论文广场 - AcademicHub

01.

arXiv (CS.AI) 2026-06-12 DOI: arXiv:2606.04935

What Type of Inference is Active Inference?

作者:

Wouter W. L. Nuijten ↗Mykola Lukashchuk ↗Thijs van de Laar ↗Bert de Vries ↗

arXiv:2606.04935v2 Announce Type: replace Abstract: Active inference casts decision-making as inference, with the Expected Free Energy (EFE) unifying goal-directed and information-seeking behavior. Recent work showed that EFE minimization can be written as Variational Free Energy (VFE) minimization on a generative model augmented with epistemic priors. We prove that the VFE of the augmented model can be rewritten as the VFE of the predictive model plus explicit entropy-correction terms, making the EFE contribution transparent. We then show that proper EFE-based planning requires combining these epistemic corrections with a planning correction that turns marginal inference into policy optimization, yielding a full variational characterization of EFE-based planning. This clarifies which corrections are needed for cross-entropy planning and for full EFE-based planning. The same entropy-corrected formulation leads to a detailed message-passing scheme for EFE-based planning together with simpler ablations. Experiments on three grid-world environments show that full EFE-based planning outperforms ablations that omit either the planning correction or the epistemic corrections.

阅读与讨论 → 访问原文 →

02.

medRxiv (Medicine) 2026-06-15 DOI: HASH:a1bc08bd0359f3dc44d745c7db841012

ICD-10 Code Ambiguity Obscures Treatment-Eligible Adults with Spinal Muscular Atrophy: A Single-Center Chart Review and Patient Outreach Study

作者:

Holly ↗Bean ↗Beshay ↗Edwards ↗Streicher ↗N. S ↗

Background. Three disease-modifying therapies (DMTs) for spinal muscular atrophy (SMA) have been approved since 2016, yet many adults remain untreated. Identifying them depends on ICD-10 codes that capture SMA but do not reliably distinguish it from other related conditions. We examined, in one U.S. health system, both patients' engagement with therapy and the accuracy of the codes used to find them. Methods. We conducted a retrospective chart review of adults in an academic health system identified by SMA-associated ICD-10 codes, with manual adjudication of diagnosis and DMT status. Confirmed SMA-positive, DMT-naive patients were invited to a structured telephone interview on treatment awareness and barriers. Results. Of 60 charts, 22 (36.7%; 95% CI 25.6-49.3%) were appropriately coded for SMA or a related disorder; only 16 (26.7%) had molecularly confirmed SMA. The other 38 (63.3%) were miscoded, spanning spinal and bulbar muscular atrophy, asymptomatic carriers, prenatal screening, and conditions unrelated to SMA. Ten of the 16 confirmed patients (62.5%) were DMT-naive; one was interviewed, one declined, and eight could not be reached. The non-response is itself a finding: the patients least visible to administrative data are the hardest to reach. Conclusions. ICD-10 ambiguity is a barrier to treatment access in adult SMA, as is loss to follow-up. We make two recommendations: continuous documentation-coding alignment that uses natural language processing to verify the genetic precondition, and type-specific SMA codes (subcodes for Types 0-4) anchored on molecular SMN1 confirmation. Together these would support cohort identification, outreach, and evidence generation without adding to clinician burden.

阅读与讨论 → 访问原文 →

03.

arXiv (CS.LG) 2026-06-11 DOI: arXiv:2606.11583

Beyond the Golden Teacher: Enhancing Graph Learning through LLM-GNN Co-teaching

作者:

Zhuoyi Peng ↗Hanlin Gu ↗Lixin Fan ↗Yi Yang ↗

arXiv:2606.11583v1 Announce Type: new Abstract: Text-attributed graphs (TAGs) underlie real-world applications such as citation networks, social media, and e-commerce. Few-shot graph learning on TAGs is hard: with only a handful of labels per class and the rest of the graph unannotated, neither GNNs nor LLMs can learn well on their own. GNNs read topology and fail on cold nodes; LLMs read text and fail on text-ambiguous nodes. Existing LLM-GNN methods all follow the same recipe: designate one model as the golden teacher and use its outputs (e.g., features or pseudo-labels) to supervise the other. We argue this golden-teacher assumption breaks under sparse supervision: neither model is golden, and treating either as such transfers its blind spots into the student. We therefore ask: can we avoid designating either model as the golden teacher, and still perform effective graph learning? We answer with LLM-GNN Co-Teaching, a bidirectional co-teaching framework in which neither model is fixed as teacher. The GNN and LLM exchange their most confident pseudo-labels under an architecture-specific small-loss criterion, and both update every round. Supervision is then mined from the trajectory: whenever a node moves from cross-model contradiction at round t to cross-model agreement at round t+1, the LLM's two answers on the same input form a preference pair (old contradicting self < new peer-endorsed self) for DPO training. We call this Round-based Pseudo-Label Preference Optimization (RPL-PO). On six benchmarks, LLM-GNN Co-Teaching consistently outperforms GNN-as-Judge and all prior methods, with absolute 3-shot gains of 7.86% on Cora and 7.73% on ogbn-arxiv; improvements carry over to 5-shot and to zero-shot cross-dataset transfer. Error-structure analysis further shows that abandoning the golden-teacher assumption substantially improves the LLM's graph learning capability on challenging samples.

