探索全球前沿学术脉络

01.

arXiv (CS.AI) 2026-06-18 DOI: arXiv:2606.19245

TxBench-PP: Analyzing AI Agent Performance on Small-Molecule Preclinical Pharmacology

作者:

Hannah Le↗Ramesh Ramasamy↗Alex Urrutia↗Mahsa Yazdani↗Tim Proctor↗Kenny Workman↗

arXiv:2606.19245v1 Announce Type: new Abstract: Artificial intelligence (AI) agents promise to accelerate drug discovery by compressing interpretation and decision-making loops, but practical deployment requires trusted evaluation on realistic program decisions. We introduce TherapeuticsBench Preclinical Pharmacology (TxBench-PP), a verifiable benchmark for small-molecule preclinical pharmacology and the first focused slice of a broader TherapeuticsBench effort across drug-discovery stages and therapeutic modalities. TxBench-PP tests whether agents can recover accurate conclusions from real-world assay data rather than memorized facts from literature. The benchmark contains 100 evaluations indexed by program stage, assay type, and task structure, spanning mechanism-of-action (MoA) and pharmacodynamic (PD) reasoning, compound-target engagement, causal target validation, developability and safety, and translational efficacy. Agents receive realistic workflow snapshots, inspect files in a coding environment, and return structured answers graded deterministically. Across 16 model-harness configurations, comprising 11 models and 4,800 trajectories, no system reliably recovered preclinical pharmacology decisions. The strongest configuration, Claude Opus 4.8 / Pi, passed 59.3\% of endpoint attempts (178/300; 95\% CI, 51.1-67.6), followed by GPT-5.5 / Pi at 55.3\% (166/300; 47.0-63.6).

阅读与讨论 → 访问原文 →

02.

arXiv (CS.CV) 2026-06-12 DOI: arXiv:2602.09730

Allure of Craquelure: A Variational-Generative Approach to Crack Detection in Paintings

作者:

Laura Paul↗Holger Rauhut↗Martin Burger↗Samira Kabri↗Tim Roith↗

Recent advances in imaging technologies, deep learning and numerical performance have enabled non-invasive detailed analysis of artworks, supporting their documentation and conservation. In particular, automated detection of craquelure in digitized paintings is crucial for assessing degradation and guiding restoration, yet remains challenging due to the possibly complex scenery and the visual similarity between cracks and crack-like artistic features such as brush strokes or hair. We propose a hybrid approach that models crack detection as an inverse problem, decomposing an observed image into a crack-free painting and a crack component. A deep generative model is employed as powerful prior for the underlying artwork, while crack structures are captured using a Mumford–Shah-type variational functional together with a crack prior. Joint optimization yields a pixel-level map of crack localizations in the painting.

阅读与讨论 → 访问原文 →

03.

medRxiv (Medicine) 2026-06-23 DOI: HASH:6fe89f1cdce1dc6d915191fb0e018e20

Uptake of minimal intervention dentistry among Romanian dental professionals and trainees: an exploratory cluster and network analysis

作者:

Simionescu↗D. P↗Sfeatcu↗Vinereanu↗

Background Minimal intervention dentistry (MID) is promoted as a prevention-oriented approach to caries management, but its integration into routine practice remains uneven. Existing research often examines MID-related knowledge, attitudes, or practices separately, offering limited insight into how these dimensions co-occur within individuals or are conditionally associated. Methods This exploratory cross-sectional survey examined multidimensional MID uptake among 327 Romanian dental students, residents, and specialists from five university centers. Ten MID-related scores were analyzed, including nine formative composites and one single-item peer-norm indicator. K-means clustering examined uptake profiles, and Gaussian graphical model network analysis with stepwise BIC selection examined conditional associations among constructs. Results A two-cluster solution was highly reproducible but modestly separated (n = 144 vs n = 183; average silhouette width = 0.13; mean Jaccard similarities = 0.92 and 0.94). The profiles reflected broadly lower versus higher uptake across knowledge-, belief-, and practice-related dimensions, while perceived peer norms for hygiene instruction showed the opposite pattern. Profile membership was not clearly patterned by gender, age band, professional status, or clinical experience. The primary network included 14 non-zero edges out of 36 possible edges, all positive; the strongest partial association linked diagnostic knowledge to diagnostic methods used in practice (partial r = .22). Familiarity, diagnostic knowledge, and general practices occupied more interconnected positions descriptively, but limited centrality stability precluded interpreting them as intervention targets. Conclusions MID uptake in this sample was better represented as a continuum of modestly differentiated profiles than as sharply separated participant types. The findings provide an exploratory map of multidimensional MID uptake and may inform future survey validation, implementation research, and dental education studies. Because the study was cross-sectional, convenience-sampled, and based on self-report, findings should be interpreted as hypothesis-generating rather than causal or population-representative.

