探索全球前沿学术脉络

01.

medRxiv (Medicine) 2026-06-10 DOI: HASH:22b69db888d8ce50e43123b7b3317f6b

Documented clinical genetic testing among carriers of hereditary breast and ovarian cancer variants: Ancestry and socioeconomic disparities in the All of Us research program

作者:

Yerukala Sathipati↗Scott↗

Importance: Hereditary breast and ovarian cancer (HBOC) variant carriers benefit from risk-reducing interventions, but only if identified. The extent to which carriers are clinically recognized, and whether recognition is equitable across diverse populations, is poorly characterized in a single large U.S. cohort. Objective: To estimate P/LP HBOC carrier prevalence across genetic ancestry groups, quantify documented clinical genetic testing among carriers, and evaluate ancestry and socioeconomic disparities in testing. Design, Setting, and Participants: Cross-sectional analysis of the All of Us Research Program Controlled Tier (Curated Data Repository v8/C2024Q3R9), comprising participants with short-read whole genome sequencing and linked electronic health record (EHR) and survey data. Carriers were ascertained from research genomic data independent of clinical testing. Exposures: Genetically inferred ancestry (African [AFR], Admixed American [AMR], East Asian [EAS], European [EUR], Middle Eastern [MID], South Asian [SAS]); self-reported household income and educational attainment. Main Outcomes and Measures: (1) Carrier prevalence with Wilson 95% CIs; (2) documented clinical genetic testing (procedure codes) among carriers; (3) adjusted odds of documented testing among women, by ancestry, before and after socioeconomic adjustment, using multivariable logistic regression. Results: Among 414,830 participants, P/LP HBOC carrier prevalence was 1.42% (95% CI, 1.38-1.45) overall and similar across ancestry groups (AFR 1.24%, AMR 1.32%, EAS 1.19%, EUR 1.52%, MID 1.68%, SAS 1.33%; overlapping CIs). Among 250,071 women in the testing analysis, documented clinical genetic testing was rare: only 74 of 5,878 carriers overall (1.3%) and 59 of 3,572 European-ancestry carriers (1.7%) had a documented test, with counts below reportable thresholds in all other ancestry groups. African-ancestry women had lower adjusted odds of documented testing than European-ancestry women (Model 1 adjusted odds ratio [aOR], 0.32; 95% CI, 0.27-0.39), an association that attenuated but persisted after adjustment for income and education (Model 2 aOR, 0.48; 95% CI, 0.40-0.58; P < 0.001); Admixed American women also had reduced adjusted odds (aOR, 0.71; 95% CI, 0.61-0.84). Lower income and lower education were independently and dose-dependently associated with lower testing odds (income

阅读与讨论 → 访问原文 →

02.

arXiv (math.PR) 2026-06-11 DOI: arXiv:2411.08653

Hilbert space embeddings of independence tests and interaction measures of several variables

作者:

Jean Carlo Guella↗

arXiv:2411.08653v2 Announce Type: replace-cross Abstract: We present a unified theoretical framework for kernel-based measures of dependence on product spaces. Building on the ideas underlying distance covariance, distance multivariance, and the Hilbert-Schmidt Independence Criterion (HSIC), we define a new family of kernels on an $n$-fold Cartesian product, termed positive definite independent of order $k$ (PDI$_{k}$ kernels). These kernels extend the concepts of positive definite and conditionally negative definite kernels to higher orders and provide the foundation for generalized independence and interaction tests, such as the generalized Lancaster interaction of order $k$ ($\Lambda_{k}^{n}$), and the Streitberg interaction ($\Sigma$). Our analysis focuses on the continuous setting, where we prove a Kernel Mean Embedding Theorem for PDI$_{k}$ kernels and establish the corresponding integrability restrictions. Based on these results, we characterize how the Kronecker products of PDI kernels behave.

阅读与讨论 → 访问原文 →

03.

arXiv (CS.AI) 2026-06-15 DOI: arXiv:2606.08881

Benchmarking Vision-Language-Action Models on SO-101: Failure and Recovery Analysis

作者:

Yi Yu↗Xinchuan Qiu↗

arXiv:2606.08881v2 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models have demonstrated strong generalization in robotic manipulation, yet existing evaluations are primarily conducted in simulation or on expensive robotic platforms, leaving their robustness on affordable real-world robots largely unexplored. We present a standardized real-world benchmark for evaluating representative VLA and imitation learning policies on the low-cost SO-101 robotic platform. The benchmark comprises four representative manipulation tasks together with unified evaluation protocols, enabling systematic comparison under embodiment uncertainty. Using real-world teleoperated demonstrations, we fine-tune and evaluate $\pi_{0.5}$, SmolVLA, Wall-X, and ACT directly on the physical platform. Beyond conventional task success rates, the benchmark incorporates a structured failure taxonomy, semantic- and execution-level failure decomposition, and recovery-aware evaluation metrics to characterize policy robustness. Experimental results show that stronger pretrained VLA policies generally outperform the imitation learning baseline, although performance remains highly task-dependent under low-cost robotic deployment conditions. Execution instability emerges as the dominant failure source, while recovery capability varies substantially across architectures. These results highlight the importance of failure and recovery analysis beyond binary task success and establish SO-101 as a practical benchmark for evaluating embodied AI systems under realistic low-cost robotic deployment conditions.