阅读与讨论 → 访问原文 →

04.

arXiv (CS.CL) 2026-06-19 DOI: arXiv:2606.19625

Where Does Social Reasoning Come From? Capability Provenance in Language Models

作者:

Glenn Matlin ↗Chandreyi Chakraborty ↗Saehee Eom ↗Mika Okamoto ↗Rayan Castilla ↗Louis Jaburi ↗Alvin Deng ↗Taywon Min ↗Lucia Quirke ↗Stella Biderman ↗Mark Riedl ↗

We use training-data attribution as an interpretable tool for capability discovery, mapping which regions of the pretraining corpus support social-reasoning versus STEM-reasoning in OLMo3-7B. Training-data attribution measures how strongly each training document influences a model's predictions on a benchmark, but document-level scores are too noisy to identify which corpus regions support which capabilities, and prior work has emphasized factual knowledge rather than reasoning. We compute gradient-based attribution (TrackStar via Bergson) over a working set drawn from the de-duplicated Dolma3 mix, aggregate influence across WebOrganizer's 24-format x 24-topic taxonomy (576 bins), and contrast benchmark pairs in a 2x2 design that varies domain (social vs. STEM) and capability type (reasoning vs. knowledge): SocialIQA and MMLU Social Sciences against ARC-Challenge and MMLU STEM. Social and STEM reasoning draw on qualitatively distinct corpus regions, and the contrast is sharper at the reasoning level than at the knowledge level. Targeted machine unlearning provides partial causal validation: forgetting high-attribution topic bins (e.g., Literature for SocialIQA) degrades the aligned benchmark more than within-bin random baselines, and we open-source all code, sampling manifests, the bin-level influence matrix, and unlearning checkpoints.

阅读与讨论 → 访问原文 →

05.

arXiv (CS.CL) 2026-06-15 DOI: arXiv:2606.14695

Persona-Pruner: Sculpting Lightweight Models for Role-Playing

作者:

Jinsu Kim ↗Jihoon Tack ↗Noah Lee ↗Jongheon Jeong ↗

Language Models (LMs) have shown remarkable potential as role-playing chatbots, delivering consistent, stylized interactions when given a specification of a character or user persona. However, applying these capabilities to real-world applications (e.g., ecosystems with numerous NPCs interacting simultaneously) exposes a critical inefficiency due to the excessive computational cost. In this paper, we question the necessity of dedicating a full, generalist model to a single persona, hypothesizing that a specific character identity relies on only a fraction of the model's total capacity. We observe that naively pruning LMs often severely degrades the role-playing performance for a specific persona; it does not distinguish between redundant knowledge and essential character traits. We propose Persona-Pruner, a framework that sculpts a lightweight role-playing model by isolating persona-specific sub-networks from a single description. Our experiments consistently show that Persona-Pruner preserves role-playing performance substantially more effectively than existing state-of-the-art LLM pruning techniques, reducing the performance drop from the dense model by up to 93.8% over the strongest baseline on RoleBench in LLM-as-a-judge score, while still maintaining general LLM capabilities. Code is available at https://github.com/jsu-kim/Persona-Pruner.

阅读与讨论 → 访问原文 →

06.

arXiv (quant-ph) 2026-06-16 DOI: arXiv:2606.15769

Noise-Adaptive Predictive Dynamical Decoupling

作者:

Ali Abu-Nada ↗Subhashish Banerjee ↗Lian-Ao Wu ↗

arXiv:2606.15769v1 Announce Type: new Abstract: Protecting quantum coherence against realistic environmental noise remains one of the fundamental obstacles to scalable quantum technologies. We develop a noise-adaptive dynamical decoupling framework that combines analytical open-quantum-system modeling with machine-learning-based forecasting for a qubit interacting with random telegraph noise. Unlike conventional dynamical decoupling protocols based on fixed pulse schedules, the proposed approach continuously forecasts short-time coherence evolution and adaptively applies control pulses according to the instantaneous noise dynamics. We investigate stationary and non-stationary environments spanning both Markovian and non-Markovian regimes. Numerical simulations demonstrate that the machine-learning-assisted adaptive control strategy substantially outperforms conventional periodic dynamical decoupling while using a comparable number of control pulses. The improvement becomes particularly pronounced in non-Markovian and non-stationary regimes, where memory effects, coherence revivals, and temporally evolving noise strongly limit the effectiveness of static pulse protocols. These results establish predictive machine-learning-assisted dynamical decoupling as a promising and scalable framework for adaptive quantum control in realistic noisy quantum devices.