阅读与讨论 → 访问原文 →

04.

arXiv (CS.AI) 2026-06-12 DOI: arXiv:2605.03460

FinSTaR: Towards Financial Reasoning with Time Series Reasoning Models

作者:

Seunghan Lee↗Jun Seo↗Jaehoon Lee↗Sungdong Yoo↗Minjae Kim↗Tae Yoon Lim↗Dongwan Kang↗Hwanil Choi↗Soonyoung Lee↗Wonbin Ahn↗

arXiv:2605.03460v3 Announce Type: replace Abstract: Time series (TS) reasoning models (TSRMs) have shown promising capabilities in general domains, yet they consistently fail in the financial domain, which exhibits unique characteristics. We propose a general 2 x 2 capability taxonomy for TSRMs by crossing 1) single-entity vs. multi-entity analysis with 2) assessment of the current state vs. prediction of future behavior. We instantiate this taxonomy in the financial domain-where the distinction between deterministic assessment and stochastic prediction is particularly critical-as ten financial reasoning tasks, forming the FinTSR-Bench benchmark based on S&P stocks. To this end, we propose FinSTaR (Financial Time Series Thinking and Reasoning), trained on FinTSR-Bench with distinct chain-of-thought (CoT) strategies tailored to each category. For assessment, which is deterministic (i.e., computable from observable data), we employ Compute-in-CoT, a programmatic CoT that enables models to derive answers directly from raw prices. For prediction, which is inherently stochastic (i.e., subject to unobservable factors), we adopt Scenario-Aware CoT, which generates diverse scenarios before making a judgment, mirroring how financial analysts reason under uncertainty. The proposed method achieves 78.9% average accuracy on FinTSR-Bench, substantially outperforming LLM and TSRM baselines. Furthermore, we show that the four capability categories are complementary and mutually reinforcing through joint training, and that Scenario-Aware CoT consistently improves prediction accuracy over standard CoT. Code is available at https://github.com/seunghan96/FinSTaR.

阅读与讨论 → 访问原文 →

05.

arXiv (CS.AI) 2026-06-12 DOI: arXiv:2605.03847

Mechanical Conscience: A Mathematical Framework for Dependability of Machine Intelligenc

作者:

Munkhdegerekh Batzorig↗Purevbaatar Ganbold↗Kyungbin Park↗Pilkong Jeong↗Kangbin Yim↗

arXiv:2605.03847v2 Announce Type: replace Abstract: Distributed collaborative intelligence (DCI), encompassing edge-to-edge architectures, federated learning, transfer learning, and swarm systems, creates environments in which emergent risk is structurally unavoidable: locally correct decisions by individual agents compose into globally unacceptable behavioral trajectories under uncertainty. Existing approaches such as constrained optimization, safe reinforcement learning, and runtime assurance evaluate acceptability at the level of individual actions rather than across behavioral trajectories, and none addresses the multi-participant, uncertainty-laden nature of DCI deployments. This paper introduces mechanical conscience (MC), a novel concept and simplified mathematical framework that operationalizes trajectory-level normative regulation for both single-agent and distributed intelligent systems. Mechanical conscience is defined as a supervisory filter that minimally corrects a baseline policy's actions to reduce cumulative deviation from a normatively admissible region, while accounting for epistemic uncertainty. We introduce associated constructs, conscience score, mechanical guilt, and resonant dependability, that provide an interpretable vocabulary and computable governance signals for this emerging field. Core theoretical properties are established: admissibility equivalence, existence of optimal regulation, and monotonic deviation reduction. Illustrative results demonstrate that MC-regulated agents maintain trajectory-level normative acceptability where conventional controllers drift outside admissible bounds, and that the framework naturally extends to suppress interaction-induced emergent risk in multi-agent DCI settings.

阅读与讨论 → 访问原文 →

06.

arXiv (CS.CL) 2026-06-19 DOI: arXiv:2606.20527

StylisticBias: A Few Human Visual Cues Drive Most Social Biases in MLLMs

作者:

Shaghayegh Kolli↗Timo Cavelius↗Nafiseh Nikeghbal↗Samantha Dalal↗Jana Diesner↗

Multimodal large language models (MLLMs) are increasingly deployed in personally and societally consequential settings, yet the visual cues that shape how these models judge people remain poorly understood. Prior work often compares different (groups of) individuals, making it difficult to separate appearance effects from identity differences. We introduce StylisticBias, a controlled benchmark for evaluating attribute-level social bias in MLLMs. We generate 500 photorealistic base faces and create about 50 single-attribute variations per face, producing about 25K images. This design keeps identity fixed and changes one visual attribute at a time. It lets us measure how specific cues shift model judgments. We evaluate six MLLMs across 25 binary social judgment scenarios. We find that age and body type dominate identity-level effects, while fashion style and other visual cues drive the largest attribute-level shifts. We further find that about 15 attributes account for nearly 80\% of the total variation, showing that bias is concentrated in a small set of visual cues. Sensitivity is strongest in judgments that are semantically aligned with appearance, especially socioeconomic and style-related judgments. We release StylisticBias as a benchmark for fine-grained bias evaluation in multimodal models. Code and dataset: https://github.com/timo-cavelius/StylisticBias and https://hf.co/datasets/shaghayegh/stylistic-bias-dataset.

