中 EN
登录 注册账户

Academic Intelligence · Curated Daily

探索全球前沿学术脉络

AcademicHub 汇聚顶级期刊与预印本平台的实时文献。定制您的专属科研雷达，利用大语言模型自动生成交叉领域文献分析简报。

登录 注册账户
01.
medRxiv (Medicine) 2026-06-15

Association of Genetic Liability to Psychiatric Disorders with Peripheral Metabolic Dysregulation

作者:
De la Hoz↗J. F↗Lee↗Y. H↗Tubbs↗J. D↗Meyerson↗Cudic↗Watts↗Feng↗Y.-C. A↗Chen↗…

Importance: Individuals with psychiatric disorders face elevated cardiometabolic risk which is linked to increased mortality. The extent to which this reflects shared pathogenesis or the downstream effects of illness and treatment remains poorly understood. Objective: To characterize the direct pleiotropic effects of psychiatric genetic liability on circulating metabolites and aggregate cardiometabolic risk, independent of psychiatric diagnosis and psychotropic medication use. Design: Cohort study. Setting: Mass General Brigham Biobank (MGBB). Participants: MGBB participants with metabolomic profiling, genomic data, and linked electronic health records. Exposures: Genetic liability to nine psychiatric disorders quantified using polygenic risk scores (PRS): attention deficit/hyperactivity disorder (ADHD), anorexia nervosa (ANO), anxiety disorder (ANX), autism spectrum disorder (ASD), bipolar disorder (BD), major depressive disorder (MDD), PTSD, schizophrenia (SCZ), and substance use disorder (SUD). Main Outcomes and Measures: 249 circulating metabolites and four metabolomic risk scores (MRS) for type 2 diabetes, myocardial infarction, ischemic stroke, and vascular dementia. PRS-metabolite associations were estimated using nested models adjusting for lifetime psychiatric diagnosis and psychotropic medication use. Results: Across 25,290 participants, we identified 604 significant PRS-metabolite associations (Bonferroni p< 1.36 x 10-4), of which 89% persisted after adjustment for lifetime diagnosis and medication use, suggesting that the direct genetic effects on metabolism are largely independent of illness or treatment. PRS for MDD, PTSD, and ADHD showed the most extensive dysregulation, with a transdiagnostic pattern of elevated lipids and systemic inflammation, specifically triglycerides ({beta} = 0.04 to 0.05, all p< 4.4 x10-13) and glycoprotein acetyls ({beta} = 0.05, all p< 2.2 x10-16). Notably, PRS for SCZ and BD showed minimal metabolite dysregulation despite having the strongest association with their target diagnoses. PRS for MDD, PTSD, ADHD, and SUD were associated with increased MRS across cardiometabolic conditions ({beta} = 0.03 to 0.08, all p< 2.1 x10-4). Sensitivity analyses controlling for BMI or excluding participants without any psychiatric history (N: 21,305 and 11,150, respectively) showed a similar pattern. Conclusions and Relevance: Psychiatric genetic liability is associated with systemic metabolic dysregulation independent of illness onset or treatment, supporting a partially pleiotropic basis for psychiatric-cardiometabolic comorbidity.

阅读与讨论 → 访问原文 →
02.
arXiv (quant-ph) 2026-06-12

Representation-Induced Symmetry Trapping in Adaptive Variational Quantum Simulations of Multi-Reference Topologies

作者:
Hermawan Kresno Dipojono↗

arXiv:2606.13387v1 Announce Type: new Abstract: Evaluating the trainability of adaptive quantum chemistry algorithms under multi-reference static correlation requires understanding how representation topologies intertwine with molecular geometry. We systematically expose a deep physical dependence on point-group symmetry by evaluating a spin-conserved SUSD operator pool across highly stretched configurations (2 x Re) of asymmetric LiH, symmetric BeH2, and asymmetric H2O. Under asymmetric distortions, the non-local mapping constraints of the Bravyi-Kitaev transformation create an optimization trapping effect–an encodement-locked manifestation of the broader barren plateau crisis. Crucially, by comparing these to the symmetrical stretching baseline of BeH2, we demonstrate that the preservation of point-group symmetry structurally protects the optimization landscape, proving that ansatz symmetry restrictions are necessary but insufficient without accounting for the underlying fermion-to-qubit representation. While current methods rely on numerical pruning to throttle pool sizes, our structural approach establishes that the mapping representation remains a critical factor in maintaining landscape trainability. Furthermore, exploiting structural overlap within our pool, we introduce a covariance-driven, adaptive shot-allocation filter. Diverging from static energy-variance minimization frameworks, our allocation engine operates as a dynamic runtime diagnostic tool. By continuously monitoring the gradient precision threshold epsilon, it aggressively prunes dead symmetry channels and triggers an automated circuit-termination sequence upon detecting representation-induced flat-lined states (dE/dtheta approx 0). This integration of algebraic measurement reuse with topology-aware statistical filtering provides a promising, resource-efficient strategy for executing deep variational algorithms on early fault-tolerant architectures.

