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。

01.
medRxiv (Medicine) 2026-06-15

Efficacy of Painhunting Therapy for Event-Related Depression: A Randomized Controlled Trial with Crossover Replication

Background. Depression affects an estimated 332 million people worldwide and is a leading cause of disability, with up to 80% of major depressive episodes preceded by an identifiable adverse life event [17,18]. First-line treatments target symptoms rather than the precipitating event and are resource-intensive: standard CBT averages roughly 12 sessions, and antidepressant discontinuation carries relapse rates near 35% at six months [8]. These limitations create a clear rationale for brief, structured interventions that address the cognitive and somatic sequelae of adverse life events directly. Painhunting therapy is one such intervention, in which each session targets a discrete adverse event through a structured incident-processing procedure. Methods. We conducted a two-arm, parallel-group, single-site randomised controlled trial comparing Painhunting therapy (Arm A, immediate; n=42) with a waitlist control (Arm B, delayed; n=42) in adults with PHQ-9 >= 9 and active psychological distress related to an adverse life event. After the primary endpoint at T2 (approximately two weeks post-randomisation), Arm B crossed over to active treatment, with T3 as the post-crossover endpoint at approximately four weeks. The primary outcome was PHQ-9 at T2 (between-arm contrast); secondary outcomes were ICG, GAD-7, WHO-DAS 2.0 (12-item), and the Global Impression of Change (GIC). Pre-specified analyses included intention-to-treat, per-protocol, and single-exclusion sensitivity populations. Results. Eighty-four participants were randomised (198 applications, 134 completed screening questionnaire, 119 passed psychometric screening). At T2, mean PHQ-9 was 2.32 (SD 2.59) in Arm A and 16.56 (SD 6.76) in Arm B, yielding an ITT between-arm Cohen d = 2.78 (95% CI 2.19-3.76, p < 0.001). Within-arm paired reductions during each arm's active-treatment window reproduced this magnitude (Arm A T0 to T2 change 14.71, Morris d = 2.80; Arm B T2 to T3 change 14.19, Morris d = 2.77, eligible n=26). Treatment gains were durable at the T4 follow-up (week 8). Aligning each arm to its own end-of-treatment timepoint, the off-treatment drift to week 8 was almost identical between arms: Arm A rose 0.78 points from T2 to T4 (2.19 to 2.97, n=37) and Arm B rose 1.59 points from T3 to T4 (4.74 to 6.33, n=27), the latter falling to 0.77 points once a single documented relapse case (R59) is excluded (4.81 to 5.58, n=26). This small off-treatment rebound then stabilised rather than continuing: Arm A was essentially unchanged from T3 to T4 (change +0.05), with concordant maintenance on ICG, GAD-7, and WHO-DAS. At T4, 68% of Arm A and 41% of Arm B remained in remission (PHQ-9 < 5). Secondary measures (ICG, GAD-7, WHO-DAS) moved in the same direction and to comparable magnitude at every timepoint. The waitlist window in Arm B showed essentially no change on any measure (PHQ-9 change 0.22, p = 0.81). Sensitivity analyses excluding six sub-threshold T2 cases, the single treated-in-error case (R82), the R59 relapse case, and one late T2 submitter left all conclusions unchanged. Conclusions. Painhunting therapy produced large and statistically robust reductions in depression, complicated grief, anxiety, and functional disability over a brief course of three to four sessions, with effect sizes substantially exceeding benchmarks reported for established first-line psychotherapies including CBT and EMDR. Critically, these gains persisted at the week-8 follow-up: depression scores in the immediate-treatment arm were essentially unchanged from four weeks to eight weeks post-randomisation, indicating that the benefit reflects durable change rather than a transient post-session dip. Treatment-window concordance between arms, durability of gains at one month off-treatment, and the flat waitlist trajectory together strengthen the evidence for genuine efficacy rather than spontaneous remission. Baseline covariates including therapeutic alliance, treatment expectancy, self-efficacy, age, and sex showed near-zero associations with outcome, reducing the plausibility of allegiance bias or expectancy effects as primary drivers. The differential retention between arms (88% vs 64% at T3) is attributable to the waitlist design and is discussed as a limitation. These findings support proceeding to a confirmatory active-comparator trial against manualized CBT. Trial registration: ClinicalTrials.gov NCT07490691, prospectively registered.