阅读与讨论 → 访问原文 →

07.

arXiv (CS.AI) 2026-06-12 DOI: arXiv:2606.09500

Deterministic Integrity Gates for LLM-Assisted Clinical Manuscript Preparation: An Auditable Biomedical Informatics Architecture

作者:

Yoojin Nam↗Jinhoon Jeong↗Namkug Kim↗

arXiv:2606.09500v3 Announce Type: replace Abstract: As autonomous research agents and AI co-scientist systems push large language models (LLMs) from drafting toward end-to-end manuscript production, the bottleneck shifts from generation to verification. Fluent LLM output can hide fabricated citations, numbers that drift from source tables, and unmet reporting-guideline items; existing tools generate without verifying, and self-critique inherits the blind spots that produce confident fabrication. We describe an architecture pairing generation with verification, resting on three principles: decompose the workflow into self-contained skills, gate every stage transition with halt-on-failure, and resolve each integrity question with the cheapest sufficient mechanism, a deterministic, re-executable check where one suffices and a prose-level probe only where interpretation is unavoidable. This determinism-where-possible split, organized as an integrity-gate taxonomy, is the core contribution. It is realized as MedSci Skills, an open-source toolkit of 43 skills with a 21-detector deterministic tier, evaluated on three public-dataset pipelines (STARD, PRISMA, STROBE) and a seeded-defect ablation. Across the three pipelines every content-hash manifest verified clean and the gates surfaced real defects; on 27 identical injected defects the deterministic gates detected all 27 with no false positives on the matched clean fixtures, whereas a single-prompt LLM reviewer detected 11, its misses in code, bibliography, and style defects the prose hides. Determinism-where-possible verification yields an auditable, re-executable trail that exposes the evidence a human needs to check an LLM-assisted manuscript: feasibility and reproducibility evidence, not a claim of human-competitive quality, which a separate blinded study addresses. MedSci Skills is MIT-licensed and archived (v3.8.0).

阅读与讨论 → 访问原文 →

08.

medRxiv (Medicine) 2026-06-17 DOI: HASH:568264cbe9466b752a9aeeb4aff35e94

Method comparisons for differentiation of Schizophrenia and Bipolar based on rs-fMRI Intrinsic and Functional Networks

作者:

Janeva↗Breyton↗Markovska-Simoska↗Guilhaumou↗Petkoski↗Iraji↗Calhoun↗Gerazov↗

Psychosis as a symptom manifests in schizophenia and bipolar disorder, two highly heterogeneous psychiatric illnesses with overlapping clinical manifestations. Resting-state functional Magnetic Resonance Imaging (rsfMRI), represents a promising tool for identifying objective biomarkers of functional brain alterations to aid differential diagnosis. In this work, we comparatively evaluate multiple rs-fMRI representations for differentiating schizophrenia and bipolar disorder using intrinsic connectivity network (ICN) temporal profiles and several functional network connectivity (FNC) approaches, including static, dynamic, and high-order connectivity analyses. The study was conducted on a cohort of 371 subjects with psychosis, while evaluation was performed using a separate held-out cohort of 315 subjects. We investigated convolutional neural network architectures applied to ICN temporal profiles, spectrograms, and scalograms, alongside classical machine learning models trained on connectivity-derived features. Across the evaluated approaches, ICN temporal profiles provided the most consistent discriminative performance, with a 1D convolutional neural network achieving the strongest overall results under the benchmark protocol. Among connectivity-based methods, static functional connectivity generally outperformed dynamic and high-order representations, suggesting that increased representational complexity did not necessarily translate into improved generalization. Although the obtained classification performance remained modest, the results highlight the challenges of robust psychosis differentiation using rs-fMRI while emphasizing the relative stability of low-order connectivity representations and temporal ICN features. These findings contribute to ongoing efforts toward reproducible and interpretable neuroimaging biomarkers for psychiatric disorders.

阅读与讨论 → 访问原文 →

09.

arXiv (CS.AI) 2026-06-19 DOI: arXiv:2606.19948

Advancing DialNav through Automatic Embodied Dialog Augmentation

作者:

Leekyeung Han↗Sangwon Jung↗Hyunji Min↗Jinseong Jeong↗Minyoung Kim↗Paul Hongsuck Seo↗

arXiv:2606.19948v1 Announce Type: new Abstract: For embodied agents capable of physical interaction, the capability to create and understand dialog is crucial to ensure both safety and effectiveness. While DialNav[han2025dialnav] provides a framework for holistic evaluation of the dialog–execution loop in photorealistic indoor navigation, its performance remains limited by a critical scarcity of training data (2K episodes). To address this, we propose an automatic generation pipeline, and construct the RAINbow dataset, a large-scale training dataset with 238K episodes for DialNav. Our pipeline converts existing VLN datasets into multi-turn dialog and creates cost-efficient and high-quality dataset. Then, we introduce two additional complementary advances to unlock the data's full potential: (1) Dual-Strategy Training, a navigation training scheme to align the navigation training with the dynamic dialog-navigation loop, and (2) a localization model that leverages VLN knowledge. By combining these complementary solutions, our model substantially outperforms the baseline in success rate on both Val Seen (58.24, +89\%) and Val Unseen (29.05, +100\%) splits, establishing a new state of the art.