阅读与讨论 → 访问原文 →
03.
arXiv (CS.CL) 2026-06-18

TW-LegalBench: Measuring Taiwanese Legal Understanding

作者:
Fei-Yueh Chen↗Chun Huang Lin↗Chan Wei Hsu↗Kuan Hsuan Yeh↗Zih-Ching Chen↗Kuan-Ming Chen↗Patrick Chung-Chia Huang↗

Large language models (LLMs) have shown impressive capabilities across diverse tasks, yet their performance on jurisdiction-specific legal reasoning remains underexplored. We present TW-LegalBench that utilizes Taiwanese legal system's rich official corpus open to the public to fill the gap in evaluating LLMs on Taiwanese law, among common-law benchmarks that focus on English sources and civil-law benchmarks focusing on sources of Simplified Chinese. TW-LegalBench comprises three task types: (1) over 16,000 multiple-choice questions (MCQs) across five years of official examinations in 18 professional domains; (2) 117 open-ended essay questions (OEQs) from examinations for legal professionals with official scoring rubrics; and (3) more than 14,000 legal judgment prediction (LJP) instances covering hundreds of crime categories. We evaluate 13 LLMs using accuracy for MCQs, a decomposed LLM-as-Judge framework based on the scoring rubric points for OEQs, and metrics for sentencing accuracy and statute citation for LJP. Our results reveal that top-performing models exceed the passing threshold for qualified lawyers (passing rate: 11%) but fall short of that for judges and prosecutors (passing rate: 1~2%). For LJP, while models demonstrate reasonable verdict type accuracy and sentence prediction capability, they struggle to cite exact legal articles. These findings highlight that reliable legal text generation remains challenging for LLMs, even though their performance on qualification examinations approaches human level.

阅读与讨论 → 访问原文 →
04.
arXiv (CS.CL) 2026-06-17

Security and Privacy Prompts in the Wild: What Users Ask LLMs and How LLMs Respond

作者:
Hobin Kim↗Xiaoyuan Wu↗Omer Akgul↗Lujo Bauer↗Nicolas Christin↗

Large language models (LLMs) are widely used to fulfill users' information needs; users ask LLMs about the weather, pose educational questions, and consult them for legal assistance. One particularly understudied area is digital security and privacy (S&P), where users may seek LLMs' help on how to secure their online accounts or protect their computers from cyber attacks. To the best of our knowledge, no prior study has collected or analyzed the S&P questions users ask LLMs; prior research on LLM response quality relied on expert-authored S&P misconceptions or FAQs rather than user queries. Drawing from WildChat, a dataset of 3.2M user-LLM conversations collected in the wild, our study identifies 14,727 S&P prompts and categorizes them into nine categories covering a wide range of S&P topics. From the S&P prompts, we sampled 450 and performed a thematic analysis to characterize the S&P questions users ask LLMs. Separate from the thematic analysis, we curated 270 advice-seeking S&P prompts, where users ask for recommendations, guidance, or specific S&P information. We measured LLM response quality and consistency when posing the prompt to LLMs 10 times. We found that commercial LLMs outperform open-weight models (GPT 5.5 provided "good enough" responses on 98% of prompts; Llama 4 on 47%). However, among prompts that received high-quality responses on average, commercial models sometimes produce contradictory responses across runs, risking confusing or misleading users.

阅读与讨论 → 访问原文 →
05.
arXiv (CS.AI) 2026-06-25

Privacy Vulnerabilities of Attention Layers in Tabular Foundation Models and Protection of High-Risk Queries

作者:
T\^ania Carvalho↗Maxime Cordy↗

arXiv:2606.26021v1 Announce Type: cross Abstract: Tabular foundation models are commonly assumed to present limited privacy concerns as they are often pre-trained on large collections of synthetic data. However, these models leverage in-context learning, where sensitive records may be provided directly at inference time as labelled context examples. In this paper, we demonstrate that predictions generated via the attention mechanism leak sufficient information to enable effective Membership Inference Attacks (MIAs). To highlight this vulnerability, we propose AMIA (Attention-based Membership Inference Attack), a shadow-model-free attack that exploits the concentration of transformer attention patterns. Our results show that attention mechanisms reveal strong membership signals, which exceed classical confidence-based attacks, achieving an average gain of 7.7\%, specially in low false-positive regimes. To mitigate this risk, we introduce an inference-time defence inspired by $k$-anonymity principles. This approach reduces the uniqueness of context-key representations without introducing random noise or retraining the model. By targeting only high-risk queries identified through AMIA scores, the defence substantially reduces membership leakage of this attack by an average of 50\% and 25\% against confidence-based attacks, while preserving predictive utility with only 3.9\% performance degradation. Beyond showing that context examples are vulnerable, we further demonstrate that fine-tuning introduces an additional source of privacy risk. In particular, samples whose prediction confidence increases after fine-tuning become more susceptible to MIAs, indicating that fine-tuning can amplify memorisation and expose sensitive training information through confidence shifts.