阅读与讨论 → 访问原文 →

07.

medRxiv (Medicine) 2026-06-24 DOI: HASH:caece6ee14cac61ca90541f7030df3e6

Genetically Proxied Interleukin-6 Inhibition and Cancer Risk: A Multi-Ancestry Drug-Target Mendelian Randomization Study of Hepatocellular Carcinoma and Colorectal Cancer

作者:

Ranjha ↗M. M. A ↗Munir ↗

Background: Interleukin-6 (IL-6) signalling drives chronic inflammation and is therapeutically targeted by tocilizumab, an approved IL-6 receptor inhibitor. Whether genetically proxied lifelong IL-6 inhibition causally influences the risk of hepatocellular carcinoma (HCC) or colorectal cancer (CRC) remains unanswered. Prior single-variant estimates from pooled observational data are methodologically limited and may reflect confounding. Methods: A two-sample drug-target Mendelian randomization (MR) study was conducted. Four independent cis-acting protein quantitative trait loci (pQTL) variants within the IL6 and IL6R gene loci (rs2228145, rs4129267, rs7529229, rs1800795) were selected as genetic instruments , with F-statistics ranging from 32.3 to 120.5, confirming instrument strength. Outcome data were obtained from four independent genome-wide association studies: HCC from BioBank Japan (BBJ; 1,866 cases, 195,745 controls), HCC from FinnGen Release 10 (674 cases, 218,118 controls), CRC from a European meta-analysis (19,948 cases, 12,124 controls), and CRC from BBJ (7,062 cases, 195,745 controls). Causal estimates were derived using inverse variance weighted (IVW) regression as the primary method, with MR-Egger and weighted median analyses as sensitivity methods. Cochran Q statistics assessed heterogeneity and MR-Egger intercept testing assessed directional pleiotropy. Results: Genetically proxied IL-6 inhibition showed no significant causal effect on HCC risk in East Asian populations (IVW odds ratio [OR] 0.997, 95% confidence interval [CI] 0.903 to 1.101, p=0.953) or European populations (IVW OR 0.984, 95% CI 0.802 to 1.208, p=0.880). Similarly, no causal effect was observed on CRC risk in European populations (IVW OR 1.015, 95% CI 0.957 to 1.075, p=0.623) or East Asian populations (IVW OR 0.999, 95% CI 0.948 to 1.052, p=0.971). Sensitivity analyses confirmed the absence of directional pleiotropy and heterogeneity across all four analyses. Leave-one-out analyses demonstrated that no single instrument drove the null findings. Conclusions: Genetically proxied IL-6 receptor inhibition, modelling the therapeutic effect of tocilizumab, showed no causal effect on HCC or CRC risk across four independent cohorts and two ancestries. These findings do not support a role for IL-6 pathway inhibition in the prevention of these cancers and provide reassuring genetic safety evidence regarding cancer risk in patients receiving tocilizumab. Larger HCC-specific GWAS are needed to definitively evaluate modest effects in this cancer type.

阅读与讨论 → 访问原文 →

08.

arXiv (CS.LG) 2026-06-11 DOI: arXiv:2601.08136

Reverse Flow Matching: A Unified Framework for Online Reinforcement Learning with Diffusion and Flow Policies

作者:

Zeyang Li ↗Sunbochen Tang ↗Navid Azizan ↗

arXiv:2601.08136v2 Announce Type: replace Abstract: Diffusion and flow policies are gaining prominence in online reinforcement learning (RL) due to their expressive power, yet training them efficiently remains a critical challenge. A fundamental difficulty that distinguishes online RL from standard generative modeling is the lack of direct samples from the target Boltzmann distribution defined by the Q-function. To address this, two seemingly distinct families of methods have been proposed for diffusion policies: a noise-expectation family, which uses a weighted average of noise as the training target, and a gradient-expectation family, which employs a weighted average of Q-function gradients. However, it remains unclear how these objectives are formally related, or whether they can be synthesized into a more general formulation. In this paper, we propose a unified framework, reverse flow matching (RFM), which rigorously addresses the problem of training diffusion and flow models without direct target samples. By adopting a reverse inferential perspective, we formulate the training target as a posterior mean estimation problem given an intermediate noisy sample. Crucially, we introduce Langevin Stein operators to construct zero-mean control variates, deriving a general class of estimators that share the same expectation. We show that existing noise-expectation and gradient-expectation methods are simply two specific instances within this broader class. This unified view yields two key advancements: it extends the capability of targeting Boltzmann distributions from diffusion to flow policies, and it enables the principled combination of Q-value and Q-gradient information to form an effective estimator, thereby improving training efficiency and stability. We instantiate RFM to train a flow policy in online RL and demonstrate improved performance on continuous-control benchmarks compared to diffusion policy baselines.