阅读与讨论 → 访问原文 →

10.

arXiv (CS.AI) 2026-06-16 DOI: arXiv:2606.14892

Relational Structural Causal Models

作者:

Adiba Ejaz↗Elias Bareinboim↗

arXiv:2606.14892v1 Announce Type: new Abstract: An artificial intelligence must have a model of its environment that is causal, supporting reasoning about interventions and counterfactuals, and also combinatorial, supporting generalization to unseen combinations of objects. In this work, we formally study when and how such a model can be learned. We develop relational structural causal models, extending structural causal models (Pearl 2009) to settings where objects and their relations vary. First, we show how answers to not only causal but also observational queries about unseen combinations of objects can not be identified without further assumptions. To enable such identification–including in the presence of unobserved confounding–we define relational causal graphs and derive symbolic identification criteria. Finally, we propose relational neural causal models, a provably correct approach that outperforms non-relational baselines on simulated traffic scenes with varying cars, signals, and pedestrians.

阅读与讨论 → 访问原文 →

11.

arXiv (CS.LG) 2026-06-18 DOI: arXiv:2412.16468

The Road to Artificial SuperIntelligence: A Comprehensive Survey of Superalignment

作者:

HyunJin Kim↗DongHyun Ryu↗Xiaoyuan Yi↗Jing Yao↗Jianxun Lian↗Muhua Huang↗Shitong Duan↗JinYeong Bak↗Xing Xie↗

arXiv:2412.16468v4 Announce Type: replace Abstract: The emergence of large language models (LLMs) has sparked discussion on Artificial Superintelligence (ASI), a hypothetical AI system that surpasses human intelligence. Although ASI remains hypothetical and far beyond current AI capabilities, discussing its potential and exploring its feasibility and potential risks is critical for the development of future AI systems. The idea of superalignment originates from scalable oversight, which studies how to supervise increasingly capable AI systems when direct human supervision becomes insufficient. In this paper, we focus on the superalignment problem: "The process of supervising, controlling, and governing artificial superintelligence." We first review scalable oversight paradigms-Sandwiching, Self-Enhancement, and Weak-to-Strong Generalization – then analyze the limitations of current paradigms through the lens of possibility and impossibility, discuss key challenges, and propose pathways for the safe and continual improvement of future AI systems.

阅读与讨论 → 访问原文 →

12.

arXiv (quant-ph) 2026-06-16 DOI: arXiv:2606.16071

Bright Emission from Dark Sources in Hyperbolic Media

作者:

Evgenii E. Narimanov↗

arXiv:2606.16071v1 Announce Type: cross Abstract: Hyperbolic media enable ultra-strong light-matter interactions through their extreme field localization and small mode volumes, but low-loss realizations are fundamentally limited to the mid-infrared, owing to the long lifetimes of optical phonons in high-quality crystals. Here we show that bright emitters operating at visible or near-infrared frequencies can be used to generate radiation in this regime by inducing mid-infrared population dynamics, thereby creating a source in the hyperbolic frequency band without a corresponding dipole transition. We demonstrate that even a source with vanishing dipole and higher multipole moments - strictly non-radiating in any isotropic medium - becomes radiatively active in a hyperbolic environment. This enables visible and near-infrared control of light-matter interactions in polaritonic hyperbolic materials, establishing a new low-loss solid-state quantum optics platform.

阅读与讨论 → 访问原文 →

13.

arXiv (CS.AI) 2026-06-24 DOI: arXiv:2605.08245

When Language Overwrites Vision: Over-Alignment and Geometric Debiasing in Vision-Language Models

作者:

Harshvardhan Saini↗Samyak Jha↗Yiming Tang↗Dianbo Liu↗

arXiv:2605.08245v4 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) increasingly power high-stakes applications, from medical imaging to autonomous systems, yet they routinely hallucinate, confidently describing content not present in the input. We investigate the root causes of these failure modes with a mechanistic analysis focusing on the decoder-based VLMs. We trace these failure modes to a geometric over-alignment: to bridge the modality gap required by attention mechanisms, decoder-based VLMs over-align visual embeddings with the text manifold, injecting a statistical linguistic bias that systematically overshadows fine-grained visual evidence. While prior work either aggressively closes this gap or suppresses hallucinations through expensive black-box decoding strategies, none addresses the underlying geometric cause. We provide the first quantitative characterization of this over-alignment, demonstrating that linguistic bias concentrates in the top principal components of a universal, dataset-agnostic text subspace. Building on this insight, we propose two complementary remedies: a training-free inference strategy and a bias-aware fine-tuning paradigm, both of which explicitly project out this subspace from visual representations. Our methods significantly reduce hallucinations across POPE, CHAIR, and AMBER benchmarks, and improve CLAIR scores on long-form captioning tasks, with the training-free variant adding no computational overhead over the base model.