阅读与讨论 → 访问原文 →
06.
arXiv (CS.AI) 2026-06-25

BrainAgent: A Large Language Model-Driven Multi-Agent Framework for Autonomous Brain Signal Understanding

作者:
Yangxuan Zhou↗Sha Zhao↗Jiquan Wang↗Shijian Li↗Gang Pan↗

arXiv:2606.25400v1 Announce Type: new Abstract: Brain-Computer Interfaces (BCIs) and brain signal understanding are pivotal for clinical health and next-generation interactions. Despite this significance, its widespread adoption in real-world scenarios remains restricted, primarily because current analytical paradigms lack sufficient agentic intelligence. First, existing methodologies impose prohibitive technical barriers, requiring extensive specialized expertise. Second, they remain inherently static and task-specific, failing to execute the complex, long-horizon workflows essential for real-world deployment. To accelerate the democratization of brain signal understanding, we draw inspiration from Large Language Models (LLMs) to introduce BrainAgent, an LLM-driven multi-agent framework designed to ground abstract natural language intent into rigorous, executable, and end-to-end processing pipelines. BrainAgent employs a hierarchical architecture where a central supervisor orchestrates specialized sub-agents for adaptive task decomposition and execution. Furthermore, we establish a comprehensive, systematic benchmark for evaluating agentic systems in brain signal analysis. Empirical results demonstrate that BrainAgent effectively automates complex workflows with superior reliability, marking a paradigm shift toward democratized brain signal understanding.

阅读与讨论 → 访问原文 →
07.
arXiv (CS.AI) 2026-06-19

Evaluating and Enhancing Negation Comprehension in Remote Sensing MLLMs

作者:
Haochen Han↗Jue Wang↗Alex Jinpeng Wang↗Fangming Liu↗

arXiv:2606.20177v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) have demonstrated remarkable success in various Remote Sensing (RS) tasks. However, their ability to comprehend negation remains underexplored, limiting deployment in real-world applications where models must explicitly identify what is false or absent, e.g., emergency responders need to locate non-flooded routes for evacuation. To comprehensively study this limitation, we introduce RS-Neg, the first benchmark to evaluate negation understanding across region-level to scene-level tasks. Specifically, we design an automated data generation pipeline for RS imagery, using LLMs to synthesize diverse negation queries, and introduce a dynamic visual focus module for verification. Our evaluation reveals that advanced RS MLLMs struggle with negation, exhibiting hallucinations and substantial performance degradation. To close this gap, we propose NeFo, a novel test-time learning method that explicitly incorporates the logical role of negation into the model optimization. Remarkably, using about 5\% unlabeled test samples, NeFo significantly improves the negation understanding of models and shows strong generalization to unseen tasks. Code and data will be released upon acceptance.

阅读与讨论 → 访问原文 →
08.
arXiv (CS.CV) 2026-06-17

Pulling The REINS: Training-Free Safety Alignment of Video Diffusion Models via Representation Steering

作者:
Rohit Kundu↗Arindam Dutta↗Sarosij Bose↗Athula Balachandran↗Amit K. Roy-Chowdhury↗

Open-weight video diffusion models can generate photorealistic unsafe content, from violence to misinformation, yet existing defenses either require expensive safety fine-tuning that degrades general capability, or apply external filters that are trivially bypassed by adversarial prompts. We present REINS (REpresentation-space INference-time Safety steering), a training-free method that aligns video diffusion models at inference time by steering their internal representations toward safe generation. Our key finding is that safety-relevant structure is linearly encoded in the hidden-state activations of video diffusion transformers, and a single direction, discovered via Supervised PCA on binary safety labels, suffices to separate safe from unsafe generation trajectories. At inference, adding this direction to hidden states at an intermediate transformer layer redirects generation from harmful content to semantically related safe alternatives, with no weight updates, no concept enumeration, and negligible computational overhead. Through mechanistic analysis, we reveal that while safety information accumulates monotonically with transformer depth, steering effectiveness peaks at intermediate layers (~50% depth), exposing a fundamental tradeoff between information availability and downstream propagation capacity. We evaluate REINS across 9 video diffusion models, multiple parameter scales (1.3B-5B), and both text-to-video and image-to-video generation, to our knowledge, the broadest safety evaluation suite in the video generation literature.