阅读与讨论 → 访问原文 →

09.

arXiv (math.PR) 2026-06-16 DOI: arXiv:2606.15203

The existence of invariant sublinear expectations for $G$-SDEs

作者:

Danyang Zhao ↗Huaizhong Zhao ↗

arXiv:2606.15203v1 Announce Type: new Abstract: In this paper, we study the existence of invariant sublinear expectations of Markovian semigroups on sublinear expectation spaces. To achieve this, we establish a complete metric space of sublinear expectations, on which we extend Harris' method to the nonlinear setting on the convergence of sublinear semigroups. We then explore two cases of $G-$diffusions by studying the Lyapunov function and the local Doeblin condition. One is the $G-$Brownian motion on the unit circle which is the case studied in Feng and Zhao [Zhaonon], but with the new method. Another is the multidimensional $G-$SDEs on the whole space $\mathbb{R}^d$. We establish, for the first time in the literature, the existence of the invariant sublinear expectation for $G-$SDEs under the non-degenerate and weakly dissipative assumption. For this, we prove that for a class of $G-$SDEs, the $G-$expectation can be represented as the supremum of the semigroup of a family of SDEs, of which the regularity is obtained by considering the Bismut-Elworthy-Li formula and the Denis-Hu-Peng representation for the distribution of $G-$Brownian motions.

阅读与讨论 → 访问原文 →

10.

arXiv (CS.CV) 2026-06-12 DOI: arXiv:2605.01733

GEASS: Gated Evidence-Adaptive Selective Caption Trust for Vision-Language Models

作者:

Zeshang Li ↗Shuoyang Zhang ↗

Vision-Language Models (VLMs) hallucinate objects that are not present, and a growing line of work tries to curb this by feeding the model its own generated caption as auxiliary evidence – assuming that a caption, once available, is something to consume. We show this fails: naively appending a caption can lower accuracy rather than raise it, dropping Qwen2.5-VL-3B$^\dagger$ on HallusionBench by nearly ten points. To understand why, we build GD-Probe, a diagnostic set that pairs a global and a detail question on the same image, so that any difference in caption effect is attributable to the question alone. Caption utility proves to be a per-query property: the same caption helps global questions and harms detail ones, through a single mechanism – an embedded caption competes with the image for attention and pulls the model's evidence onto its own text – whose sign is set by whether the caption covers the queried content. Crucially, this regime is readable from quantities the decoder already emits, with no attention access or grounding. We turn this into GEASS (Gated Evidence-Adaptive Selective Caption Trust), a training-free, logit-level module that decides per query how much of the caption to trust, gating it by the clean path's confidence, weighting it by the entropy reduction it induces, and raising the evidence bar when the two pathways disagree. Across four VLMs and two benchmarks (POPE and HallusionBench), GEASS improves over both vanilla inference and contrastive decoding under a single fixed setting, adding only two forward passes and no parameters.

阅读与讨论 → 访问原文 →

11.

arXiv (quant-ph) 2026-06-16 DOI: arXiv:2606.16784

Quantum Nonlocal Games on Graph Ensembles

作者:

arXiv:2606.16784v1 Announce Type: new Abstract: Quantum entanglement is one of the most striking discoveries in all of science. This effect allows, for instance, two spatially separated agents to coordinate their actions, without communication, to an extent that is both counter-intuitive, and provably impossible by any other physical means. A recently discovered example is that of mobile agents (players) performing spatial coordination tasks such as rendezvous, where the agents aim to meet on a network without communication. Until now, demonstrations of this advantage have relied on highly idealized conditions: agents are assumed to have complete knowledge of the topography, and experiments have been restricted to simulations using data generated by qubits within a single quantum processor. Here we address both limitations by developing a theory for graph ensembles that capture topographical uncertainty and by experimentally demonstrating the advantage in rendezvous scenarios between physically separated ion-trap systems with access to remote entanglement. Moreover, we simulate a broader set of problems on superconducting hardware. Surprisingly, when players are given the ability to gather more local information the quantum advantage increases – a feat impossible by classical means. Our findings establish a concrete route toward practical quantum advantages in motion coordination problems. More broadly, they point to a new way of using portable quantum devices to enhance collective decision-making in uncertain environments.

阅读与讨论 → 访问原文 →

12.

arXiv (CS.AI) 2026-06-18 DOI: arXiv:2606.19168

Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection

作者:

Jinhan Li ↗Kexian Tang ↗Yihan Xu ↗Zhuorui Ye ↗Kaifeng Lyu ↗

arXiv:2606.19168v1 Announce Type: new Abstract: To achieve deeper safety alignment for large language models (LLMs), recent efforts have studied how to push safety interventions earlier into the pretraining stage, primarily by filtering unsafe data or rewriting it into safer forms. We argue that pretraining-stage alignment should go beyond making the data safe: LLMs may compose seemingly benign knowledge and capabilities into unsafe behaviors. To this end, we propose Safety Reflection Pretraining, a pretraining-stage alignment method which regularly inserts short safety reflections into pretraining corpora to integrate self-monitoring directly into language modeling, establishing a foundational capability that is subsequently reinforced by compatible post-training. Our experiments with 1.7B models pretrained on FineWeb-Edu show that Safety Reflection Pretraining improves safety classification accuracy and substantially reduces the success rates of inference-stage and finetuning attacks. Complementary to our real-world experiments, we also introduce a fully controlled synthetic environment, MedSafetyWorld, with a clear definition of safety and a reasoning structure under which models can easily generalize unsafe behaviors from safe data. Ablations in MedSafetyWorld further demonstrate a clear advantage of Safety Reflection Pretraining in preventing models from acting on unsafe behaviors generalized from safe data, compared with data filtering and rewriting. Taken together, our findings suggest that pretraining alignment should not only make the training data safe, but also shape the behaviors that models are likely to acquire from safe data.