阅读与讨论 → 访问原文 →

14.

arXiv (CS.CV) 2026-06-11 DOI: arXiv:2409.18478

Temporal2Seq: A Unified Framework for Temporal Video Understanding Tasks

作者:

Min Yang↗Zichen Zhang↗Qian Dang↗Limin Wang↗

With the development of video understanding, there is a proliferation of tasks for clip-level temporal video analysis, including temporal action detection (TAD), temporal action segmentation (TAS), and generic event boundary detection (GEBD). While task-specific video understanding models have exhibited outstanding performance in each task, there remains a dearth of a unified framework capable of simultaneously addressing multiple tasks, which is a promising direction for the next generation of AI. To this end, in this paper, we propose a single unified framework, coined as Temporal2Seq, to formulate the output of these temporal video understanding tasks as a sequence of discrete tokens. With this unified token representation, Temporal2Seq can train a generalist model within a single architecture on different video understanding tasks. In the absence of multi-task learning (MTL) benchmarks, we compile a comprehensive co-training dataset by borrowing the datasets from TAD, TAS, and GEBD tasks. We evaluate our Temporal2Seq generalist model on the corresponding test sets of three tasks, demonstrating that Temporal2Seq can produce reasonable results on various tasks and achieve advantages compared with single-task training on this framework. We also investigate the generalization performance of our generalist model on new datasets from different tasks, which yields superior performance to the specific model.

阅读与讨论 → 访问原文 →

15.

arXiv (CS.LG) 2026-06-18 DOI: arXiv:2602.11467

PRISM: A 3D Probabilistic Neural Representation for Interpretable Shape Modeling

作者:

Yining Jiao↗Sreekalyani Bhamidi↗Carlton Jude Zdanski↗Julia S Kimbell↗Andrew Prince↗Cameron P Worden↗Samuel Kirse↗Christopher Rutter↗Benjamin H Shields↗Jisan Mahmud↗Marc Niethammer↗

arXiv:2602.11467v2 Announce Type: replace Abstract: Understanding how anatomical shapes evolve in response to developmental covariates - and quantifying their spatially varying uncertainties - is critical in healthcare research. Existing approaches typically rely on global time-warping formulations that ignore spatially heterogeneous dynamics. We introduce PRISM, a novel framework that bridges implicit neural representations with uncertainty-aware statistical shape analysis. PRISM models the conditional distribution of shapes given covariates, providing spatially continuous estimates of both the population mean and covariate-dependent uncertainty at arbitrary locations. A key theoretical contribution is a closed-form Fisher Information metric that enables efficient, analytically tractable local temporal uncertainty quantification via automatic differentiation. Experiments on three synthetic datasets and one clinical dataset demonstrate PRISM's strong performance across diverse tasks - from modeling shape evolution to personalized shape prediction and anomaly detection - within a unified framework, while providing interpretable and clinically meaningful uncertainty estimates.

阅读与讨论 → 访问原文 →

16.

arXiv (CS.AI) 2026-06-19 DOI: arXiv:2606.19595

IHBench: Evaluating Post-Interruption Recovery in Voice Agents with Structured Workflows

作者:

Ahmad Salimi↗Wentao Ma↗Yuzhi Tang↗Dongming Shen↗Mu Li↗Alex Smola↗

arXiv:2606.19595v1 Announce Type: cross Abstract: Voice agents deployed in structured workflows (customer service, healthcare scheduling, account management) must handle frequent user interruptions while maintaining progress through multi-step procedures. Existing benchmarks for speech-capable models focus on the timing of interruptions: barge-in detection, endpointing, and turn-taking dynamics. They leave unmeasured what happens after the interruption: does the agent resume the workflow at the correct step? Does it address the user's interjection? Does it avoid re-delivering content the user already heard? We introduce IHBench (Interruption Handling Benchmark), a benchmark that evaluates post-interruption recovery in voice agents executing state-machine-driven workflows across 10 enterprise domains. Six interruption types are injected at controlled points mid-utterance, with per-interruption evaluation rubrics generated alongside the data. Each interruption is scored on two axes: task fulfillment and recovery quality. We evaluate 27 audio-language model configurations from OpenAI, Google, and the open-weight community. Models vary widely, and recovery quality depends strongly on the interruption type. Across our experiments, closed-weight models are consistently more robust to interruptions than open-weight ones: they win far more often on task fulfillment, degrade roughly 3.3x more slowly as conversations grow longer, and show no audio-versus-text modality gap, whereas the open-weight models lose ground on all three. A human study validates the LLM judge against human annotators, and a cross-benchmark analysis against AudioMultiChallenge indicates that recovery quality is a largely distinct capability axis.

