaluations show frontier models return the correct verdict 89-92% of the time. Notably, performance degrades (22-40% correct) on entries without sufficient information to claim a verdict, even with explicit instructions to abstain in such cases. In all, DeoGen presents an automatic, scalable, and parametric method for generating deontic reasoning benchmarks.

en
arXiv AI Papers

Weaken to Strengthen: Weakening-Induced On-Policy Self-Distillation for Large Reasoning Models

Self-distillation offers a way to improve reasoning models when a suitable external teacher is unavailable. However, common approaches rely on privileged information to construct a stronger self-teacher, limiting their use when only unlabeled prompts are available. In this paper, we introduce Weakening-Induced On-Policy Self-Distillation (WISD), which instead constructs a weaker counterpart of the original model and uses the contrast between the two as a training signal without requiring privileged information. We explore two ways to construct the weaker model: low-bit quantization and magnitude pruning. We evaluate WISD across diverse architectures and scales on mathematical and coding reasoning tasks. Across these benchmarks, quantization and pruning improve average accuracy over the original models by 4.0 and 5.5 percentage points, respectively, achieving state-of-the-art performance even without privileged information. Notably, pruning raises Qwen3.5-4B's accuracy from 57.7% to 70.5% on HMMT 2025 and from 39.1% to 56.7% on LiveCodeBench v6. Quantization raises Qwen3.6-35B-A3B's accuracy from 68.5% to 74.6% and from 61.8% to 69.5% on the same benchmarks, respectively. These results establish WISD as a novel and effective self-distillation framework for large reasoning models.

en
AWS Machine Learning Blog

Building an Agentic AI Platform on Amazon Bedrock AgentCore

Stand up a governed, multi-account Agentic AI Platform on Amazon Bedrock AgentCore. This post walks through the architecture and an open-source reference blueprint: the operating model, control plane, per-Workstream cells, separate inference and tool gateways, and a CI/CD evaluation pipeline that gates each agent before production.

en
arXiv AI Papers

Information-Theoretic Reliability--Safety Bounds and Finite-Alignment Dynamics in AI Knowledge Distillation

The reliability and safety of artificial intelligence (AI), together with their transfer and evolution through knowledge distillation, have attracted growing attention. We establish an information-theoretic bound on the reliability--safety trade-off and characterize the dynamics of this trade-off under knowledge distillation. We first prove a universal trade-off inequality that is independent of the specific model form. Its boundary is determined solely by a single parameter Kh, which quantifies the hidden layer's ability to distinguish benign from hazardous inputs. We then develop a canonical Boltzmann-machine model of AI response behavior and derive two exact relations between reliability and safety. The trade-off curve is determined by the hazard discrimination capacity K, while the overall answer tendency controls the operating point along the curve. These relations agree well with existing experimental data. In knowledge distillation, we find that teacher responses induce a renormalization of the student model's trade-off parameters. We further use a renormalization-group approach to study multigenerational distillation, deriving exact recursion relations and the associated renormalization-group-like flows for two representative direct intergenerational distillation protocols at arbitrary teacher--student alignment strength. These flows reveal fixed-point structures, long-term dynamics, and critical behavior, yielding testable predictions for the critical transition between capability loss and stable transmission.

en
arXiv AI Papers

On-Policy Self-Distillation with Teacher Evolution

On-Policy Self-Distillation (OPSD), which aligns a student model's token-level distribution with that of its self-teacher prompted by privileged context, has emerged as a promising direction to mitigate reward sparsity in sequence-level post-training paradigms. However, OPSD often fails to achieve continuous performance gains during training. In this paper, we find that the model's limited self-teaching capability underlies this phenomenon by generating unreliable guidance on certain samples. To address this problem, we propose Teacher-Evolution Self-Distillation (TESD), which enables the self-teacher to generate rollouts and receive environmental feedback based on task correctness and context independence. First, using these rewards as training objectives, we perform teacher-oriented optimization to continuously enhance the model's self-teaching capability during distillation. Second, by estimating teacher guidance reliability from these rewards, we apply adaptive student-oriented optimization to down-weight samples with unreliable guidance. Extensive experiments across multiple datasets and model scales show that TESD consistently outperforms representative token-level and sequence-level methods, highlighting the necessity of improving supervision quality in OPSD.

en
arXiv AI Papers

Curriculum-Based Deep Reinforcement Learning for Robust Autonomous Rocket Landing under Atmospheric Uncertainty