阅读与讨论 → 访问原文 →

04.

arXiv (CS.CL) 2026-06-15 DOI: arXiv:2602.04879

Rethinking the Trust Region in LLM Reinforcement Learning

作者:

Penghui Qi↗Xiangxin Zhou↗Zichen Liu↗Tianyu Pang↗Chao Du↗Min Lin↗Wee Sun Lee↗

Reinforcement learning (RL) has become a cornerstone for fine-tuning Large Language Models (LLMs), with Proximal Policy Optimization (PPO) serving as the de facto standard algorithm. Despite its ubiquity, we argue that the core ratio clipping mechanism in PPO is structurally ill-suited for the large vocabularies inherent to LLMs. PPO constrains policy updates based on the probability ratio of sampled tokens, which serves as a noisy single-sample Monte Carlo estimate of the true policy divergence. This creates a sub-optimal learning dynamic: updates to low-probability tokens are aggressively over-penalized, while potentially catastrophic shifts in high-probability tokens are under-constrained, leading to training inefficiency and instability. To address this, we propose Divergence Proximal Policy Optimization (DPPO), which substitutes heuristic clipping with a more principled constraint based on a direct estimate of policy divergence (e.g., Total Variation or KL). To avoid huge memory footprint, we introduce the efficient Binary and Top-K approximations to capture the essential divergence with negligible overhead. Extensive empirical evaluations demonstrate that DPPO achieves superior training stability and efficiency compared to existing methods, offering a more robust foundation for RL-based LLM fine-tuning. Our code is available at https://github.com/sail-sg/Stable-RL.

阅读与讨论 → 访问原文 →

05.

arXiv (CS.CL) 2026-06-24 DOI: arXiv:2606.23695

Quantifying Prior Dominance in RAG Systems

作者:

Barak Or↗

Retrieval-Augmented Generation (RAG) grounds Large Language Models in external knowledge, yet current evaluations rely on discrete heuristics that suffer from ''epistemic blindness'' - failing to distinguish genuine contextual information extraction from parametric memory recall. To address this, we introduce the Normalized Context Utilization (NCU) metric, leveraging continuous token log-probabilities across zero-shot, oracle, and adversarial conditions to strictly quantify contextual information gain. Evaluating architectures ranging from 1.5B to 72B parameters alongside a proprietary commercial API reveals that for strict factual extraction (without Chain-of-Thought reasoning), traditional scaling laws exhibit extreme diminishing returns: highly efficient Small Language Models (SLMs) match or outperform high-capacity architectures. Furthermore, we demonstrate that ``Prior Dominance'' correlates with model scale and proprietary alignments. The evaluated commercial API not only overrode explicit external evidence in nearly half of adversarial conflicts, but also frequently suffered from systemic confidence collapse (Negative Transfer) when its parametric priors were contradicted. Our findings highlight the structural epistemic advantage and superior contextual adherence of SLMs in strict extraction workflows.

阅读与讨论 → 访问原文 →

06.

arXiv (CS.AI) 2026-06-24 DOI: arXiv:2606.23742

Low-power analogue neural networks with trainable nonlinear connections for continuous control

作者:

Ian T. Vidamour↗Fernando Aguirre↗Thomas J. Hayward↗Matthew O. A. Ellis↗Charles Swindells↗Alexander McDonnell↗Martin Trefzer↗Finley Robins↗Luca Manneschi↗Susan Stepney↗Tony Kenyon↗Oliver J. Sutton↗…

arXiv:2606.23742v1 Announce Type: cross Abstract: Physical neural networks promise low-power machine learning by computing directly with analogue device physics, but most architectures force nonlinear device responses to act as scalar weights. Inspired by Kolmogorov-Arnold networks, we place trainable nonlinear functions on the connections, making each physical connection a learnable computational element. Realising these functions as analogue band-pass filters on field-programmable analogue arrays, we find that the benefit is task-dependent and follows from the smoothness of the physical basis: the networks represent smooth, continuously valued targets, including robotic kinematics, continuous control, and photovoltaic maximum-power-point tracking, with far fewer nodes and connections than multilayer perceptrons, but offer no parameter-efficiency advantage on classification-like decision boundaries. Trained networks transfer to hardware across approximately 35,000 connections with quantified fidelity, and a dedicated CMOS implementation is projected to operate at approximately 30 microwatts. A memristive realisation reproduces the same behaviour in simulation, indicating that the advantage comes from placing trainable nonlinearity on connections, rather than from a particular device.