阅读与讨论 → 访问原文 →
10.
arXiv (CS.AI) 2026-06-17

Multiple cyclicity and Wavelet Decomposition with Channel Correlation for Long-term Time Series Forecasting

作者:
Bin Wang↗Heming Yang↗Jinfang Sheng↗

arXiv:2606.17996v1 Announce Type: cross Abstract: Cyclicity and trend are important components of time series data and many studies based on cyclicity and trend have achieved good results in long-term time series forecasting. However, we believe that current work neglects the influence of real-world inter-channel correlations in time series data which leads to suboptimal predictions. Furthermore, these models rely on complex designs to capture diverse information so that resulting in low computational efficiency. To address this challenge, we propose McWC, a long-term time series forecasting model that separately models the cyclicity, trend, and inter-channel correlations. Specifically, McWC first decouples cyclical information from data using a multi-layer cyclicity construction module. Then, it extracts inter-channel correlations using multi-layer perceptron. Next, it models and fuses the multi-layer high-frequency and low-frequency information from data using a multi-level wavelet decomposition module. Finally, it aggregates the results of different components to obtain the output. Simultaneously, we decouple intra-channel autocorrelations by calculating a loss function in the frequency domain. Experiments on six real-world datasets demonstrate that McWC achieves state-of-the-art performance, exhibiting excellent computational efficiency and historical information extraction capabilities.

阅读与讨论 → 访问原文 →
11.
arXiv (CS.CV) 2026-06-24

DiffusionBench: On Holistic Evaluation of Diffusion Transformers

作者:
Xingjian Leng↗Jaskirat Singh↗Zhanhao Liang↗Ethan Smith↗Martin Bell↗Aninda Saha↗Yuhui Yuan↗Liang Zheng↗

Diffusion transformer (DiT) research on image generation has converged to a single evaluation setup: class-conditional generation on ImageNet. While methods improve the FID and related metrics, it is increasingly unclear whether they reflect real progress in generative modeling. The natural alternative, i.e., text-to-image (T2I) generation, is perceived as too costly or inconvenient to train and evaluate and is often skipped. We argue that this perception no longer holds. We introduce NanoGen, a unified DiT training and evaluation framework. NanoGen matches state-of-the-art DiT baselines on ImageNet and, with 12 lines of configuration change, also trains competitive text-to-image models. It currently supports RAE, VAE, pixel-space, and MeanFlow diffusion methods under both ImageNet and T2I setups. Under NanoGen, training T2I requires comparable compute to ImageNet. After training 21 latent diffusion models with NanoGen, we observe that method ranking shows no strong correlation between ImageNet and T2I generation: Pearson correlation is between -0.377 and -0.580 across three metrics. This suggests that a method which improves class-conditional ImageNet FID may show no corresponding improvement on T2I, clearly indicating the necessity of evaluating DiTs on both tasks. To this end, we summarize ImageNet and text-to-image results, which yields DiffusionBench, a holistic benchmark for DiT research. We recommend reporting DiffusionBench in place of ImageNet alone: methods that improve DiffusionBench are more likely to reflect broader progress.

阅读与讨论 → 访问原文 →
12.
arXiv (CS.CL) 2026-06-19

Granularity-Regulated Adaptive Computational Efficiency for Optimal Verification in Test-Time Scaling

作者:
Ardit Krasniqi↗Luan Vejsiu↗Elira Dervishi↗

Test-time scaling (TTS) has emerged as a powerful paradigm for improving the reasoning performance of large language models (LLMs) by investing additional compute at inference time. A central component of TTS is the verifier, which selects or scores candidate solutions to guide the search process. While prior work has explored the benefit of verification, a fundamental question remains underexplored: what is the optimal granularity of verification under a given compute budget? Coarse-grained outcome reward models (ORMs) and fine-grained process reward models (PRMs) represent two extremes, yet neither alone achieves compute-optimality across all regimes. In this paper, we establish a unified theoretical framework, called GRACE (\underline{G}ranularity-\underline{R}egulated \underline{A}daptive \underline{C}omputational \underline{E}fficiency), that characterizes the optimal verification granularity as an explicit function of problem difficulty, verifier accuracy, and compute budget. We prove that there exists a phase transition: fine-grained verification dominates when either the compute budget is large or the problem is hard, whereas coarse-grained verification is preferred in the low-budget, easy-problem regime. Our theory unifies Best-of-$N$, beam search, and step-level MCTS within a single Pareto-optimality framework, and motivates an adaptive granularity strategy that provably achieves the compute-performance Pareto frontier. Empirical results on MATH-500, GSM8K, and AIME benchmarks corroborate all four theoretical claims, with our adaptive strategy outperforming fixed-granularity baselines by up to 3.1\% accuracy at matched compute.