阅读与讨论 → 访问原文 →

13.

arXiv (CS.AI) 2026-06-19 DOI: arXiv:2606.19899

Measuring Biological Capabilities and Risks of AI Agents

作者:

Patricia Paskov ↗Jeffrey Lee ↗Kyle Brady ↗Alyssa Worland ↗

arXiv:2606.19899v1 Announce Type: cross Abstract: This paper addresses a rapidly emerging policy challenge: how to generate and interpret credible evidence about the biological capabilities and risks of AI scientists, or agentic AI systems capable of autonomously or collaboratively performing multi-step scientific tasks. As these systems enter real research workflows, decision-makers increasingly face evaluation results whose meaning depends on underlying design choices that are often implicit or under-documented. We synthesize current evidence on AI-enabled biological risks and introduce biological agentic evaluations as a promising, but interpretation-sensitive, tool for assessing these systems. Our central contribution is a set of practical, experience-grounded considerations – drawing from our own evaluations – that show how choices around defining, designing, running, scoring, and documenting evaluations materially shape what results do and do not imply about risk. The analysis is intended to help policymakers interpret biological evaluation outputs with appropriate caution; guide public and private funders toward high-leverage investments in AI-biology evaluation research; and support biosecurity practitioners assessing emerging AI systems. A secondary audience includes researchers designing or conducting agentic evaluations within frontier AI labs, AI providers, scientific institutions, and third-party evaluation organizations.

阅读与讨论 → 访问原文 →

14.

arXiv (CS.CV) 2026-06-16 DOI: arXiv:2606.14732

Steady-Forcing: Balancing Spatial Persistence and Motion Continuity in Long-Horizon Nature Video Diffusion

作者:

Matiur Rahman Minar ↗Seunghun Oh ↗GangHyeon Jeong ↗Unsang Park ↗

Autoregressive video diffusion models enable streaming generation but often degrade over long rollouts: static scene layouts drift, while mechanisms that improve spatial stability tend to suppress motion, causing natural flows such as water, fire, or smoke to stagnate. We study this stability-motion trade-off in fixed-camera long-horizon nature video generation, where the two failure modes can be more clearly separated than in moving-camera settings. We propose Steady-Forcing, a memory and training framework combining a persistent visual anchor (V-Sink), an exponential moving-average motion memory (EMA-Sink), block-relative temporal encoding, periodic cache purification, and distillation from a Wan2.1-14B teacher with motion-rewarded priors under task-focused configurations. Together, these components are designed to preserve background identity while sustaining visually plausible fluid dynamics over multi-minute autoregressive rollouts. Evaluations across seven baselines show that Steady-Forcing improves long horizon background consistency and imaging quality, while a blind user study indicates stronger perceived stability and motion continuity. The benchmark evaluation further suggest that generic VBench aggregate scores under-penalize fixed-camera artifacts as well as rewarding drift-induced optical flow as Dynamic Degree while not directly penalizing texture hardening or flow stagnation - motivating future task-specific benchmarks for static-camera nature-flow evaluation. Project page: https://minar09.github.io/steadyforcing/

阅读与讨论 → 访问原文 →

15.

arXiv (CS.LG) 2026-06-16 DOI: arXiv:2606.14909

Audited Conformal Prediction for Classification under Unknown Distribution Shift

作者:

Yanfei Zhou ↗Rizal Fathony ↗Nam H. Nguyen ↗Matteo Sesia ↗

arXiv:2606.14909v1 Announce Type: cross Abstract: We consider the problem of uncertainty quantification for a pretrained classification model deployed under unknown distribution shift. We propose Audited Conformal Prediction (ACP), a method that leverages a small labeled dataset from the target population to train an auxiliary audit model identifying inputs where the legacy model is likely to fail. By integrating the audit model's outputs into the conformal prediction framework, ACP produces prediction sets that guarantee marginal coverage while achieving substantially higher conditional coverage in practice than existing approaches. We develop and analyze two complementary integration strategies – one targeting marginal coverage with improved conditional performance, the other providing explicit group-conditional coverage guarantees – and establish theoretical guarantees for both. Experiments on synthetic and real-world datasets validate the method and illustrate trade-offs between prediction set size and conditional coverage.