02.
arXiv (CS.CL) 2026-06-15

Beyond Perplexity: UTF-8 Validity in Byte-aware Language Models

Byte-level tokenization enables language models to handle any Unicode input, but models can generate invalid UTF-8 sequences when encountering rare or unseen characters. We investigate the relationship between training scale and UTF-8 generation reliability with a 355M parameter model trained on 80B tokens from a balanced multilingual corpus of English, Japanese, Korean, and Chinese. We introduce multiple evaluation protocols that isolate UTF-8 structural validity from language modeling. UTF-8 validity convergence lags perplexity by a roughly a factor of two: perplexity stabilizes after 2.1B tokens, but UTF-8 validity requires 4.2B tokens. In context-free generation, rare characters achieve higher structural validity than common characters, suggesting over-specialization of frequent character representations. Through experiments, we observed that reliable UTF-8 generation is a distinct capability requiring evaluation beyond perplexity.

03.
arXiv (quant-ph) 2026-06-24

Universal Extraction of Quantum Critical Exponents and Phase Transitions via Tailored Hilbert Space

作者:

arXiv:2606.24312v1 Announce Type: cross Abstract: Finite-size scaling and the renormalization group form the central toolkit for analyzing quantum phase transitions (QPTs). In this Letter, we introduce a novel Hilbert-space tailoring scheme to probe quantum critical phenomena. Applied to the second-order QPT of the one-dimensional (1D) XY model, our method yields precise critical points and exponents on lattices containing merely 50 unit cells. We further establish the universal applicability of this framework via investigations of the Berezinskii-Kosterlitz-Thouless transition in the 1D XXZ chain: critical parameters are recovered with as few as 12 lattice sites. This technique may open an alternative, efficient route to universally characterize QPT across many-body lattice systems.

04.
arXiv (CS.CV) 2026-06-16

Divide-and-Denoise: A Game-Theoretic Method for Fairly Composing Diffusion Models

The abundance of pre-trained diffusion models provides an opportunity for composition. Combining several models, however, runs the risk of one model dominating or models disagreeing with each other. Here, we propose Divide-and-Denoise, a method for coordinating multiple pre-trained diffusion models during sampling. Much like managing a specialized workforce, our method creates a fair but efficient division of labor across models. Central to our method is the notion of an allocation which defines the responsibility of each model to every region of the noisy sample. At every timestep, we then denoise by (i) updating the allocation by solving a fair division game, where we divide the sample into regions that maximize total utility under fairness constraints, and (ii) aligning the models with this allocation, where we guide each model to denoise within its assigned region. This leads to a new composite denoising process that evolves in tandem with a division process. We evaluate Divide-and-Denoise on conditional image generation. Across several quality metrics, including the GenEval benchmark, our method outperforms baselines and resolves common failures including missing objects and mismatched attributes. Experiments show that Divide-and-Denoise utilizes each model's expertise without neglecting any other model.

05.
arXiv (quant-ph) 2026-06-16

Boson Sampling as a Probe of Chaotic and Integrable Quantum Dynamics in a Photonic Chip

arXiv:2605.25398v2 Announce Type: replace Abstract: Quantum chaos plays a key role in understanding complex quantum dynamics, while integrated photonics offers unique advantages for quantum applications, including high-speed operation, scalability, and programmable unitary transformations. However, integrated photonic approaches to probing quantum chaos remain largely unexplored, owing to the absence of a clear connection between programmable photonic dynamics and established chaos diagnostics. In this work, we establish Fock-state boson sampling as a practical probe of quantum chaos by exploiting the sensitivity of multiphoton interference to the random-matrix properties of underlying single-particle unitary dynamics. More importantly, we design and fabricate a programmable quantum photonic chip to experimentally implement this framework, achieving the first integrated-photonic demonstration of quantum-chaos probes based on boson sampling. Experimental results show that the three complementary probes proposed in this work, namely the distance to Porter–Thomas statistics, Shannon entropy, and Out-of-Time-Ordered-Correlator-equivalent observables, exhibit close agreement with theoretical predictions and consistently distinguish chaotic and integrable dynamics. Our work provides a scalable route for investigating complex quantum dynamics on programmable photonic platforms while leveraging the intrinsic advantages of boson sampling through multiphoton interference and complex output statistics.

06.
arXiv (CS.LG) 2026-06-16

Your Privacy My Cloak: Backdoor Attacks on Differentially Private Federated Learning