阅读与讨论 → 访问原文 →

17.

arXiv (CS.CL) 2026-06-11 DOI: arXiv:2510.01157

Where Do Backdoors Live? A Component-Level Analysis of Backdoor Propagation in Speech Language Models

作者:

Alexandrine Fortier↗Thomas Thebaud↗Jes\'us Villalba↗Najim Dehak↗Patrick Cardinal↗Peter West↗

Speech language models (SLMs) are systems of systems: independent components that unite to achieve a common goal. Despite their heterogeneous nature, SLMs are often studied end-to-end; how information flows through the pipeline remains obscure. We investigate this question through the lens of backdoor attacks. We first establish that backdoors can propagate through the SLM, leaving all tasks highly vulnerable. From this, we design a component analysis to discover the role each component takes in backdoor learning. We find that backdoor persistence or erasure is highly dependent on the targeted component. Beyond propagation, we examine how backdoors are encoded in shared multitask embeddings, showing that poisoned samples are not directly separable from benign ones, challenging a common separability assumption used in filtering defenses. Our findings emphasize the need to treat multimodal pipelines as intricate systems with unique vulnerabilities, not solely extensions of unimodal ones.

阅读与讨论 → 访问原文 →

18.

arXiv (CS.AI) 2026-06-11 DOI: arXiv:2606.11537

MoCA-Agent: A Market-of-Claims Code Agent for Financial and Numerical Reasoning

作者:

Abdelrahman Abdallah↗AbdelRahim A. Elmadany↗Sameh Al Natour↗Hasan Cavusoglu↗Adam Jatowt↗Muhammad Abdul-Mageed↗

arXiv:2606.11537v1 Announce Type: new Abstract: Financial and tabular question answering requires more than fluent reasoning: answers must be grounded in the exact facts, formulas, units, signs, and scales that support them. A single misread cell or incorrect operation can silently produce a plausible but wrong result. We introduce \textsc{MOCA-Agent}, a market-of-claims code agent that replaces free-form multi-agent debate with claim-level verification. The system decomposes each question into typed atomic claims, asks specialist trader agents to buy or sell those claims, clears their orders into confidence-weighted accept/reject decisions, and synthesizes an executable Python program from market-supported evidence. A code-aware verifier then checks the program for execution, structural consistency, and common financial reasoning errors, with at most one market-aware repair round. Across ten public benchmarks spanning financial numerical reasoning, general tabular reasoning, ESG question answering, and multimodal chart reasoning, \textsc{MOCA-Agent} achieves strong performance using a fixed Qwen3.6-27B backbone, including $78.3\%$ on FinQA, $76.0\%$ on FinanceMath, $71.2\%$ on MultiHiertt, $86.9\%$ on ESGenius, and $85.6\%$ average on FinChart-Bench. These results show that aggregating evidence at the level of atomic claims, rather than whole answers, improves robustness in high-stakes numerical reasoning.\footnote{The code and data are available: https://github.com/UBC-NLP/MoCA-Agent.

阅读与讨论 → 访问原文 →

19.

arXiv (CS.CL) 2026-06-17 DOI: arXiv:2606.18216

Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients

作者:

Byung-Kwan Lee↗Ximing Lu↗Shizhe Diao↗Minki Kang↗Saurav Muralidharan↗Karan Sapra↗Andrew Tao↗Pavlo Molchanov↗Yejin Choi↗Yu-Chiang Frank Wang↗Ryo Hachiuma↗

Knowledge distillation transfers a teacher's competence to a small student but is brittle in the small-student regime: forcing the student to imitate logits from a much larger teacher concentrates it on the teacher's sharpest modes, hurting generalization on benchmark families beyond the training corpus. Reinforcement learning (RL) avoids logit imitation by training on the student's own rollouts. However, on questions where every rollout fails-yielding zero advantage and being silently discarded-injecting a stronger teacher's response into the policy gradient breaks the on-policy assumption and induces drift. We introduce Zone of Proximal Policy Optimization (ZPPO), inspired by Vygotsky's zone of proximal development, which keeps the teacher inside the prompt rather than the policy gradient. On hard questions, ZPPO constructs two reformulated prompts: a Binary Candidate-included Question (BCQ) pairs one correct teacher response with one incorrect student response as anonymized candidates the student must discriminate, and a Negative Candidate-included Question (NCQ) aggregates the student's wrong rollouts into a single prompt to surface their shared failure modes. A prompt replay buffer recirculates each hard question until it either graduates-the student's mean rollout accuracy on it reaches half- or is FIFO-evicted under finite capacity, amplifying BCQ and NCQ inside the student's current zone of proximal development. On the Qwen3.5 family at four student scales (0.8B-9B) with a 27B teacher, post-trained as vision-language models and evaluated on a 31-benchmark suite (16 VLM, 10 LLM, 5 Video), ZPPO outperforms off/on-policy distillation and GRPO, with the largest gains at the smallest scale.