阅读与讨论 → 访问原文 →

07.

arXiv (CS.LG) 2026-06-11 DOI: arXiv:2601.21817

A Judge-Aware Ranking Framework for Evaluating Large Language Models without Ground Truth

作者:

Mingyuan Xu↗Xinzi Tan↗Jiawei Wu↗Doudou Zhou↗

arXiv:2601.21817v3 Announce Type: replace-cross Abstract: Evaluating large language models (LLMs) on open-ended tasks without ground-truth labels is increasingly done via the LLM-as-a-judge paradigm. A critical but under-modeled issue is that judge LLMs differ substantially in reliability; treating all judges equally can yield biased leaderboards and misleading uncertainty estimates. More data can make evaluation more confidently wrong under misspecified aggregation. We propose a judge-aware ranking framework that extends the Bradley-Terry-Luce model by introducing judge-specific discrimination parameters, jointly estimating latent model quality and judge reliability from pairwise comparisons without reference labels. We establish identifiability up to natural normalizations and prove consistency and asymptotic normality of the maximum likelihood estimator, enabling confidence intervals for score differences and rank comparisons. Across multiple public benchmarks and a newly collected dataset, our method improves agreement with human preferences, achieves higher data efficiency than unweighted baselines, and produces calibrated uncertainty quantification for LLM rankings.

阅读与讨论 → 访问原文 →

08.

arXiv (CS.CL) 2026-06-24 DOI: arXiv:2601.11957

PEARL: Self-Evolving Assistant for Time Management with Reinforcement Learning

作者:

Bingxuan Li↗Jeonghwan Kim↗Cheng Qian↗Xiusi Chen↗Eitan Anzenberg↗Niran Kundapur↗Heng Ji↗

Overlapping calendar invitations force busy professionals to repeatedly decide which meetings to attend, reschedule, or decline. We refer to this preference-driven decision process as calendar conflict resolution. Automating this decision process is crucial yet challenging. Scheduling logistics can drain hours, and human delegation often fails at scale, which motivates us to ask: Can we trust large language models (LLMs) or language agents to manage time? To enable a systematic study of this question, we introduce CalConflictBench, a benchmark for long-horizon calendar conflict resolution. In CalConflictBench, conflicts are presented to agents round-by-round over a calendar year, requiring them to infer and adapt to user preferences progressively. Our experiments show that current LLM agents perform poorly with high error rates, e.g., Qwen-3-30B-Think has an average error rate of 35%. To address this gap, we propose PEARL, a reinforcement-learning framework that (i) augments the language agent with an external preference memory that stores and updates inferred strategies (e.g., attendee priorities, topic importance, time/location preferences), and (ii) optimizes the agent with round-wise rewards that directly supervise decision correctness, ranking quality, and memory usage across rounds. Experiments on CalConflictBench show that PEARL achieves an error reduction rate of 0.76 and a 55% improvement in average error rate compared to the strongest baseline.

阅读与讨论 → 访问原文 →

09.

arXiv (CS.AI) 2026-06-24 DOI: arXiv:2606.24388

PHANTOM: A Large-Scale Dataset of Multimodal Adversarial Attacks for Vision-Language Models

作者:

Simone Gallivanone↗Hossein Khodadadi↗Mauro Dore↗Mauro Medda↗Nicola Franco↗

arXiv:2606.24388v1 Announce Type: new Abstract: We introduce a large-scale, open-source dataset of pre-generated adversarial attacks for vision-language models (VLMs). The dataset is designed to be diverse, representative, and practical, extending existing benchmarks by covering 10 high-level categories and 55 subcategories of harmful intents. Our primary goal is to make adversarial data accessible to the research community, given the computational cost and complexity of generating large numbers of attacks. The dataset comprises 47 524 adversarial samples, generated using state-of-the-art attack strategies from recent literature. Our work complements existing efforts by consolidating and extending prior benchmarks from multiple established sources, resulting in 7 826 intents, and introduce an additional category to broaden coverage. This provides realistic evaluation resources for studying model robustness and alignment. Our dataset intends to enable researchers and practitioners to systematically evaluate the robustness and safety of VLMs, fine-tune attack-generation models, and develop or stress-test defensive guardrails under diverse adversarial conditions. By releasing this resource, we aim to lower the barrier to adversarial research and foster more reproducible, comprehensive, and comparable evaluations of VLM safety.