arXiv:2606.17035v1 Announce Type: new Abstract: Prior research suggests that differential privacy (DP) inherently enhances the robustness of federated learning (FL) against backdoor attacks. In this paper, we challenge this assumption. Through an empirical analysis of two baseline attack strategies, we uncover a fundamental tension in DP-FL: while bypassing DP allows state-of-the-art defenses to detect and filter malicious updates, complying with DP inadvertently masks their distinguishing statistical characteristics. Consequently, existing defenses become ineffective as DP reduces the raw backdoor signal. Building on this masking effect, we propose RING, a novel attack that explicitly exploits DP to conceal malicious contributions while maximizing attack impact. By collaboratively crafting adversarial perturbations, compromised clients reconstruct a strong backdoor signal during aggregation without triggering anomaly detection. RING operates as a perturbation layer that is agnostic to the underlying backdoor technique, making it broadly applicable and composable with existing attacks – a property that significantly amplifies the threat it poses to DP-FL. Extensive evaluations across four image and text datasets under non-iid distributions show that RING achieves an average attack success rate of 90.3% against six state-of-the-art defenses under a moderate privacy budget, an improvement of up to 26.08x over baseline strategies. Finally, we evaluate potential countermeasures and find that mitigating this threat incurs significant utility trade-offs, exposing a fundamental security gap in the deployment of differentially private FL.

07.
arXiv (CS.CL) 2026-06-18

Notation Matters: A Benchmark Study of Token-Optimized Formats in Agentic AI Systems

Large language models in Agentic AI systems consume tool schemas and execution results and emit tool invocations as structured data. The default language for that exchange, JSON, was designed for application-to-application interchange rather than token efficiency, so its structural elements impose substantial token overhead. Recent work proposes token-optimized alternatives such as TOON (Token-Oriented Object Notation) and TRON (Token Reduced Object Notation) as more compact replacements, but these formats have been evaluated only on isolated comprehension or generation tasks. Whether their token reductions hold inside end-to-end agentic loops therefore remains an open question. We evaluate TOON and TRON on four agentic benchmarks (BFCL, MCPToolBenchPP, MCP-Universe, StableToolBench) and five open-weight LLMs, decoupling input compression from output compression to measure comprehension and generation independently. TRON reduces tokens by up to 27% with accuracy within 14pp of the JSON baseline. TOON achieves up to 18% reduction at a similar 9pp accuracy cost, but additionally cascades on multi-turn parsing failures and collapses parallel tool-call output for most models. The code is available at: https://github.com/lkutschka/notation-matters

08.
arXiv (CS.AI) 2026-06-12

WISE: A Long-Horizon Agent in Minecraft with Why-Which Reasoning

arXiv:2606.12852v1 Announce Type: new Abstract: Rapid advances have been made in developing general-purpose embodied agent in environments like Minecraft through the adoption of LLM-augmented hierarchical approaches. Despite their promise, low-level controllers often become performance bottlenecks due to repeated execution failures. We argue that a key limitation is not only the lack of episodic memory, but also the decoupling of what-where-when memory from which-why reasoning. To address this, we propose WISE (Which-Why Informed Semantic Explorer), a long-horizon agent framework with an enhanced low-level controller equipped with a Causal Event Graph that augments episodic memory with explicit causal structure linking observations to task relevance. Unlike prior work such as MrSteve, which relies on feature similarity for retrieval, WISE enables robust recall under viewpoint changes and supports opportunistic task reordering through causal reasoning. Building on this memory, we propose an Opportunistic Task Scheduler that dynamically re-prioritizes subtasks when causally relevant opportunities are detected. We further equip WISE with a multi-scale progressive exploration strategy to provide spatially comprehensive observations for downstream reasoning. Experiments show that WISE largely improves task success and efficiency on long-horizon sparse tasks, particularly in settings requiring adaptive decision-making.

09.
arXiv (CS.CV) 2026-06-17

MuseVLA: An Adaptive Multimodal Sensing Vision-Language-Action Model for Robotic Manipulation

Humans naturally leverage diverse sensing modalities to interact with the physical world, while most Vision-Language-Action (VLA) models for robotics rely solely on RGB observations. This limits their ability to perceive physical properties that are difficult or impossible to infer from RGB cameras, such as temperature, sound, or radar response. We present MuseVLA, an adaptive multimodal sensing VLA model that integrates novel sensors as on-demand tools for robotic manipulation. Given a task instruction and visual context, MuseVLA first generates a sensor token and target description that select the sensing modality to invoke and what to attend to, analogous to a tool call with arguments. It then converts the selected sensor measurement into a grounded sensor image, a unified intermediate representation that encodes heterogeneous readings for multimodal fusion and action generation. This design decouples sensor-specific processing from the VLA backbone, enabling efficient integration of diverse modalities. To reduce the need for expensive multisensory robot datasets, we further introduce a data synthesis pipeline that augments existing RGB video datasets with grounded sensor images, enabling generalization to unseen sensor-guided tasks. We evaluate MuseVLA on a real-world robot across challenging dexterous hand manipulation tasks that require multimodal sensing inputs, including temperature-guided pick-and-place, audio-driven object search, and radar-assisted hidden object retrieval. MuseVLA achieves 80.6% success rate on average, outperforming RGB-only and multisensory VLA baselines significantly, and exhibits strong zero-shot capabilities on unseen tasks.