阅读与讨论 → 访问原文 →

20.

arXiv (CS.CL) 2026-06-18 DOI: arXiv:2606.19266

Trade-offs in Medical LLM Adaptation: An Empirical Study in French QA

作者:

Ikram Belmadani↗Oumaima El Khettari↗Carlos Ramisch↗Frederic Bechet↗Richard Dufour↗Benoit Favre↗

The development of large language models (LLMs) has led to an increased focus on their adaptation to specialized domains and languages, yet the effectiveness of domain adaptation strategies remains unclear. We present a study of medical domain adaptation using French medical question-answering (QA) as a case study. We compare continual pretraining (CPT), supervised fine-tuning (SFT), and their combination across three model families, multiple sizes, and three initialization types, explicitly disentangling adaptation effects from base model choice. We evaluate both multiple-choice (MCQA) and open-ended QA (OEQA) under greedy and constrained decoding using automatic metrics and LLM-as-a-Judge evaluation. For MCQA, CPT+SFT most often achieves the best scores, but gains over SFT are small and frequently not statistically significant, making SFT a strong and cost-effective default. For OEQA, CPT consistently improves overlap-based metrics, while SFT often degrades generation quality; instruction tuning and CPT+SFT are preferred by LLM-based evaluation. Cross-lingual experiments further show effective transfer from French adaptation to English benchmarks. Overall, we provide practical guidelines for selecting adaptation strategies under computational constraints.

阅读与讨论 → 访问原文 →

21.

arXiv (CS.CL) 2026-06-16 DOI: arXiv:2606.16934

Exploring Extrinsic and Intrinsic Properties for Effective Reasoning with Code Interpreter

作者:

Patomporn Payoungkhamdee↗Napat Laosaengpha↗Jenta Wonglertsakul↗Pittawat Taveekitworachai↗Pume Tuchinda↗Panjapong Poobanchuen↗Ekapol Chuangsuwanich↗Can Udomcharoenchaikit↗Samuel Cahyawijaya↗Peerat Limkonchotiwat↗Sarana Nutanong↗

Reasoning with a Code Interpreter (CI) has emerged as an effective paradigm for enhancing the reasoning capabilities of large language models (LLMs) through executable computation and iterative verification. Despite its growing adoption, the behavioral properties underlying effective code reasoning remain largely underexplored. In this work, we investigate code reasoning from two distinct perspectives inspired by prior studies of natural language reasoning: extrinsic properties, represented by crucial tokens, and intrinsic properties, represented by code-specific cognitive behaviors. Across multiple LLMs, we find that stronger CI reasoning models consistently exhibit a higher prevalence of crucial tokens and cognitive behaviors, particularly verification, backtracking, and backward chaining. Building on these observations, we examine how these properties can be leveraged during both inference and training. At inference time, appending code-specific crucial tokens improves performance on several reasoning capabilities, including mathematical, ordering, and optimization, while yielding limited benefits elsewhere. At training time, augmenting a state-of-the-art framework with code-specific cognitive behaviors improves supervised fine-tuning and reinforcement learning performance in two of three evaluated models. Further analysis shows that these behaviors reduce overthinking in incorrect responses and improve token efficiency, while also revealing factors that limit gains in a certain model. Our findings provide the first systematic characterization of effective reasoning with CI and demonstrate both the potential and limitations of leveraging key properties to improve CI-based reasoning.

阅读与讨论 → 访问原文 →

22.

arXiv (CS.CV) 2026-06-16 DOI: arXiv:2606.16533

Kairos: A Native World Model Stack for Physical AI

作者:

Kairos Team↗Fei Wang↗Shan You↗Qiming Zhang↗Tao Huang↗Zuoyi Fu↗Zhisheng Zheng↗Yunlong Xi↗Feng Lv↗Xiaoming Wu↗Zeyu Liu↗Cong Wan↗…

World models are transitioning from passive visual generators to foundational, operational infrastructure for Physical AI: they must natively acquire world knowledge from heterogeneous experience, maintain persistent states over long horizons, and execute efficiently within real deployment constraints. We introduce Kairos, a native world model stack designed around these requirements. (1) Kairos learns the world by pioneering a Native Pre-training Paradigm governed by a Cross-Embodiment Data Curriculum, which organizes open-world videos, human behavioral data, and robot interactions into a progressive developmental pathway. (2) Kairos maintains the world by unified world understanding, generation, and prediction within a Native Unified Architecture equipped with Hybrid Linear Temporal Attention, where sliding-window attention captures local dynamics, dilated sliding windows capture mid-range dependencies, and gated linear attention maintains persistent global memory. We establish formal theoretical bounds demonstrating that this temporal factorization strictly limits error accumulation, mathematically guaranteeing state propagation across extended horizons. (3) Kairos runs the world by incorporating a Deployment-Aware System Co-Design to support low-latency rollout generation on server and consumer-grade hardware for real-world observation-action-feedback loops. Experiments on embodied world-model, long-horizon, and action-policy benchmarks show that Kairos achieves top level performance while offering a strong efficiency-capability trade-off. Together, these results position Kairos as a cohesive operational foundation for future self-evolving physical intelligence.