阅读与讨论 → 访问原文 →
13.
arXiv (CS.AI) 2026-06-16

S-SPPO: Semantic-Calibrated Self-Play Preference Optimization

作者:
Xiwen Chen↗Wenhui Zhu↗Jingjing Wang↗Peijie Qiu↗Zhipeng Wang↗Huayu Li↗ZhengXiao He↗Xuanzhao Dong↗Prayag Tiwari↗Mingkun Xu↗Yujian Xiong↗Feng Luo↗…

arXiv:2606.01561v2 Announce Type: replace Abstract: Aligning Large Language Models (LLMs) with human preferences is often formulated via Direct Preference Optimization (DPO). However, the standard Bradley-Terry instantiation of DPO is limited in modeling common departures from transitivity in human preferences. To address this, recent work has introduced Self-Play Preference Optimization (SPPO), which iteratively refines the policy by training on self-generated win-lose pairs. Our investigation, however, reveals a critical instability in SPPO: the optimization is prone to policy degeneration when the preference oracle assigns overly confident wins to semantically indistinguishable responses. To mitigate this, we propose S-SPPO, a dual-space semantic calibration framework comprising: i) Supervision Calibration via semantic gating, which anneals win rate targets toward the maximum-entropy baseline as semantic overlap increases; and ii) Representation Calibration via latent repulsion to enforce geometric diversity to prevent manifold collapse and maintain latent diversity between chosen and rejected samples. Theoretically, we show that the calibration preserves the constant-sum game structure, facilitating convergence to a Nash Equilibrium. Empirically, S-SPPO avoids the performance degradation seen in prior methods, achieving 52.19% win rate and 47.46% length-controlled win rate on AlpacaEval 2.0 with Llama-3-8B, without using additional human-annotated preferences during training. The code will be available at https://github.com/xiwenc1/s-sppo.

阅读与讨论 → 访问原文 →
14.
arXiv (CS.CL) 2026-06-11

Automated Creativity Evaluation of Language Models Across Open-Ended Tasks

作者:
Min Sen Tan↗Zachary Kit Chun Choy↗Syed Ali Redha Alsagoff↗Nadya Yuki Wangsajaya↗Mohor Banerjee↗Swaagat Bikash Saikia↗Alvin Chan↗

Large language models (LLMs) have achieved remarkable progress in language understanding, reasoning, and generation, sparking growing interest in their creative potential. Realizing this potential requires systematic and scalable methods for evaluating creativity across diverse tasks. However, most existing creativity metrics are tightly coupled to specific tasks, embedding domain assumptions into the evaluation process, and limiting scalability and generality. To address this gap, we introduce an automated, domain-agnostic framework for quantifying LLM creativity across open-ended tasks. Our approach separates the measurement apparatus from the creative task itself, enabling scalable, task-agnostic assessment. Divergent creativity is measured using semantic entropy, a reference-free and robust metric for novelty and diversity, validated against human annotations, LLM-based novelty judgments and baseline diversity measures. Convergent creativity is assessed via a novel retrieval-based multi-agent judge framework that delivers context-sensitive evaluation of task fulfilment with over 60% improved efficiency. We validate our framework in three qualitatively distinct domains: problem-solving (MacGyver), research ideation (HypoGen), and creative writing (BookMIA), using a broad suite of LLMs. Empirical results show that our framework reliably captures key facets of creativity, including novelty, diversity, and task fulfilment, and reveal how model properties, such as size, temperature, recency, and reasoning, impact creative performance. Our work establishes a reproducible and generalizable standard for automated LLM creativity evaluation, paving the way for scalable benchmarking and accelerating progress in creative AI.

阅读与讨论 → 访问原文 →
15.
arXiv (quant-ph) 2026-06-25

Towards Robust Optimal Measurements Against Noise in Quantum Metrology

作者:
Xinglei Yu↗Xinzhi Zhao↗Liangsheng Li↗Stanis{\l}aw Kurdzia{\l}ek↗Chengjie Zhang↗Chuan-Feng Li↗Guang-Can Guo↗

arXiv:2606.25638v1 Announce Type: new Abstract: Quantum parameter estimation utilizes quantum mechanical effects to attain higher measurement precision than classical schemes. In practical implementations, however, noise is inevitably present during the measurement process, causing a decrease in precision. Quantifying the impact of noise on different measurements is of considerable significance. Here, we experimentally investigate robust optimal measurements based on the theory of Fisher information measurement noise susceptibility (FI MENOS), which quantifies how susceptible a measurement is to noise. By constructing a polarizing Mach-Zehnder interferometer, we implement phase estimation under controlled noise. Our results indicate that different measurements exhibit distinct sensitivities to noise. To assess the influence of diverse noise types on precision, we further construct an experimental setup capable of introducing various forms of noise. The experimental results affirm that FI MENOS represents the worst-case scenario for estimation precision, enabling us to evaluate the noise immunity of optimal measurements. Our work provides a deeper insight into quantum metrology with noise, marking a notable advancement in quantifying the robustness of quantum estimation schemes against measurement noise effects.