10.
arXiv (CS.CL) 2026-06-18

VISUALSKILL: Multimodal Skills for Computer-Use Agents

Computer-use agents (CUAs) approach human-level performance on standardised benchmarks but still struggle on long-horizon tasks and unseen software. Existing skill libraries address this with reusable skills, but represent the skill artifact as text only, despite the visual nature of GUI interaction. We propose VISUALSKILL: a hierarchical multimodal skill, tailored to each target application and organised as a central index over per-topic files, which the agent consumes through a load_topic MCP tool that fetches the relevant topic's text and figures on demand. We construct each skill with a two-stage pipeline that combines authored documentation with live-application UI exploration. On two CUA benchmarks, CUA-World and OSExpert-Eval, a Claude Code CLI agent backed by Claude Opus 4.6 reaches an average score of 0.456 with VISUALSKILL, a +15.3 point absolute lift over the no-skill baseline (0.303). Against a matched text-only skill that is generated from the same source content and differs from VISUALSKILL only in modality, VISUALSKILL yields a further +8.3 point absolute gain over the matched text-only skill (0.373 vs. 0.456), providing direct evidence that retaining visual figures in the skill artifact, rather than verbalizing them away, helps the agent both identify UI elements and verify workflow state after each action. Our code is available at https://github.com/XMHZZ2018/VisualSkills.

11.
arXiv (CS.CL) 2026-06-11

Decoding Multimodal Cues: Unveiling the Implicit Meaning Behind Hateful Videos

Hateful videos have become prevalent on online platforms, highlighting an urgent need for effective detection. However, existing studies primarily focus on binary classification and fail to provide contextual rationales that reveal the implicit meanings behind these judgments, significantly undermining model explainability. To fill this gap, we aim to achieve explainable hateful video detection, enabling models to provide contextual rationales that integrate relevant evidence and logical reasoning alongside decisions. This approach can comprehensively enhance the understanding of video content and the explainability of the decision-making process. We first introduce two datasets, Ex-HateMM and Ex-ImpliHateVid, for explainable hateful video detection. Each dataset provides fine-grained annotations of multimodal harmful elements, along with contextual rationales. We then propose an Information Augmentation and Reasoning Enhancement (IARE) framework designed for explainable detection. The framework employs an information augmentation phase that leverages the multimodal chain-of-thought to integrate harmful elements, thereby enriching rationale evidence. Additionally, IARE incorporates a reasoning enhancement phase, in which Direct Preference Optimization guides the model toward correct reasoning paths and away from incorrect ones, thereby improving the logical coherence of its justifications. We conduct extensive experiments on the two datasets, comparing multiple baselines with our proposed IARE framework. The results demonstrate that IARE achieves state-of-the-art performance while also generating accurate rationales.

12.
arXiv (CS.LG) 2026-06-12

GenAutoML: An Agentic Framework for Dynamic Architecture Generation and Optimization in Time-Series Analysis

arXiv:2606.05860v2 Announce Type: replace Abstract: Designing neural architectures for time-series forecasting and anomaly detection remains a resource-intensive task that often requires substantial domain expertise. Traditional Automated Machine Learning (AutoML) systems typically rely on static, predefined search spaces, limiting their ability to adapt to diverse data characteristics. We present GenAutoML, an agentic framework that leverages Large Language Models (LLMs) as neural architects to bridge natural-language requirements and executable PyTorch implementations. The framework incorporates a Sandboxed Reflection Loop for autonomous code refinement and a Signature-Aware Runtime that enforces architectural consistency and execution safety. To improve robustness under non-stationary conditions, we further introduce a Dynamic Reversible Instance Normalization (Dyn-RevIN) wrapper. Experiments on the ETTh1, ETTm1, and Weather benchmarks demonstrate that GenAutoML can dynamically generate task-specific neural architectures tailored to dataset characteristics. Among the generated models, WaveInterferenceNet achieves inference latency below 0.01 ms per sample while maintaining competitive predictive performance. By emphasizing computational efficiency, architectural adaptability, and stable optimization behavior, GenAutoML enables the creation of ultra-lightweight neural networks suitable for resource-constrained and latency-sensitive Edge AI deployments.