阅读与讨论 → 访问原文 →

23.

arXiv (CS.AI) 2026-06-12 DOI: arXiv:2603.00610

CMI-RewardBench: Evaluating Music Reward Models with Compositional Multimodal Instruction

作者:

Yinghao Ma↗Haiwen Xia↗Hewei Gao↗Weixiong Chen↗Yuxin Ye↗Yuchen Yang↗Sungkyun Chang↗Mingshuo Ding↗Yizhi Li↗Ruibin Yuan↗Simon Dixon↗Emmanouil Benetos↗…

arXiv:2603.00610v3 Announce Type: replace-cross Abstract: While music generation models have evolved to handle complex multimodal inputs mixing text, lyrics, and reference audio, evaluation mechanisms have lagged behind. In this paper, we bridge this critical gap by establishing a comprehensive ecosystem for music reward modeling under Compositional Multimodal Instruction (CMI), where the generated music may be conditioned on text descriptions, lyrics, and audio prompts. We first introduce CMI-Pref-Pseudo, a large-scale preference dataset comprising 110k pseudo-labeled samples, and CMI-Pref, a high-quality, human-annotated corpus tailored for fine-grained alignment tasks. To unify the evaluation landscape, we propose CMI-RewardBench, a unified benchmark that evaluates music reward models on heterogeneous samples across musicality, text-music alignment, and compositional instruction alignment. Leveraging these resources, we develop CMI reward models (CMI-RMs), a parameter-efficient reward model family capable of processing heterogeneous inputs. We evaluate their correlation with human judgment scores on musicality and alignment on CMI-Pref along with previous datasets. Further experiments demonstrate that CMI-RM not only correlates strongly with human judgments, but also enables effective inference-time scaling via top-k filtering. Code is available at GitHub (https://github.com/Haiwen-Xia/CMI-RewardBench). Model weights: CMI-RM (https://huggingface.co/HaiwenXia/CMI-RM). Datasets: CMI-Pref-Pseudo (https://huggingface.co/datasets/HaiwenXia/cmi-pref-pseudo) and CMI-Pref (https://huggingface.co/datasets/HaiwenXia/cmi-pref)

阅读与讨论 → 访问原文 →

24.

arXiv (CS.CL) 2026-06-11 DOI: arXiv:2603.08501

Fanar-Sadiq: A Multi-Agent Architecture for Grounded Islamic QA

作者:

Ummar Abbas↗Mourad Ouzzani↗Mohamed Y. Eltabakh↗Omar Sinan↗Gagan Bhatia↗Hamdy Mubarak↗Majd Hawasly↗Mohammed Qusay Hashim↗Kareem Darwish↗Firoj Alam↗

Large language models (LLMs) can answer religious knowledge queries fluently, yet they often hallucinate and misattribute sources, which is especially consequential in Islamic settings where users expect grounding in canonical texts (Qur'an and Hadith) and jurisprudential (fiqh) nuance. Retrieval-augmented generation (RAG) improves grounding, however, a single retrieve-then-generate pipeline is insufficient for diverse Islamic queries, including verbatim scripture, citation-grounded guidance, and rule-constrained computations such as zakat and inheritance. To address these challenges, we present Fanar-Sadiq, a bilingual Arabic-English Islamic QA system built on a multi-agent, tool-augmented architecture. It is a core component of the Fanar AI platform. Fanar-Sadiq routes Islamic queries to specialized modules within an agentic tool architecture. It supports intent-aware routing, retrieval-grounded fiqh answers with normalized citations and verification traces, exact verse lookup with quotation validation, and deterministic Sunni zakat and inheritance calculators with madhhab-sensitive branching. We evaluate the end-to-end system on public Islamic QA benchmarks and show strong effectiveness and efficiency. It is publicly accessible through an API and Web application and has received over 1.9M accesses in less than a year (https://api.fanar.qa/docs).

阅读与讨论 → 访问原文 →

25.

arXiv (CS.CV) 2026-06-19 DOI: arXiv:2606.19944

Timage: A Generative Text-in-Image Paradigm for Fine-Tuning Vision-Language Models

作者:

Yifeng Wu↗Huimin Huang↗Ruiluo Wu↗Chunyi Lin↗Guanhua Chen↗Xian Wu↗Wang Song↗Ruize Han↗

Multimodal Large Language Models (MLLMs) often lose track of the right image regions during fine-grained spatial reasoning, because a textual query rarely carries any explicit geometric anchor into the pixel domain. Prevailing remedies either rewire the model's weights or pad the prompt with verbose instructions, yet neither reliably pins the language to the correct visual coordinates without eroding the backbone's general competence. We introduce Timage, a paradigm that recasts multimodal understanding as an alignment problem solved at the input: the query is drawn, as a typeset overlay, onto the image itself. The placement and appearance of this overlay are produced by a Constrained Schrödinger Bridge (cSB), an entropic optimal-transport sampler that factorizes layout synthesis into two coupled stochastic stages. The first stage, Region Search, transports noise toward query-aligned image zones while obeying a hard occlusion barrier that protects salient foreground content; the second stage, Appearance Shaping, sizes the glyphs through an ``ink-budget'' regularizer so that the rendered text stays legible and visually balanced. The resulting overlay behaves as an explicit attention beacon that channels the model's focus along spatial semantics. On the VMCBench suite, Timage paired with a modest 7B backbone clearly overtakes far larger proprietary systems as well as parameter-tuned baselines. The study positions deliberate input reconstruction as a powerful, architecture-neutral lever for strengthening multimodal reasoning.

阅读与讨论 → 访问原文 →