阅读与讨论 → 访问原文 →
16.
arXiv (CS.CL) 2026-06-16

Dealing with Annotator Disagreement in Hate Speech Classification

作者:
Somaiyeh Dehghan↗Mehmet Umut Sen↗Berrin Yanikoglu↗

Hate speech detection is a crucial task, especially on social media where harmful content can spread quickly. Collecting social media content (tweets etc.) to train machine learning models is easy, but detecting and categorizing hate speech can be difficult due to the inherently subjective nature. This subjectivity leads to frequent disagreement among annotators, particularly for subtle or borderline content. Traditional approaches either discard non-consensus samples or force a ''gold standard'' through expert adjudication, ignoring valuable information about uncertainty and diverse human perspectives. We examine the largely overlooked problem of annotator disagreement in hate speech classification and evaluate a range of aggregation methods, including majority voting, ordinal strategies (minimum, maximum, and mean), and analyze their impact across binary, 4-class, and 6-class classification tasks. In addition, we leverage annotators' perceived hate speech strength scores to explore regression-based and hybrid modeling approaches. Among others, we show that filtering non-consensus samples results in over-optimistic results and that the perceived strength provides a complementary signal that enhance classification performance. Finally, we establish new state-of-the-art results for hate speech detection in Turkish tweets, and demonstrate that annotator disagreement, when properly modeled, is a valuable resource for building more robust and reliable systems.

阅读与讨论 → 访问原文 →
17.
medRxiv (Medicine) 2026-06-22

Discovering Novel intracranial EEG Biomarkers of Seizure Generating Tissue through Time-Frequency Analysis

作者:
Romero Mila↗Hoang↗N. P↗Pinto-Orellana↗Daida↗Kanai↗Kuroda↗Hussain↗S. A↗Shrey↗D. W↗Asano↗…

Objective: EEG biomarkers for seizure-generating tissue have historically been identified visually, which lacks objectivity and limits utility of automated approaches. For example, high frequency oscillations and interictal epileptiform discharges were promising markers to improve surgical outcomes for refractory epilepsy, but low specificity has hindered clinical implementation, and automated algorithms have not improved this. Methods: We developed Intracranial EEG Pattern Identification and Categorization, an automated, data-driven time-frequency framework for EEG biomarker discovery. It detects transient high-power intracranial EEG waveforms (1-500 Hz) and characterizes them using eight features. In seizure-free patients, waveforms occurring predominantly in resected intracranial EEG channels are candidate biomarkers. Results: In retrospective data from 14 seizure-free post-surgical patients from University of California, Los Angeles, we identified 9 waveform categories strongly associated with resected intracranial EEG channels. These included beta, gamma, and ripple band bursts, sometimes co-occurring with interictal epileptiform discharges; however, many were visually imperceptible in the broadband EEG. Using a support vector machine, we generated a unified classification metric based on these waveforms and tested it on 87 seizure-free subjects from Detroit Medical Center. This metric achieved higher area under the precision-recall curve than six state-of-the-art benchmark algorithms (p

阅读与讨论 → 访问原文 →
18.
arXiv (CS.CL) 2026-06-16

Revisiting the Systematicity in Negation in the Era of In-Context Learning

作者:
Hitomi Yanaka↗Taisei Yamamoto↗

Understanding the meaning of negated sentences remains one of the challenges for language models, even in the era of large language models (LLMs). We analyze systematicity regarding LLM understanding of negation from two perspectives: behavioral systematicity and representational systematicity. For behavioral systematicity, we confirm that through demonstrations and in-context learning, LLMs can recognize negation expressions and scope within sentences to some extent, but they fail to achieve perfect performance. In particular, the difficulty of the negation scope recognition for models varies depending on the output format. For representational systematicity, we analyze the extent to which function vectors can be robustly constructed from in-context examples for tasks that are essential to understanding negation. The experiments suggest that while function vectors can be composed for negation cue extraction tasks, extracting function vectors for recognizing scope is more challenging.

阅读与讨论 → 访问原文 →
19.
arXiv (CS.CL) 2026-06-24

MEMPROBE: Probing Long-Term Agent Memory via Hidden User-State Recovery

作者:
Enze Ma↗Yufan Zhou↗Wei-Chieh Huang↗Jie Yang↗Huanhuan Ma↗Zixuan Wang↗Chengze Li↗Chunyu Miao↗Philip S. Yu↗Zhen Wang↗