阅读与讨论 → 访问原文 →

10.

arXiv (CS.AI) 2026-06-19 DOI: arXiv:2606.19613

StaminaBench: Stress-Testing Coding Agents over 100 Interaction Turns

作者:

Vlad Sobal↗Shuo Yang↗Yuting Zhang↗Wei Xia↗Stefano Soatto↗

arXiv:2606.19613v1 Announce Type: cross Abstract: We introduce StaminaBench, a benchmark that measures the stamina of coding agents: how many consecutive interaction turns (change requests) they can handle before failing. Unlike the prevailing fraction-of-tasks-solved metric, this matches real vibe-coding where sessions run dozens or hundreds of turns. In StaminaBench, agents implement a REST API server and modify it across a tunable number of procedurally generated follow-up change requests - 100 in our experiments, resulting in codebases of up to 6,000 lines. Tests are generated fully programmatically without LLM involvement, ensuring reproducibility and reliability; change sequences are drawn from either a hardcoded or LLM-driven sampler, both constrained to a structured action space to ensure changes are valid. The agent and the server run in an isolated environment and communicate with the benchmark through HTTP, making testing fully black-box and language-agnostic. We evaluate six agent harnesses paired with seven open-source LLMs across 20 scenarios of 100 turns each and find that: (1) all the tested models fail within 5-6 turns, confirming that vibe-coding-style programming without thorough testing produces bugs; (2) passing test feedback back to the agent and allowing it to retry improves passed turn count by up to 12x; and (3) a good harness is required for strong performance: stronger models exhibit up to a 6x gap between their best and worst harness, while weaker models fail with any harness. We release the benchmark and the generated tasks to enable further research into multi-turn coding agent behavior. Benchmark code and data: github.com/amazon-science/StaminaBench.

阅读与讨论 → 访问原文 →

11.

arXiv (CS.CV) 2026-06-12 DOI: arXiv:2606.13676

Modality Forcing for Scalable Spatial Generation

作者:

Bardienus Pieter Duisterhof↗Deva Ramanan↗Jeffrey Ichnowski↗Justin Johnson↗Keunhong Park↗

Text-to-image (T2I) models contain rich spatial priors. Synthesizing photorealistic, cluttered scenes requires an understanding of geometry, including perspective and relative scale. Prior works adapt T2I models to leverage this prior for depth prediction, but they require dense depth data and involve complex recipes. We propose Modality Forcing, a simple, scalable post-training recipe for joint image-depth generation using a single DiT trained on sparse depth data. Modality Forcing enables conditional and joint generation of image and depth in any permutation by assigning separate noise levels per modality. Per-modality decoders let us train on sparse, real-world depth and achieve strong, generalizable depth prediction. We further show that Modality Forcing inherits the scalability of T2I pre-training: by training a set of T2I models from scratch (370M to 3.3B parameters), we find that larger models trained on more image data produce more accurate depth. Our strongest model is competitive with state-of-the-art monocular depth estimators and reduces AbsRel by 57% relative to existing joint image-depth generative models. These results provide strong evidence that image generation is a scalable pre-training objective for spatial perception. https://modality-forcing.github.io/

阅读与讨论 → 访问原文 →

12.

arXiv (quant-ph) 2026-06-16 DOI: arXiv:2604.25953

Entanglement as a Witness of Quantum Coherence: A Bipartite Monty-Hall Protocol

作者:

Jorge Meza-Dom\'inguez↗

arXiv:2604.25953v3 Announce Type: replace Abstract: We present a bipartite protocol inspired by the Monty Hall puzzle that operationally distinguishes quantum coherence from classical ignorance. A principal qutrit is entangled with an ancillary qutrit via a controlled unitary, preparing $|\Psi\rangle = \frac{1}{\sqrt{3}}(|A,0\rangle + |B,1\rangle + |C,2\rangle)$. A rank-1 projective discard then eliminates one basis state, leaving a coherent superposition of the two remaining states. Finally, the ancilla and qutrit are measured, yielding joint probabilities that encode the interplay between superposition and measurement back-action. We show that the conditional probability $P(B|anc=0)$ takes the value $1/4$ in both quantum mechanics and the classical ignorant-host model, making it unsuitable as a witness. The true quantum-classical separation emerges in conditional joint probabilities that correlate ancilla outcomes with specific discard operations. We define witnesses $\mathcal{W}_{i,j} = P(anc=i, qutrit=j \mid discard k)$ where $j$ differs from the ancilla-implied state. Quantum mechanics predicts $\mathcal{W} = 1/4$, while any classical epistemic model with perfect initial correlations yields $\mathcal{W} = 0$. We provide the explicit $9 \times 9$ unitary matrix, a complete analysis of all measurement outcomes, and a detailed proof of the violation. The witness is fully immune to white noise and robust against moderate dephasing. The protocol requires only a single pair of entangled qutrits and sequential measurements – no spatial separation, no multiple copies, and no complex sets of incompatible observables. This makes it suitable for advanced undergraduate laboratories and provides a pedagogically accessible test of the ontic-epistemic distinction in quantum foundations.