13.
arXiv (CS.AI) 2026-06-24

HiPath: Hierarchical Vision-Language Alignment for Structured Pathology Report Prediction

arXiv:2603.19957v2 Announce Type: replace-cross Abstract: Pathology reports are structured, multi-granular documents encoding diagnostic conclusions, histological grades, and ancillary test results across one or more anatomical sites; yet existing pathology vision-language models (VLMs) reduce this output to a flat label or free-form text. We present HiPath, a lightweight VLM framework built on frozen UNI2 and Qwen3 backbones that treats structured report prediction as its primary training objective. Three trainable modules totalling 15M parameters address complementary aspects of the problem: a Hierarchical Patch Aggregator (HiPA) for multi-image visual encoding, Hierarchical Contrastive Learning (HiCL) for cross-modal alignment via optimal transport, and Slot-based Masked Diagnosis Prediction (Slot-MDP) for structured diagnosis generation. Trained on 749K real-world Chinese pathology cases from three hospitals, HiPath achieves 68.9% strict and 74.7% clinically acceptable accuracy with a 97.3% safety rate, outperforming all baselines under the same frozen backbone. Cross-hospital evaluation confirms generalisation with only a 3.4pp drop in strict accuracy while maintaining 97.1% safety.

14.
arXiv (quant-ph) 2026-06-12

First-order and interior-point methods for entanglement detection

arXiv:2508.05854v3 Announce Type: replace Abstract: Quantum entanglement lies at the heart of quantum information science, yet its reliable detection in high-dimensional or noisy systems remains a fundamental computational challenge. Semidefinite programming (SDP) hierarchies, such as the Doherty-Parrilo-Spedalieri (DPS) and Extension (EXT) hierarchies, offer complete methods for entanglement detection, but it is well known that their practical use is limited by exponential growth in problem size if implemented naively. We make three contributions. First, we introduce a new SDP hierarchy, PST, that is sandwiched between EXT and DP – offering a tighter approximation to the set of separable states than EXT, while incurring significantly lower computational overhead than DPS. Second, we explicitly construct compact, polynomially-scalable descriptions of EXT and PST using partition mappings and operators. These descriptions in turn yield formulations that satisfy desirable properties such as the Slater condition and are well-suited to both first-order methods (FOMs) and interior-point methods (IPMs). Third, we design a suite of entanglement detection algorithms: three FOMs (Frank-Wolfe, projected gradient, and fast projected gradient) based on a least-squares formulation, and a custom primal-dual IPM based on a conic programming formulation. These methods are numerically stable and capable of producing entanglement witnesses or proximity measures, even in cases where states lie near the boundary of separability. Numerical experiments on benchmark quantum states demonstrate that our algorithms improve the ability to solve deeper levels of the SDP hierarchy.

15.
arXiv (CS.CV) 2026-06-12

On the Reliability of Cue Conflict and Beyond

Understanding how neural networks rely on visual cues offers a human-interpretable view of their internal decision processes. The cue-conflict benchmark has been influential in probing shape-texture preference and in motivating the insight that stronger, human-like shape bias is often associated with improved in-domain performance. However, we find that the current stylization-based instantiation can yield unstable and ambiguous bias estimates. Specifically, stylization may not reliably instantiate perceptually valid and separable cues nor control their relative informativeness, ratio-based bias can obscure absolute cue sensitivity, and restricting evaluation to preselected classes can distort model predictions by ignoring the full decision space. Together, these factors can confound preference with cue validity, cue balance, and recognizability artifacts. We introduce REFINED-BIAS, an integrated dataset and evaluation framework for reliable and interpretable shape-texture bias diagnosis. REFINED-BIAS constructs balanced, human- and model- recognizable cue pairs using explicit definitions of shape and texture, and measures cue-specific sensitivity over the full label space via a ranking-based metric, enabling fairer cross-model comparisons. Across diverse training regimes and architectures, REFINED-BIAS enables fairer cross-model comparison, more faithful diagnosis of shape and texture biases, and clearer empirical conclusions, resolving inconsistencies that prior cue-conflict evaluations could not reliably disambiguate.