Autonomous far-range rocket landing is a challenging control problem because navigation, descent-rate regulation, attitude management, and fuel use must be coordinated in the presence of uncertain atmospheric and vehicle dynamics. This paper presents a curriculum-driven proximal policy optimization (PPO) framework for three-dimensional rocket landing in a stochastic environment incorporating altitude-dependent aerodynamic drag, wind shear, temporally correlated turbulence, discrete gusts, uncertain propulsion and aerodynamic parameters, thrust-vector misalignment, engine lag, actuator-gain variation, noisy observations, and command latency. Learning is organised into far-range approach, soft descent, and integrated landing stages. Policy transfer, critic warm-up after reward changes, rehearsal of earlier tasks, and progressive expansion of weather severity are used to support learning across these coupled phases. Across 100,000 evaluation episodes, the policy achieves 96,656 successful landings, corresponding to a success rate of 96.656\% and a 95\% Wilson confidence interval of 96.543--96.766\%. Success decreases from 98.84\% under zero-disturbance conditions to 88.07\% in the highest weather-severity band. Successful touchdowns exhibit small miss distances and low residual horizontal and vertical velocities. An exploratory logistic-regression analysis identifies mean wind speed as the strongest measured correlate of success, with initial distance, initial altitude, initial vertical velocity, and wind-shear exponent also associated with outcomes. The results quantify both the effectiveness of the learned policy over the evaluated disturbance distribution and the degradation that occurs under severe compound atmospheric conditions.

en
arXiv AI Papers

Think Before You Act: Decision Gate for LLM Agents

LLM agents increasingly perform consequential actions through external tools, yet their proposed actions are often executed without independent verification of policy compliance or task completion. We introduce JevAct, a runtime verification framework that intercepts selected agent actions before execution and enables feedback-driven correction. JevAct compiles domain policies into executable state checks and structured questions evaluated by a prefill-only decision model, which produces probability-based judgments without relying on LLM generation. By combining deterministic verification of tool-state facts with probabilistic assessment of conversational context, the framework identifies potentially invalid actions and premature task termination, providing targeted feedback for bounded revision. JevAct operates independently of the underlying agent model and requires no modifications to its parameters or tools. We evaluate JevAct across eleven LLM agents on τ^2-bench Airline and Telecom tasks and BFCL. Results show improvements in agent reliability across multiple models and benchmarks, with particularly consistent gains on Telecom and BFCL, while performance on Airline is more mixed. Overall, these findings suggest that combining structured decision models with executable verification can improve the reliability of tool-using LLM agents.

en
arXiv AI Papers

How Deterministic SAST and AI Security Agents Divide the Work: Coverage, Harness Policy and Cost in Snyk VulnBench 2.0

We ask what AI coding agents and AI security harnesses find that deterministic static application security testing (SAST) does not, and what they miss that SAST catches. VulnBench 2.0 is a benchmark of 108 hand-curated, attacker-reachable vulnerabilities in 20 open-source web applications in Go, JavaScript/TypeScript, Python and Java; 62 flow across more than one file. A finding is credited only if it locates the vulnerability's attacker-controlled source and its dangerous sink, each within two lines. We evaluate nine productized configurations under one blind protocol: the Snyk Code static analyzer, Claude Code's /security-review with three model and effort settings (one with the Snyk MCP server), Codex Security with GPT-5.6 Luna, Terra and Sol, and the DeepSec scanner with two model backends. Snyk Code reached 82.4% macro recall (89 of 108) and Claude Code with Opus 5 at xhigh effort 79.3% (85 of 108); they disagree on 28 vulnerabilities, split 16 to 12, too few to tell them apart. Neither engine was sufficient alone: together they reached 101 of 108, the ceiling of all seven comparable configurations combined, and the AI configurations also reported at least seven real vulnerabilities that the reference set did not contain and Snyk Code did not report. What an AI engine misses depends on its harness as much as on its model: without Claude Code's built-in review policy, the same model matched 16 instead of 9 of the 20 vulnerabilities in the classes that policy excludes or restricts (supplementary run on 19 of 20 applications); Codex Security with GPT-5.6 Sol located 80 vulnerabilities at file level but 34 at line level; DeepSec's model-free file selection kept a file from only 56 of the 108 flows. Coverage and cost were only loosely related. The harness, fixtures, reference sets and all run data are public.

en
arXiv AI Papers

Repairs for Whom? Evaluator-Executor Mismatch in Selecting Repairs for Failed LLM Agents