阅读与讨论 → 访问原文 →

13.

arXiv (CS.CV) 2026-06-16 DOI: arXiv:2509.25594

K-Prism: A Knowledge-Guided and Prompt Integrated Universal Medical Image Segmentation Model

作者:

Bangwei Guo↗Yunhe Gao↗Meng Ye↗Difei Gu↗Yang Zhou↗Leon Axel↗Dimitris Metaxas↗

Medical image segmentation is fundamental to clinical decision-making, yet existing models remain fragmented. They are usually trained on single knowledge sources and specific to individual tasks, modalities, or organs. This fragmentation contrasts sharply with clinical practice, where experts seamlessly integrate diverse knowledge: anatomical priors from training, exemplar-based reasoning from reference cases, and iterative refinement through real-time interaction. We present $K-Prism$, a unified segmentation framework that mirrors this clinical flexibility by systematically integrating three knowledge paradigms: (i) $semantic priors$ learned from annotated datasets, (ii) $in-context knowledge$ from few-shot reference examples, and (iii) $interactive feedback$ from user inputs like clicks or scribbles. Our key insight is that these heterogeneous knowledge sources can be encoded into a dual-prompt representation: 1-D sparse prompts defining $what$ to segment and 2-D dense prompts indicating $where$ to attend, which are then dynamically routed through a Mixture-of-Experts (MoE) decoder. This design enables flexible switching between paradigms and joint training across diverse tasks without architectural modifications. Comprehensive experiments on 18 public datasets spanning diverse modalities (CT, MRI, X-ray, pathology, ultrasound, etc.) demonstrate that K-Prism achieves state-of-the-art performance across semantic, in-context, and interactive segmentation settings.

阅读与讨论 → 访问原文 →

14.

arXiv (CS.CL) 2026-06-24 DOI: arXiv:2606.24453

Bayesian control for coding agents

作者:

Theodore Papamarkou↗Vladislav Smirnov↗Viktor Mazanov↗Artem Vazhentsev↗Preslav Nakov↗Timothy Baldwin↗Artem Shelmanov↗

Modern coding agents pair LLM generators with various tools, including cheap diagnostics and expensive verifiers. The tool-use decisions are typically governed by orchestrators that often use fixed rules and ignore uncertainty. We formulate orchestration as cost-sensitive sequential hypothesis testing: a Bayesian controller maintains a belief over candidate correctness and dynamically decides whether to gather more evidence, refine the candidate, verify it, or stop. Across six generators and nine coding benchmarks, Bayesian control proves to be most valuable when verification is costly and critics are informative but imperfect. Beyond control, the belief state yields an interpretable correctness score that outperforms token-probability and raw tool-success baselines for uncertainty quantification.

阅读与讨论 → 访问原文 →

15.

arXiv (CS.CV) 2026-06-17 DOI: arXiv:2606.17386

TerraTransfer: Learning End-to-End Driving Policies Without Expert Demonstrations

作者:

Zikang Xiong↗Weixin Li↗Zhouchonghao Wu↗Akshay Rangesh↗Saarth Bonde↗Grantland Hall↗Chen Tang↗Yihan Hu↗Wei Zhan↗

End-to-end autonomous driving has achieved state-of-the-art performance on benchmarks and real-world deployments. Its standard training recipe, however, is expensive across all stages: collecting and labeling millions of driving frames is costly, and closed-loop RL on images is bottlenecked by the per-step cost of photorealistic rendering plus a forward pass through a large vision backbone. Self-play in vectorized simulators changes the economics: millions of rollout steps per second, and a state distribution naturally rich in collisions, near-misses, and recoveries that no driving log contains. Our approach exploits this asymmetry by decoupling learning to drive from learning to see. We pretrain a single policy by self-play, then align its latent space with a pretrained vision backbone, through the action KL divergence and a batch-relational low-rank structural loss. The action target comes from the self-play policy, so alignment never supervises against a logged trajectory: a paired dataset of (image, scene-state) frames suffices, with no need for the curated expert demonstrations that imitation pretraining is built on. On photorealistic 3D Gaussian splatting closed-loop scenarios, the resulting end-to-end policy matches or exceeds prior end-to-end methods.

阅读与讨论 → 访问原文 →

16.

arXiv (CS.LG) 2026-06-19 DOI: arXiv:2606.19367

Weibull Weight-Scale Parameter Evolution under AdamW Training Dynamics

作者:

Tiexin Ding↗

arXiv:2606.19367v1 Announce Type: new Abstract: Building on a two-parameter Weibull framework for diagnosing transformer weight distributions, we study why the Weibull weight-scale parameter $\lambda$ grows, overshoots, and then relaxes during AdamW training. We derive a leading-order three-force decomposition of the squared weight norm from the AdamW update: an alignment force measuring the correlation between weights and the adaptive update direction, an injection force from adaptive step magnitude, and a decay force from decoupled weight decay. On self-trained Pythia-70M models with ground-truth optimizer moments, alignment dominates the rise phase, contributing 88-94% of the absolute force budget across four random seeds and remaining robust to super-weight removal. Near saturation, alignment and decay approach balance, explaining the transition from weight-scale growth to relaxation. These force dynamics directly govern the squared-norm component underlying $\lambda(t)$; the remaining RMS-to-Weibull reconstruction offset is measurable and decomposes into bridge and integration components, totaling approximately 5-6% in densely sampled regions. To extend the analysis to real models where optimizer moments are unavailable, we introduce a spline displacement method that recovers the alignment force from sparse checkpoints with approximately 92-94% accuracy, about twice the naive two-point baseline. We further observe that the peak value of $\lambda(t)$ varies with training-data coherence in our experiments, suggesting a data-dependent component of weight-scale growth that we leave to a controlled follow-up study. Code and data are available at https://github.com/tiexinding/NPM-Weibull-public.