16.
arXiv (CS.LG) 2026-06-18

EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts

arXiv:2606.18967v1 Announce Type: new Abstract: Reinforcement learning (RL) has become a representative post-training paradigm for LLMs, enabling strong reasoning and agentic capabilities. However, rollout generation remains a dominant latency bottleneck because autoregressive sampling decodes responses sequentially and a small number of long-tailed generations often determine completion time. Speculative decoding (SD) offers a natural way to address this bottleneck, as it is a well-established technique for serving fixed LLMs that reduces latency by rapidly drafting tokens and accepting them through parallel verification while preserving the target-model distribution. However, its practical speedups do not directly carry over to RL rollouts: (i) the evolving target policy makes any fixed drafter increasingly mismatched with the policy's output distribution; and (ii) active batch sizes shrink throughout rollout decoding, shifting decoding from compute-bound to memory-bound regimes where parallel verification can exploit underutilized compute. Therefore, accelerating RL rollouts requires both a drafter that remains effective under long, high-temperature generations from an evolving policy and system-aware use of SD that avoids compute-bound regimes. We present EfficientRollout, a system-aware self-SD framework designed to address this gap for RL rollouts. EfficientRollout induces a quantized drafter from the target model (i.e. self-speculative decoding), keeping it coupled to the evolving policy without separate drafter pretraining or online adaptation. It further coordinates a system-aware SD toggle policy with acceptance-aware draft-length adaptation, enabling speculation only in beneficial regimes while matching the drafting budget to evolving drafter quality. EfficientRollout reduces rollout and end-to-end latency by up to 19.6% and 12.7%, respectively, over an accelerated AR rollout baseline, while preserving final model quality.

17.
arXiv (CS.LG) 2026-06-17

RadSEM: A Finding-by-Finding Metric for Clinical Consistency in Radiology Reports

arXiv:2606.17062v1 Announce Type: cross Abstract: Radiology report evaluation must distinguish clinical compatibility from surface similarity, because negation, laterality, or normal-abnormal polarity can reverse a finding. We propose RadSEM (Radiology Sentence-Level Evaluation Metric), a constrained LLM-assisted metric for reference-based evaluation of radiology Findings. RadSEM rewrites reference and generated reports into ordered atomic finding sentences, each expressing one site-finding proposition. It then performs contradiction-constrained many-to-many matching: incompatible pairs such as "effusion" and "no effusion" receive no credit, while compatible granularity differences can receive partial credit. A deterministic stage weights pairs by part-whole and abnormal-detail relationships, counts unmatched findings, and produces an abnormal-focused weighted F1 score. Thus, the LLM supports structured rewriting and local alignment rather than acting as an opaque judge. We evaluate RadSEM with SSREE, a controlled monotonicity stress test built from 2,448 de-identified reports expanded into five graded corruption levels. RadSEM achieves Kendall tau_b of 0.957, all-pairs concordance of 97.8%, adjacent concordance of 95.0%, and strict five-level ordering for 81.9% of reports, outperforming radiology-specific and general text metrics while avoiding the failure in which polarity-inverted reports regain lexical overlap. On the same SSREE set, RadSEM outperforms the Ref-anchored RadSEM-Alt policy, improving adjacent concordance from 90.7% to 95.0% and strict ordering from 67.2% to 81.9%. On a 599-triplet synonym/antonym subset, RadSEM prefers synonyms in 597 cases (99.67%). These results suggest that explicit finding units, contradiction-aware matching, and abnormal-focused deterministic scoring make report scoring more interpretable and sensitive to clinically meaningful errors. Code is available at https://github.com/jdh-algo/RadSEM.

18.
arXiv (CS.AI) 2026-06-24

GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents

arXiv:2606.24551v1 Announce Type: new Abstract: Computer-use agents can execute software tasks through either graphical interfaces or programmatic command interfaces, but existing evaluations confound interaction modality with differences in tasks, initial states, verifiers, and permitted actions. We introduce a matched execution-layer benchmark of 440 desktop tasks across 18 applications and 12 workflow categories, where screen-only GUI agents and skill-mediated CLI agents receive identical goals, states, and final-state verifiers while being restricted to modality-native actions. In this controlled setting, the strongest GUI agent reaches a 59.1% full pass rate, outperforming the strongest original-skill CLI agent at 48.2%; however, verifier-guided skill augmentation raises CLI success to 69.3%, showing that much of the CLI deficit comes from incomplete skill coverage rather than model capability alone. These results suggest that GUI and CLI expose different execution bottlenecks: GUI agents are limited by reliable grounded interaction over long-horizon workflows, whereas CLI agents are limited by the coverage and scalability of their skill interfaces.

19.
medRxiv (Medicine) 2026-06-15

Entity-Aware Generation of Synthetic Clinical Progress Notes for Prostate Cancer using Large Language Model