阅读与讨论 → 访问原文 →

16.

arXiv (math.PR) 2026-06-16 DOI: arXiv:2606.15425

Well-posedness of stochastic parabolic equations with gradient nonlinearities and applications to phase-field models

作者:

Amjad Saef ↗

arXiv:2606.15425v1 Announce Type: new Abstract: We study well-posedness of stochastic parabolic equations with gradient nonlinearities. Our analysis is based on recent maximal-regularity frameworks for nonlinear stochastic parabolic equations in critical spaces. We extend the existing results by controlling drift and noise coefficient separately. This way we can allow for less regular driving noise in case of subcritical dispersion coefficients. Our approach, based on gluings of local solutions, moreover implies new continuation criteria. We then apply our existence result and the continuation criteria to show global well-posedness of phase-field models of moving boundary problems.

阅读与讨论 → 访问原文 →

17.

arXiv (CS.AI) 2026-06-15 DOI: arXiv:2606.14211

Closing the Reflection Gap: A Free Calibration Bonus for Agentic RL

作者:

Yinglun Zhu ↗

arXiv:2606.14211v1 Announce Type: new Abstract: LLMs are increasingly deployed as agents that interact with external environments and observe feedback such as execution results, error messages, and tool outputs. A well-functioning agent should be able to leverage this feedback to accurately assess its own performance. Yet we find a persistent reflection gap: LLM agents tend to mis-assess their own outputs after observing concrete environment feedback – even for questions they correctly answered – and standard RL barely helps due to a credit-assignment mismatch. To close this gap, we propose RefGRPO, a simple yet effective fix that augments standard RL algorithms with two key ingredients: a free calibration bonus computed by contrasting the agent's own reflection with the actual outcome (requiring no additional reward model, LLM judge, or external annotation), and a dynamic schedule on its coefficient. Compared to standard RL baselines, our method simultaneously improves reflection calibration (e.g., reduces underconfidence rate $44.4\% \to 7.7\%$) and task accuracy (e.g., $75.1\% \to 76.5\%$) on text-to-SQL across five benchmarks. The resulting calibrated reflection turns the agent into its own verifier grounded in environment feedback, which further enables (i) better self-improvement that uses reflections as pseudo-rewards without outcome supervision, and (ii) more effective test-time selective prediction by committing only to rollouts flagged as correct.

阅读与讨论 → 访问原文 →

18.

arXiv (CS.AI) 2026-06-15 DOI: arXiv:2606.13732

When Sample Selection Bias Precipitates Model Collapse

作者:

Xinbao Qiao ↗Xianglong Du ↗Wei Liu ↗Jingqi Zhang ↗Peihua Mai ↗Meng Zhang ↗Yan Pang ↗

arXiv:2606.13732v1 Announce Type: new Abstract: The proliferation of recursive training on synthetic data can alleviate data scarcity but risks model collapse, where repeated training erodes distributional tails and homogenizes outputs. Data selection is widely viewed as a remedy, yet its reliability depends critically on the reference distribution used by the verifier. We show that in low-resource verification regimes, where each verifier observes only a small, fragmented, and biased slice of the target manifold, selection itself becomes biased. This situation naturally arises in low-resource data silos such as healthcare consortia or proprietary financial institutions, where raw data cannot be pooled and local references are inherently incomplete. As a result, selection preferentially retains samples aligned with the local manifold while pruning globally relevant tail modes, turning from a safeguard against collapse into a mechanism that precipitates it. We theoretically prove that such siloed selection accelerates collapse and induces power-law diversity decay. As an initial mitigation, we construct Wasserstein proxy references from multiple silos without sharing raw data. Empirical results confirm that local-reference selection fails on skewed distributions, whereas collaborative proxy references mitigate diversity degradation, suggesting that recursive synthetic-data pipelines require particular caution when real-data coverage is fragmented or scarce.

阅读与讨论 → 访问原文 →

19.

arXiv (CS.AI) 2026-06-16 DOI: arXiv:2606.16532

Dual-Granularity Orthogonal Disentanglement for Generalizable Audio Deepfake Detection

作者:

Zhuodong Liu ↗Hugen Lv ↗Xiangyu Li ↗Chunhong Yuan ↗

arXiv:2606.16532v1 Announce Type: cross Abstract: Audio deepfake detectors often fail to generalize across speakers, as they learn speaker-identity features rather than synthesis artifacts, known as implicit identity leakage. Existing methods address this but incur architectural complexity or training instability. This paper proposes a dual-granularity orthogonal disentanglement framework enforcing feature independence at two levels: sample-level cosine orthogonality captures directional decorrelation, while batch-level cross-covariance regularization eliminates linear correlations across embedding dimensions. A curriculum disentanglement schedule progressively strengthens the orthogonality constraint without auxiliary networks or adversarial dynamics. Experiments on ASVspoof 2019 LA, ASVspoof 2021 DF, and In-the-Wild datasets demonstrate that the proposed method achieves 1.35%, 7.88%, and 21.58% equal error rates (EER), respectively, surpassing gradient reversal disentanglement by 2.60% absolute on cross-dataset transfer.