阅读与讨论 → 访问原文 →

17.

arXiv (CS.CL) 2026-06-17 DOI: arXiv:2606.17229

Rift: A Conflict Signature for Deception in Language Models

作者:

Petr Nyoma↗

A model that lies while knowing the truth is the central case ELK cannot handle with behavioral evaluation alone. We ask whether such deception leaves an internal signature distinguishing it from honest error. Our key move is a control for wrongness: we contrast a sleeper agent (knows the truth, lies on trigger) against a naive liar (fine-tuned to emit the same wrong answers with no honest training). Both produce identical wrong outputs; any difference is about knowledge conflict, not incorrectness. We find deceptive forward passes carry a conflict signature - 2.1-2.3x higher residual rank than naive-liar passes on the same wrong answer - strong enough to identify which of two responses is the lie with 100% accuracy and no labels, across GPT-2 small/medium (three seeds) and three instruct models. Across Qwen2.5-1.5B/7B and Phi-3-mini, instructed deception raises residual rank on every tested fact (18/18, 40/40, 34/34); on Phi-3, lies separate perfectly from both honest answers and hallucinations (AUC 1.0, Wilcoxon p~6e-11). The signature survives strategic self-constructed deception (model invents its own lie, AUC 1.0), active concealment attempts (AUC 1.0), and length-controlled replication (20/20, AUC 1.0, p~1e-6). Using basis-free relative representations, a probe trained on one model family detects deception in two other families zero-shot (mean AUC 0.933), surviving simultaneous architecture and format change (AUC 0.821), and transfers across five languages (AUC 1.000, length-controlled). The signature is read-only: detectable but not injectable (0/8 both directions). Honest limitations and six negative experiments are documented in full.

阅读与讨论 → 访问原文 →

18.

arXiv (math.PR) 2026-06-12 DOI: arXiv:2606.13313

Sphere Packings in Higher Dimension (after Boaz Klartag)

作者:

Guillaume Aubrun↗

arXiv:2606.13313v1 Announce Type: cross Abstract: Let $\delta_n^L$ be the maximal density of a lattice sphere packing in the $n$-dimensional Euclidean space. We explain how Boaz Klartag proved the inequality $\delta_n^L \geq c n^2 2^{-n}$ where $c>0$ is a universal constant. In higher dimension, even for non-lattice sphere packings, this new lower bound is a substantial improvement. Klartag's proof uses the probabilistic method in two different ways. The first, very standard, relies on the statistical properties of a uniformly chosen random lattice. The second, completely new, studies the stochastic evolution of an ellipsoid constrained to contain non nonzero lattice points in the interior.

阅读与讨论 → 访问原文 →

19.

medRxiv (Medicine) 2026-06-15 DOI: HASH:f66d9805092bf802a75698dfcd5dc4ec

SPIRIT-CONSORT-ELM: Element-Level Assessment of Randomized Controlled Trial Reporting Using Large Language Models

作者:

Jiang↗Ying↗Brown↗A. W↗Lan↗Song↗Menke↗Vorland↗Mayo-Wilson↗Kilicoglu↗

Randomized controlled trials (RCTs) play a central role in assessing the benefits and harms of interventions. Incomplete reporting in RCT publications can compromise the verifiability and usefulness of RCTs. SPIRIT and CONSORT reporting guidelines aim to improve the completeness of RCT protocols and results publications, respectively. However, many RCTs are not reported completely. Checking manuscripts automatically could help authors improve the completeness of reports prior to publication. We previously annotated SPIRIT-CONSORT-TM, a corpus of 200 articles (comprising 100 protocol-results publication pairs) using 83 checklist items drawn from SPIRIT 2013 and CONSORT 2010. We also trained machine learning models to automatically assess reporting at the item level. Each checklist item can include multiple constituent elements (i.e., specific details required for that item), and an item might be considered fully reported when all of its elements are present. However, prior work does not explicitly capture or evaluate reporting at the element level. To address this gap, we extended SPIRIT-CONSORT-TM by incorporating element-level annotations and using them to assess reporting completeness (SPIRIT-CONSORT-ELM). We formulated element-level assessment as a machine reading comprehension task, operationalized through 119 questions, where each question targets a specific reporting element within a checklist item. Using the 200 articles included in SPIRIT-CONSORT-TM, two annotators independently answered 119 questions for 50 articles (25 protocol-results pairs) and resolved any discrepancies through discussion; the remaining 150 articles (75 protocol-results pairs) were assessed by a single annotator. We then developed an automated pipeline for element-level assessment using SPIRIT-CONSORT-ELM. The pipeline first applies a PubMedBERT-based model to identify sentences containing item-level reporting information, then it uses a generative large language model (LLM; GPT-5) with chain-of-thought reasoning to answer element-level questions based on the retrieved evidence. Agreement between the two annotators was high (Gwet's AC1: 0.782) and our pipeline achieved high accuracy in identifying element-level reporting evidence (F1: 0.822, Gwet's AC1: 0.796). Ablation studies indicate that chain-of-thought reasoning and the inclusion of illustrative in-context examples modestly improve LLM performance on the machine reading comprehension task. SPIRIT-CONSORT-ELM provides a benchmark for evaluating reporting guideline completeness at the element level, enabling assessment of RCT transparency beyond the simple presence or absence of checklist items and is publicly available at https://osf.io/kznx4/. The automated pipeline establishes a robust baseline for assessing RCT reporting and demonstrates potential as a practical aid for authors, reviewers, and editors to identify and address gaps in completeness and transparency of RCT reports.