Objectives: This study investigates large language models (LLMs) for clinical entity projection across substantial textual transformation. Specifically, we evaluate whether entities annotated in Spanish prostate cancer case reports can be preserved and explicitly projected when the source narratives are transformed into hospital-style clinical progress notes. Entity projection is treated as a generation-driven task, allowing paraphrase, condensation and narrative reorganisation, providing that clinically relevant entities remain recoverable as structured annotations. Methods: A corpus of 109 Spanish prostate cancer case reports was annotated using a silver-standard pipeline combining Spanish biomedical named-entity recognition with rule-based prostate-specific antigen (PSA) and Gleason extractors. The resulting silver-standard annotations were validated on a subset of generated notes against a gold-standard consensus produced by medical experts in prostate cancer. Four LLMs were evaluated for note generation and entity projection: GPT-5.4 Nano, Qwen 3.5:35B-A3B, GLM5 and Claude Sonnet 4.6. Entity-to-Entity (E2E) generation used XML-annotated cases as RAG-supported input, whereas Text-to-Entity (T2E) generation required models to generate and annotate notes directly from plain text cases. Zero-shot and few-shot prompting were tested. Projection quality was measured using precision, recall and F1-score, and complemented by LLM-as-a-judge evaluation using Kimi K2.6. Results: E2E consistently outperformed T2E, indicating that explicit entity-enriched in- put substantially facilitates entity preservation and localisation. GLM5 achieved the best E2E zero-shot result (F1 = 0.915), followed by Claude Sonnet 4.6 (F1 = 0.896). In T2E, few-shot prompting improved performance, with Claude Sonnet 4.6 reaching the highest score (F1 =0.718). Age, Gleason, Disease, Procedure, Duration and negation-related entities were robustly projected, whereas PSA and Dose showed less stable behaviour. Conclusion: LLMs can generate clinically plausible synthetic prostate cancer evolution notes while preserving a substantial proportion of source entities, particularly when explicit semantic annotations are provided as input. However, the lower and more variable performance observed in T2E highlights the difficulty of jointly generating clinical narratives and projecting entities without source-side information, especially for numerical and measure-related entities.

20.
arXiv (CS.CV) 2026-06-24

Do Foundation Models See Biology? Evaluating Attention Coherence with Spatial Transcriptomics in Glioblastoma

Whether attention maps from pathology foundation models capture genuine biology remains unknown, yet this question is critical for clinical trust and regulatory approval. We propose a spatial transcriptomics-based framework for orthogonal, hypothesis-free evaluation of attention and apply it to five pathology foundation models (CONCH v1.5, UNI v2, Virchow2, GigaPath, H-Optimus-1) and a ResNet50 baseline. Using attention-based multiple instance learning, we train single-task and multi-task models to predict five molecular alterations in glioblastoma on the CPTAC cohort, validate on an independent TCGA cohort, and evaluate biological coherence of attention maps against 87 transcriptional signatures using co-registered Visium spatial transcriptomics data from 18 samples. Internally, no single encoder dominates across all tasks, and external validation inverts internal performance rankings. Attention maps show a five-fold enrichment gradient from pathways (Cohen's d=0.329) to individual genes (d=0.055), indicating that attention captures emergent multi-gene transcriptional programs rather than individual molecular events. Spatially smooth attention maps do not imply biological coherence, and different encoders attend to distinct biological compartments. Our framework provides objective, quantitative assessment of what foundation models learn from histopathology, moving the field beyond qualitative saliency map review.

21.
arXiv (CS.AI) 2026-06-18

Guava: An Effective and Universal Harness for Embodied Manipulation

arXiv:2606.18363v1 Announce Type: cross Abstract: Language models trained on large-scale vision-language data have demonstrated strong potential for embodied agents. Harnessing models through embodied tools use offers a promising alternative to end-to-end vision-language-action systems by combining high-level reasoning with external modules for perception, planning, and control. However, it remains unclear what makes an effective harness for embodied manipulation, and to what extent such a harness can unlock embodied capabilities in a wide range of reasoning models. In this work, we present Guava, a harness framework for embodied tool use developed through systematic exploration of the design space of agent workflows, action spaces, and observation spaces. Our study identifies three key ingredients for effective embodied agents: iterative perception-reasoning-action loops, semantic action abstractions, and multimodal observations. To understand whether these design principles are universal even to small models, we develop an end-to-end training pipeline that distills embodied manipulation capabilities into a 4B open-source model using fewer than 2K trajectories collected entirely in simulation. Experimental results in both simulation and real-world environments show performance comparable to frontier proprietary models while exhibiting strong generalization to unseen objects, novel instructions, and long-horizon tasks. Results suggest that a well-designed harness can serve as a scalable, model-agnostic interface for embodied manipulation, enabling strong emergent embodied capabilities in compact open-source models with minimal training data.