阅读与讨论 → 访问原文 →

20.

arXiv (CS.LG) 2026-06-19 DOI: arXiv:2604.06464

Weighted Bayesian Conformal Prediction

作者:

Xiayin Lou ↗Peng Luo ↗

arXiv:2604.06464v2 Announce Type: replace Abstract: Conformal prediction provides distribution-free prediction intervals with finite-sample coverage guarantees, and recent work by Snell \& Griffiths reframes it as Bayesian Quadrature (BQ-CP), yielding powerful data-conditional guarantees via Dirichlet posteriors over thresholds. However, BQ-CP fundamentally requires the i.i.d. assumption. Meanwhile, weighted conformal prediction handles distribution shift via importance weights but remains frequentist, producing only point-estimate thresholds. We propose Weighted Bayesian Conformal Prediction (WBCP), which generalizes BQ-CP to arbitrary importance-weighted settings by replacing the uniform Dirichlet $\Dir(1,\ldots,1)$ with a weighted Dirichlet $\Dir(\neff \cdot \tilde{w}_1, \ldots, \neff \cdot \tilde{w}_n)$, where $\neff$ is Kish's effective sample size. We prove four theoretical results: (1)~$\neff$ is the unique concentration parameter matching frequentist and Bayesian variances; (2)~posterior standard deviation decays as $O(1/\sqrt{\neff})$; (3)~BQ-CP's stochastic dominance guarantee extends to per-weight-profile data-conditional guarantees; (4)~the HPD threshold provides $O(1/\sqrt{\neff})$ improvement in conditional coverage. We instantiate WBCP for spatial prediction as Geographical BQ-CP, where kernel-based spatial weights yield per-location posteriors with interpretable diagnostics. Experiments on synthetic and real-world spatial datasets demonstrate that WBCP maintains coverage guarantees while providing substantially richer uncertainty information.

阅读与讨论 → 访问原文 →

21.

arXiv (CS.CV) 2026-06-25 DOI: arXiv:2606.25652

Auto-Labelling-Based Domain Transfer for 3D Object Detection on a Bicycle-Mounted LiDAR Platform

作者:

Mario Finkbeiner ↗Max A. Buettner ↗Kanak Mazumder ↗Fabian B. Flohr ↗

Reliable 3D perception of vulnerable road users (VRUs) such as cyclists and pedestrians is essential for their safety in urban traffic and a core requirement for autonomous driving (AD). Alongside advances in vehicle-based perception, research increasingly equips bicycles with sensors to study traffic from a perspective native to VRUs. Such platforms still rely on LiDAR detectors originally trained on vehicle data, yet annotated 3D data from a cyclist's perspective is scarce. How well these detectors generalise to this setting has not been evaluated. We present a 3D object detection benchmark of 1,027 annotated LiDAR keyframes (over 18,000 3D bounding boxes) from the FUSE-Bike platform in urban Munich. We evaluate four nuScenes-pre-trained detectors against 1,854 human-verified ground-truth (GT) boxes both in their original form and after finetuning on training labels produced by a VRU-dedicated auto-labelling pipeline that requires no manual annotation. The zero-shot domain gap is concentrated on the VRU classes. Finetuning recovers most of it, improving mean average precision (mAP) by up to 23.4 points with the largest gains on pedestrians and cyclists, and the adapted detectors even surpass the quality of the auto-labels they were trained on. The benchmark provides a reproducible baseline for VRU-centric 3D detection and shows that auto-labels are a viable substitute for manual annotation when adapting vehicle-trained detectors to a cyclist platform.

阅读与讨论 → 访问原文 →

22.

arXiv (quant-ph) 2026-06-24 DOI: arXiv:2606.24048

An Analysis of Speculative Window Decoders for Quantum Error Correction

作者:

Jocelyn Li ↗Margaret Martonosi ↗

arXiv:2606.24048v1 Announce Type: new Abstract: Fault-tolerant quantum computing is essential for realizing the substantial computational speedups that quantum computing can bring, but it requires real-time error decoding with high performance. Speculative window decoding improves performance by reducing the time spent waiting for dependencies from prior decoding windows. However, speculative decoders have only been evaluated under the regime of superconducting qubits with fast gate speeds, surface codes, and matching decoders. Since different quantum technologies can have slower gate speeds, we evaluate the performance of speculative decoding under slow gate speeds. We also examine its sensitivity to speculation accuracy, decoder latency, processor count, and workload parallelism, which can vary across different quantum error correction codes, decoders, and hardware platforms. This work presents design principles for identifying when speculative decoding yields the greatest performance improvements. It also reveals the conditions under which non-speculative decoders outperform speculative decoders.