阅读与讨论 → 访问原文 →

20.

arXiv (CS.AI) 2026-06-17 DOI: arXiv:2606.17122

TrustErase: Auditable Instant Machine Unlearning with Passport-Embedded Representations

作者:

Rutger Hendrix↗Leonardo G. Russo↗Concetto Spampinato↗Matteo Pennisi↗Giovanni Bellitto↗

arXiv:2606.17122v1 Announce Type: cross Abstract: The demand for privacy-compliant AI has amplified the need for machine unlearning; yet, existing retraining or distillation-based methods remain unverifiable and computationally costly. We introduce TrustErase, a verifiable, data-free unlearning framework leveraging passport-embedded representations for instant, modular, and auditable forgetting. By treating passports as cryptographic keys within parameter-efficient adaptation layers, TrustErase enables the removal of specific classes or datasets through simple deactivation, without retraining, fine-tuning, or access to the original data. A singular value based decomposition conceals passports within model weights, ensuring that unlearning actions remain transparent and provably compliant. Evaluations on MNIST, CIFAR10 and CIFAR100 show that TrustErase matches or exceeds state-of-the-art benchmarks such as DELETE, L2UL, and Boundary Shrink, while operating in a strictly data-free regime. Ultimately, TrustErase establishes a new paradigm for trustworthy, accountable, and instantly forgettable AI systems.

阅读与讨论 → 访问原文 →

21.

arXiv (CS.CL) 2026-06-16 DOI: arXiv:2606.15405

T-Mem: Memory That Anticipates, Not Archives

作者:

Weidong Guo↗Dakai Wang↗Zixuan Wang↗Hui Liu↗Yu Xu↗

Long-term memory is essential for conversational agents to remain coherent across extended dialogues, follow through on commitments made many sessions earlier, and adapt their behaviour to each user. Current LLM-backed long-term conversational memory, however, is reachability-bounded by the similarity between a query and stored content, both lexical and dense-vector. The approach is effective when query and memory share surface features such as wording or named entities (we call this descriptive). But it misses another, equally valuable class of cases, where query and memory do not share surface features and are tied only by a latent semantic arc (associative). On this regime prevailing long-term memory systems collectively fail. Covering this other half is what allows an assistant, for the first time, to actively draw on past dialogue as a semantic asset. On the memory side, this is the engineering counterpart of what cognitive science calls episodic future thinking: rehearsing past experience for the future contexts under which it will need to be found. We call these write-time rehearsals triggers. We propose T-Mem, the first long-term conversational memory architecture that covers both descriptive and associative recall. At each of two evidence granularities, single facts and full exchanges, T-Mem instantiates one descriptive trigger family and one associative trigger family, so that every memory remains reachable from both surface-similar and relevance-bound queries. As empirical validation, T-Mem reaches state-of-the-art on both LoCoMo and LoCoMo-Plus.

阅读与讨论 → 访问原文 →

22.

arXiv (CS.LG) 2026-06-11 DOI: arXiv:2606.11982

PAWS: Preference Learning with Advantage-Weighted Segments

作者:

Aleksandar Taranovic↗Onur Celik↗Niklas Freymuth↗Ge Li↗Serge Thilges↗Huy Le↗Tai Hoang↗Rania Rayyes↗Gerhard Neumann↗

arXiv:2606.11982v1 Announce Type: new Abstract: Preference-based reinforcement learning (PbRL) learns policies from human trajectory-level comparisons, avoiding explicit reward design and expert demonstrations. Existing methods typically train utility functions on trajectory or segment-level preferences while relying on per-step utility estimates during policy optimization. This training and inference mismatch induces a distribution shift that severely degrades temporal credit assignment and limits policy learning. We analyze this issue and propose PAWS, a segment-based preference learning method that performs policy updates directly using segment-level advantage functions. By aligning utility training with policy optimization, PAWS preserves trajectory-level preference information and avoids unreliable per-step learning signals. Experiments on simulated robotic manipulation and locomotion tasks demonstrate that PAWS consistently outperforms existing PbRL approaches, highlighting the importance of distribution-consistent preference learning.