When an LLM agent fails, repair methods choose which action to replace and what to put in its place. They often score candidate repairs with greedy replay, a stronger model, an environment oracle, or a judge that reads the failed log. We ask whether these evaluators pick repairs that help the agent that will actually execute them. Holding failed trajectories and candidate single-step repairs fixed, we label every candidate with independent rollouts of the deployed policy and measure how much deployment value each evaluator's choice loses. On failures not screened for repair value, pre-specified tests show that greedy replay loses 0.08-0.13 in expected net success gain in ALFWorld and 0.06-0.13 in net reward in WebShop, relative to selection with the deployed agent's own rollouts. At an equal rollout budget, selecting repairs for Qwen2.5-7B with Qwen2.5-14B rollouts loses 0.25 in ALFWorld and 0.07 in WebShop, twenty 14B rollouts per candidate are worth no more than two of the 7B agent's own, and fresh rollouts confirm that the two models prefer opposite repairs on selected disagreement cases. Between similar-sized models from different families, the losses appear in both directions in both environments (0.07-0.17). The scoring executor also changes repair-location labels and method rankings. Repairs should be verified with the executor that will run them, and repair labels should name their executor.

en
arXiv AI Papers

Transportable Offline Reinforcement Learning in Presence of Unmeasured Confounding

Offline reinforcement learning seeks to learn decision policies from previously collected data, but deployment environments may differ from the domains in which the data were collected, and logged variables may omit confounders that affected both historical actions and outcomes. We study offline policy learning under the joint challenges of domain discrepancy and unmeasured confounding, where multiple heterogeneous source datasets provide complementary information about a target domain. We develop a causally adjusted transportability framework combining transportability theory with instrumental variables, negative controls, and proxies to identify target-domain causal quantities. For confounded contextual bandits, identification reduces to systems of bridge-function integral equations, enabling minimax estimation from finite samples. Confidence sets for the identified quantities are then used for pessimistic policy optimization, yielding CAT-CCB. We extend the framework to episodic confounded Markov decision processes, where complementary sources identify target reward and transition mechanisms under structural and completeness conditions. Under Gaussian dynamics and linear function approximation, this yields CAT-CMDP. We establish high-probability finite-sample suboptimality guarantees that quantify the effects of approximation, estimation, source-to-target shift, offline coverage, and sample size. For bounded finite-dimensional linear classes, the contextual-bandit guarantee recovers the standard {O}(N^{-1/2}) rate. Numerical experiments support the effectiveness of the framework for learning target-domain policies from heterogeneous confounded data. Together, the results provide a principled approach when no single source dataset suffices for causal identification and policy optimization.

en
arXiv AI Papers

Keepace.ai: A Technical Report on a Large Language Model for Sports and Health

Large language models hold promise for delivering safe, scientifically grounded, and personalized exercise guidance at scale, yet the sports-and-health domain still lacks curated training corpora and rigorous evaluation benchmarks. Keepace.ai is a domain-specialized LLM built on Qwen3.5-27B through post-training, currently optimized for Chinese-language exercise guidance across three tasks: workout plan generation, sports-and-health knowledge QA, and personal sports data analysis. In internal rubric-based evaluations designed with domain experts, Keepace.ai outperforms its base model and larger general-purpose baselines, with the most pronounced improvements in exercise safety, boundary compliance, structured output completeness, and scientific accuracy. This report summarizes the current design, data pipeline, safety-first evaluation framework, and evaluation results of Keepace.ai.

en
arXiv AI Papers

Variational Consequence-Driven Offline Reinforcement Learning

Offline reinforcement learning (RL) typically mitigates distribution shift by imposing divergence constraints on the action distributions of the learned and behavior policies. However, this paradigm cannot ensure the underlying dynamic consistency structure: even minor action deviations can lead to drastically different transitions, inadvertently driving the agent away from the offline dataset. To address this, we explore policy constraints through their induced state-consequence distributions rather than relying on pure action-distribution matching. Nevertheless, directly enforcing such consequence-aware constraints is computationally intractable without access to the true underlying dynamics model. To bypass this issue, we propose Structural Behavior Regularization (StBR), which replaces the intractable objective with a closed-form latent divergence. Structuring this latent space to reflect transition dynamics renders the latent divergence a theoretically bounded surrogate for the true consequence-distribution divergence. By regularizing the learned policy with this latent divergence, StBR adaptively tightens the behavior constraint in dynamics-sensitive regions, effectively preventing deviations outside the offline dataset. Empirically, we demonstrate that StBR achieves superior performance across diverse offline RL benchmarks.

en