Long-term memory promises LLM agents that grow more capable across sessions, maintaining an accurate, evolving understanding of the user that interaction forms. In practice, however, this memory is evaluated mostly through downstream behavior, such as later answers, personalization quality, or task success, which tests that understanding only indirectly and leaves the memory artifact itself largely unaudited. We argue that long-term memory should instead be evaluated as an auditable post-interaction artifact: after ordinary assistance, what structured user state can be reconstructed from the memory the agent leaves behind? We instantiate this view in MEMPROBE, a benchmark in which a memory-equipped agent assists simulated users, each carrying a hidden, taxonomy-anchored user-state bank, across a trajectory of leak-controlled tasks, after which that bank is reconstructed from the agent's resulting memory under both full-store and top-k access. Built on synthetic ground truth for efficient, scalable measurement, MEMPROBE spans 50 simulated users with 31 hidden dimensions each (1,550 recovery targets) and tests 5 representative memory systems. Testing state-of-the-art memory agents, we find that successful assistance and recoverable memory behave as distinct capabilities. Task completion nearly saturates, even for a memoryless baseline, while category-balanced recovery stays moderate (about 0.6) and drops further under top-k retrieval. MEMPROBE is the first benchmark to study memory recovery directly, reconstructing the user state a system retains and scoring it against ground truth. We see recovery as a concrete objective for future memory agents to optimize, and MEMPROBE as a step toward an environment where agents are trained to remember their users, growing more faithful the longer they know them.

阅读与讨论 → 访问原文 →
20.
arXiv (CS.CL) 2026-06-24

Co-occurring associated retained concepts in Diffusion Unlearning

作者:
Miso Kim↗Georu Lee↗Yunji Kim↗Hoki Kim↗Jinseong Park↗Woojin Lee↗

Unlearning has emerged as a key technique to mitigate harmful content generation in diffusion models. However, existing methods often remove not only the target concept, but also benign co-occurring concepts. As illustrated in Fig.1, unlearning nudity can unintentionally suppress the concept of person, preventing a model from generating images with person. We define these undesirably suppressed co-occurring concepts that must be preserved CARE (Co-occurring Associated REtained concepts). Then, we introduce the CARE score, a general metric that directly quantifies their preservation across unlearning tasks. With this foundation, we propose ReCARE (Robust erasure for CARE), a framework that explicitly safeguards CARE while erasing only the target concept. ReCARE automatically constructs the CARE-set, a curated vocabulary of benign co-occurring tokens extracted from target images, and leverages this vocabulary during training for stable unlearning. Extensive experiments across various target concepts (Nudity, Van Gogh style, and Tench object) demonstrate that ReCARE achieves overall state-of-the-art performance in balancing robust concept erasure, overall utility, and CARE preservation.

阅读与讨论 → 访问原文 →
21.
arXiv (CS.CL) 2026-06-19

DeFrame: Debiasing Large Language Models Against Framing Effects

作者:
Kahee Lim↗Soyeon Kim↗Steven Euijong Whang↗

As large language models (LLMs) are increasingly deployed in real-world applications, ensuring their fair responses across demographics has become crucial. Despite many efforts, an ongoing challenge is hidden bias: LLMs appear fair under standard evaluations, but can produce biased responses outside those evaluation settings. In this paper, we identify framing – differences in how semantically equivalent prompts are expressed (e.g., "A is better than B" vs. "B is worse than A") – as an underexplored contributor to this gap. We first introduce the concept of "framing disparity" to quantify the impact of framing on fairness evaluation. By augmenting fairness evaluation benchmarks with alternative framings, we find that (1) fairness scores vary significantly with framing and (2) existing debiasing methods improve overall (i.e., frame-averaged) fairness, but often fail to reduce framing-induced disparities. To address this, we propose a framing-aware debiasing method that encourages LLMs to be more consistent across framings. Experiments demonstrate that our approach reduces overall bias and improves robustness against framing disparities, enabling LLMs to produce fairer and more consistent responses.

阅读与讨论 → 访问原文 →
22.
arXiv (CS.CL) 2026-06-19

MENTOR: Reinforcement Learning via Flexible Teacher-Optimized Rewards for Tool-Use Distillation

作者:
ChangSu Choi↗Hoyun Song↗Dongyeon Kim↗WooHyeon Jung↗Minkyung Cho↗Sunjin Park↗NohHyeob Bae↗Seona Yu↗KyungTae Lim↗

Distilling the tool-use capabilities of large language models (LLMs) into small language models (SLMs) is essential for their practical application. The predominant approach, supervised fine-tuning (SFT), suffers from poor out-of-domain (OOD) generalization due to its rigid alignment with static teacher trajectories. While reinforcement learning (RL) offers an alternative, the capacity limitations of SLMs pose a severe dilemma: sparse outcome rewards provide insufficient guidance, whereas strict trajectory matching imposes overly restrictive constraints. To bridge this capacity-driven gap, we propose MENTOR, which introduces a flexible yet process-aware reward structure. Instead of enforcing rigid replication, MENTOR uses the teacher's reference to guide tool-use behavior, balancing behavioral alignment with downstream performance. Extensive experiments on controlled executable-tool benchmarks demonstrate that MENTOR improves OOD tool-use performance compared to SFT and strict RL baselines. Our findings suggest that within verifiable tool-use environments, flexible tool-use alignment offers a more effective approach than strict trajectory replication for developing adaptable small models.