22.
arXiv (CS.CL) 2026-06-11

Quantifying Subliminal Behavioral Transfer Ratios in Language Model Distillation

Distillation of a language model intended to transfer benign behavior to a student model may also transfer undesirable characteristics, if they are present in the teacher model, a phenomenon known as subliminal learning. While qualitative evidence supports the existence of this effect, its magnitude has not been systematically characterized. This study quantifies subliminal behavioral transfer ratios by steering two teacher models (Llama-2-7B-Chat and Qwen2.5-7B-Instruct) at varying steering strengths and distilling student models using only benign data. Evaluation on 100 JailbreakBench prompts with GPT-4.1, serving as the evaluator, indicates that transfer is robust but exhibits distinct scaling behaviors. Llama-2 demonstrates a sharp threshold ($\tau = {0.25,0.32} \ beyond \ \alpha = -0.15$), whereas Qwen2.5 displays continuous and higher levels of transfer ($\tau$ up to $0.61$).

23.
arXiv (CS.LG) 2026-06-16

Size Doesn't Matter: Cosine-Scored Sparse Autoencoders

arXiv:2606.15054v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) detect features via inner product, so a feature's activation scales with both its directional alignment and the input's norm. Under BatchTopK, high-norm tokens inflate all pre-activations simultaneously, claiming dictionary slots regardless of content alignment. This matters because sublayer normalization has already discarded the magnitude the score measures, so the encoder detects a quantity the model does not read. We replace the score with a learned blend of cosine similarity and input magnitude, letting the optimizer choose how much norm to use; a per-feature extension lets each feature decide independently. In both regimes, training is free to recover inner product but never does, with no feature ever choosing more than half-magnitude dependence. At matched reconstruction, the cosine encoder learns features that align with human-recognizable concepts far more often than standard, filling dictionary slots that inner product wastes on norm detectors. Loss reweighting that equalizes gradients barely closes the gap, confirming forward-pass score geometry as the lever. The advantage is not universal across tasks or depths, but we believe cosine scoring should be the default for dictionary learning on normalized representations.

24.
arXiv (CS.CV) 2026-06-12

MinhwaNet: Faithful but Insufficient Object Grounding in Korean Folk Painting

Korean folk painting (minhwa) is built from a small vocabulary of auspicious symbols, a tiger for protection, a pair of birds for marital harmony, a peony for wealth, that recur across many of its painted genres. This suggests an obvious computational approach, identify which symbols appear in a painting and read the genre from the inventory. Working with a public corpus that pairs whole paintings, eight-field bilingual curatorial captions, and a separate set of expert object crops, we find that this approach does not work. A model given only a list of which symbols a painting contains predicts the genre far worse than a model that fuses the image with the curatorial text, and forcing the genre representation to be object-grounded actively hurts accuracy. The visual evidence on which the genre prediction rests is nonetheless localized and inspectable. A leakage-safe object evidence map projected from a part-level detector is spatially faithful to where curators isolated symbolic objects and to a patch-based surrogate's own gradient saliency. We name this configuration a faithful-but-insufficient dissociation. The part-level explanation is honest about what the part-level model sees, yet the genre target turns on how symbols are arranged rather than on which ones appear. The same lens separates a content label that survives transfer to held-out source institutions, genre, from a style label that does not, era, a prediction we confirm on two further labels in the corpus. We release the multimodal system, a worked-example reading of one painting's evidence map against its catalogue, and a set of evaluation cautions that recur in long-tailed heritage collections.

25.
medRxiv (Medicine) 2026-06-22

Pump-Free Patient-Derived Human Proximal Tubule Microphysiological System for Modeling Flow-Dependent Epithelial Maturation and Cisplatin Injury

Recent initiatives by the U.S. Food and Drug Administration and the National Institutes of Health to reduce animal testing in drug development have highlighted the need for in vitro platforms that better recapitulate human biology for preclinical safety assessment. Drug-induced nephrotoxicity remains a major cause of drug attrition, underscoring the need for human-relevant kidney models. To address this, a pump-free human patient-derived proximal tubule microphysiological system was developed by integrating human renal proximal tubular epithelial cells (hRPTECs), isolated from non-tumorous nephrectomy cortex, with a porous membrane-based microfluidic device. Expanded hRPTECs were cultured for 10 days under static conditions or rocker-driven shear stress approximating physiological proximal tubular flow. Shear stress increased epithelial density, enhanced proximal tubule marker expression (Na+/K+-ATPase and aquaporin-1), and improved Zonula occludens-1 and occludin localization. Bulk RNA sequencing demonstrated transcriptomic changes associated with enhanced apical maturation and epithelial signature. In cisplatin-induced injury assays, shear-conditioned epithelia exhibited reduced cell density and increased {gamma}H2AX staining, indicating greater sensitivity to nephrotoxicity. These findings demonstrate that rocker-driven shear stress promotes epithelial maturation in patient-derived hRPTECs. The pump-free human patient-derived proximal tubule microphysiological system offers a practical, scalable, and physiologically relevant platform for modeling flow-dependent proximal tubule biology and assessing human-relevant nephrotoxicity.