阅读与讨论 → 访问原文 →

23.

arXiv (CS.AI) 2026-06-24 DOI: arXiv:2606.24429

Detecting AI Coding Agents in Open Source: A Validated Multi-Method Census of 180 Million Repositories

作者:

Arsham Khosravani ↗Audris Mockus ↗

arXiv:2606.24429v1 Announce Type: cross Abstract: Generative AI coding agents are entering the open-source supply chain, yet their diverse and often invisible traces leave their prevalence poorly understood. We introduce a multi-layered detection framework that integrates configuration-file scanning, commit-message analysis, author-identity matching, and bot-signature lookup across World of Code (180M+ Git repositories), classifying agent traces into four behavioral types. No single method captures more than a fraction of activity: multi-method detection identifies 850,157 Claude Code commits in one snapshot, of which bot-account lookup_the signal most adoption studies rely on_recovers only 28,154 (3.3%), a 30x relative-recall gap, so single-signal prevalence estimates are biased low by at least this factor. Every detection pattern is hand-validated (495 labels) with per-cell precision and Wilson confidence intervals. Across snapshots from December 2024 to April 2026, commit-attributed agents generate over 320,000 commits per month; Claude Code leads (886,122 commits across 17,295 projects) and dominates silent, configuration-file-only adoption (21,078 projects). Compared against an independent pull-request census (AIDev), the two channels capture nearly disjoint agent populations_a PR census misses 79% of commit-detected Claude Code adopters and essentially all Codex adopters_and different kinds of work: PR-deployed cloud agents (Codex, Cursor) surface as feature work, while commit-deployed in-editor agents (Claude Code, OpenHands, Aider) surface as maintenance. The observed work profile follows deployment and detection mode rather than the tool itself, so no single channel is representative.

阅读与讨论 → 访问原文 →

24.

arXiv (CS.AI) 2026-06-15 DOI: arXiv:2603.03970

Generative AI for Managerial Decision-Making under Ambiguity and Sycophancy

作者:

Sule Ozturk Birim ↗Fabrizio Marozzo ↗Yigit Kazancoglu ↗

arXiv:2603.03970v2 Announce Type: replace Abstract: Generative artificial intelligence (GenAI) is increasingly being integrated into complex business workflows, fundamentally shifting the boundaries of managerial decision-making. However, the reliability of its strategic advice in ambiguous business contexts remains a critical knowledge gap. To address this gap, this study compares multiple GenAI models in their ability to detect ambiguity, examines whether a systematic ambiguity-resolution process improves response quality, and investigates their susceptibility to sycophantic behavior when confronted with flawed managerial directives. Using a novel four-dimensional business ambiguity taxonomy, we conducted a human-in-the-loop experiment across strategic, tactical, and operational scenarios. The resulting decisions were assessed through a human-validated automated evaluation framework based on agreement, actionability, justification quality, and constraint adherence. The results show that our approach not only distinguishes different types of ambiguity, but also reveals how ambiguity resolution systematically changes model behavior. In particular, resolving ambiguities improved decision quality across all managerial levels, with the strongest gains observed in constraint adherence. The analysis further showed that sycophantic behavior is not uniform across models: some models challenged flawed assumptions, whereas others tended to comply with them. This study contributes to the bounded rationality literature by positioning GenAI as a cognitive scaffold that can detect and resolve ambiguities managers might overlook, while demonstrating that its artificial limitations require human oversight to ensure its reliability as a strategic partner.

阅读与讨论 → 访问原文 →

25.

arXiv (CS.LG) 2026-06-16 DOI: arXiv:2605.03573

Stochastic Schrödinger Diffusion Models for Pure-State Ensemble Generation

作者:

Jian Xu ↗Wei Chen ↗Shigui Li ↗Chao Li ↗Jingyuan Zheng ↗Delu Zeng ↗John Paisley ↗Qibin Zhao ↗

arXiv:2605.03573v3 Announce Type: replace-cross Abstract: Quantum machine learning increasingly relies on pure-state representations, motivating generative models that sample directly in quantum representation space rather than perturbing classical inputs and re-encoding. We introduce Stochastic Schrödinger Diffusion Models (SSDMs), a score-based generative framework that defines diffusion, scores, and reverse-time sampling intrinsically on the complex projective manifold $\mathbb{CP}^{d-1}$ under the Fubini–Study metric. SSDMs combine a Riemannian Ornstein–Uhlenbeck forward diffusion with a stochastic Schrödinger realization, and learn reverse-time dynamics driven by the Riemannian score. Our central technical contribution is a local-time learning objective that exploits the local Euclidean OU limit of intrinsic manifold diffusions in Fubini-Study normal coordinates to obtain an analytic teacher score, bypassing the intractable transition densities that limit existing Riemannian score-based models. Across synthetic, physics-inspired (TFIM, XXZ), and quantum feature-state benchmarks up to $14$ qubits, SSDMs match target pure-state ensembles by orders of magnitude on MMD and observable statistics over both ambient Euclidean and matched Riemannian score-based baselines, and improve representation-level diagnostics for downstream quantum kernel methods.

阅读与讨论 → 访问原文 →

探索全球前沿学术脉络