阅读与讨论 → 访问原文 →
23.
arXiv (CS.AI) 2026-06-17

MapSatisfyBench: Benchmarking Satisfaction-Aware Map Agents through Behavior-Grounded Implicit Decision Factors

作者:
Lubin Bai↗Mengyu Cao↗Sixue Wang↗Zhongwei Wan↗Yue Pan↗Jiale Hou↗Xiang Li↗Xiuyuan Zhang↗

arXiv:2606.17453v1 Announce Type: new Abstract: Large language model agents are increasingly integrated into map services. Since map services are embedded in everyday-life scenarios rather than professional task settings, users often express their needs informally, resulting in underspecified queries with many unspoken needs, namely, implicit decision factors that are critical for user satisfaction. Although clarification is an effective way to mitigate this issue, it increases user burden in daily interaction, and a capable agent should first proactively recover such factors from available information sources. However, evaluating this ability is challenging. The first challenge is to determine which implicit decision factors are suitable for evaluation. A factor is evaluable only if it affects user acceptance and can be recovered from information available to the agent before it responds. Second, user satisfaction cannot be reliably represented by a single reference answer, requiring a benchmark that converts satisfaction-relevant factors into objective and quantifiable evaluation targets. To address these challenges, we propose a restore-identify-filter framework that reconstructs complete user needs from behavior-chain evidence, identifies implicit decision factors, and retains only those supported by pre-query evidence. Building on this methodology, we construct MapSatisfyBench from large-scale, real-world anonymized user data and annotate ground truth from five dimensions and enables full-chain evaluation of satisfaction-aware map agents. Experiments show that current agents generally perform well on explicit task completion, but remain limited in satisfying implicit decision factors and proactively acquiring the evidence needed for satisfaction-aware decisions. These findings establish MapSatisfyBench as a benchmark for shifting map-agent evaluation from task completion toward satisfaction-aware spatial decision making.

阅读与讨论 → 访问原文 →
24.
arXiv (CS.AI) 2026-06-16

The Energy Blind Spot: NVIDIA's Flagship Edge AI Hardware Cannot Support Process-Level Energy Attribution

作者:
Deepak Panigrahy↗Aakash Tyagi↗

arXiv:2605.27599v2 Announce Type: replace-cross Abstract: Agentic AI workloads - where a single user goal triggers multi-step orchestration, tool calls, retries, and failure recovery - are being targeted for edge deployment, with NVIDIA, Dell, HP, ASUS, MSI, Acer, and Gigabyte all shipping GB10-based desktop AI systems in 2026. We recently demonstrated that orchestration structure dominates agentic energy cost, with workflows consuming 4.33x more energy per successful goal than linear baselines and OOI reaching 7.63x for multi-step reasoning tasks. Separately, Raj et al. show that CPU-side processing accounts for up to 90.6% of total latency and 44% of total dynamic energy in agentic workloads. We report a systematic energy-observability audit of the ASUS Ascent GX10 (GB10 SoC) and find that the platform exposes no CPU energy counter, no INA power-rail monitor, no IPMI/BMC, and no SCMI powercap protocol through any supported software interface. The only on-device energy telemetry is instantaneous GPU power via NVML. We further discover that the MediaTek firmware already computes per-rail energy internally via an undocumented ACPI interface (SPBM), but NVIDIA states there are "no plans to expose CPU rail information." On-device per-process energy attribution - as performed on x86 via RAPL - is therefore not reproducible on this platform through supported interfaces. We formalize a hardware requirements specification for energy-attributed AI, propose an interim calibration bridge for per-domain energy decomposition - confirmed on the Acer Veriton GN100 where CPU energy accumulators are live - and identify a standards-track path via SCMI powercap. Our findings motivate the low-carbon computing community to demand energy observability as a first-class hardware requirement.

阅读与讨论 → 访问原文 →
25.
arXiv (CS.CL) 2026-06-16

SHARD: Safe and Helpful Alignment via Self-Reframing Distillation

作者:
Viswonathan Manoranjan↗Amogh Gupta↗Anvesh Rao Vijjini↗Thomas Hofweber↗Snigdha Chaturvedi↗

Large language models often struggle with sensitive prompts. They may refuse outright, provide generic safety boilerplate, or fail to address the user's legitimate informational needs that can be answered safely. We introduce SHARD, a self-reframing distillation method to improve safe-helpfulness. It first rewrites sensitive prompts to surface benign intent using philosophical guidelines, then reframes its original responses into safe, more helpful ones, and finally fine-tunes the model on its self-reframed responses. Across DNA and the English subset of LINGUASAFE, SHARD improves helpfulness for most model families while preserving safety. It also remains competitive with distillation from a larger teacher model, suggesting that models can internalize safe and helpful behavior elicited from their own. Warning: This paper contains content that may be offensive or harmful.

阅读与讨论 → 访问原文 →