阅读与讨论 → 访问原文 →

23.

arXiv (CS.CL) 2026-06-11 DOI: arXiv:2606.11202

One Jailbreak, Many Tongues: Learning Language-Insensitive Intention Representations for Multilingual Jailbreak Detection

作者:

Shuyu Jiang↗Kaiyu Xu↗Xingshu Chen↗Hao Ren↗Rui Tang↗Yi Zhang↗Tianwei Zhang↗Hongwei Li↗

Large language models (LLMs) are increasingly deployed in applications for global multilingual users, yet safety training remains concentrated in dominant languages and has not progressed in parallel with multilingual capability, creating exploitable gaps for jailbreak attacks. Current jailbreak defenses are largely developed and evaluated in dominant languages, and their effectiveness is limited by the scarcity of aligned multilingual supervision and representations dispersion caused by language variation. To address this issue, we propose MLJailDe, a multilingual jailbreak detection framework designed to improve both multilingual robustness and cross-lingual generalization. MLJailDe first introduces a multilingual back-translation data augmentation algorithm to construct a semantically consistent and functionally effective dataset spanning 11 languages, consisting of 2,232 benign and 1,239 jailbreak samples. On this basis, MLJailDe employs relative-distance constraints to reduce cross-lingual representation dispersion and encourage jailbreak prompts with similar intent to form consistent clusters across languages, while an imbalance-aware classification objective is further used to alleviate class imbalance and learn more reliable multilingual decision boundaries. Experimental results show that MLJailDe outperforms state-of-the-art baselines across multiple languages, achieving an F1 score of 98.5\%, and obtains an average F1 score of 97.1\% on unseen languages, demonstrating strong effectiveness and cross-lingual generalization.

阅读与讨论 → 访问原文 →

24.

Nature (Science) 2026-06-10 DOI: HASH:978a9ee7f0f72bbb0f019e8b69731468

Two-component exciton condensates in an electron–hole bilayer

作者:

Ruishi Qi↗

Macroscopic quantum coherence emerges when bosons condense into a Bose–Einstein condensate (BEC)1–5. Excitons are a long-sought solid-state route to high-temperature BECs with strong interactions, electrical tunability and potentially multicomponent spinor order, but conclusive evidence for equilibrium condensation has remained elusive. Here we report evidence for two-component exciton BECs in MoSe2/hBN/WSe2 electron–hole bilayers6–9 by probing the spin–valley susceptibility of constituent electrons and holes. This heterostructure hosts equilibrium exciton fluids with four spin–valley flavours. Magneto-optical spectroscopy in a dilution refrigerator reveals three exciton condensate phases with distinct flavour polarizations. At zero magnetic field, the many-body ground state is a coherent superposition of two condensed intravalley exciton flavours. Under a magnetic field, the intravalley exciton condensate first switches to a two-component intervalley condensate through a first-order quantum phase transition at a weak critical field and then turns into a fully polarized single-component condensate at high fields. The condensate signatures form a dome in density–temperature space, persisting up to approximately 1.8 K. Our results establish van der Waals electron–hole bilayers as a versatile platform for strongly interacting, multicomponent exciton BECs. Macroscopic quantum coherence arises in two-component exciton Bose–Einstein condensates within MoSe2/hBN/WSe2 electron–hole bilayers, exhibiting distinct spin–valley polarized phases, quantum phase transitions under magnetic fields and stable condensate behaviour up to approximately 1.8 K.

阅读与讨论 → 访问原文 →

25.

arXiv (CS.CL) 2026-06-15 DOI: arXiv:2604.24942

Independent-Component-Based Encoding Models of Brain Activity During Story Comprehension

作者:

Kamya Hari↗Taha Binhuraib↗Jin Li↗Cory Shain↗Anna A. Ivanova↗

Encoding models provide a powerful framework for linking continuous stimulus features to neural activity; however, traditional voxelwise approaches are limited by measurement noise, inter-subject variability, and redundancy arising from spatially correlated voxels encoding overlapping neural signals. Here, we propose an independent component (IC)-based encoding framework that dissociates stimulus-driven and noise-driven signals in fMRI data. We decompose continuous fMRI data from naturalistic story listening into ICs using one subset of the data, and train encoding models on independent data to predict IC time series from large language model representations of linguistic input. Across subjects, a subset of ICs exhibited consistently high predictivity. These ICs were spatially and temporally consistent across subjects and included cognitive networks known to respond during story listening (auditory and language). Auditory component time series were strongly correlated with acoustic stimulus features, highlighting the interpretability of identified component time series. Components identified as noise or motion-related artifacts by ICA-AROMA showed uniformly poor predictive performance, confirming that highly predicted components reflect genuine stimulus-related neural signals rather than confounds. Overall, IC-based encoding models enable analyses at the level of functional networks, accommodating the variability in network locations across individuals and providing interpretable results that are easy to compare across subjects. Code provided at: https://github.com/kamyahari/IC-Encoding-Models.git

阅读与讨论 → 访问原文 →