探索全球前沿学术脉络

01.

medRxiv (Medicine) 2026-06-22 DOI: HASH:7c412bb86a6b9bb1ed944ef1b709f1f1

Spatial Analysis and Multilevel Determinants of Hypertension in Zambia: Analysis of the 2017 WHO STEPS Survey

作者:

Mutasha↗Simukoko↗Nkandu↗

Background: Hypertension is the leading modifiable cardiovascular risk factor globally, with the fastest-growing burden in low- and middle-income countries. This study aimed to estimate national hypertension prevalence, map provincial patterns, assess spatial clustering, and identify individual and community-level determinants among Zambian adults using the 2017 WHO STEPS survey. Methods: This cross-sectional study used data from the 2017 WHO STEPS survey, a nationally representative sample of 4,301 adults aged 18-69 years. Hypertension was defined as systolic BP [&ge;]140 mmHg, diastolic BP [&ge;]90 mmHg, or current antihypertensive use. Spatial autocorrelation was assessed via Moran's I and LISA. Four nested generalised linear mixed models with PSU-level random intercepts identified individual and community-level determinants. Results: Overall weighted hypertension prevalence was 24.0%. Lusaka recorded the highest prevalence (30.2%), followed by Southern (29.9%) and Muchinga (28.3%) provinces; Western Province had the lowest (12.4%). Spatial clustering was statistically significant but modest (Moran's I = 0.0247, p < 0.001). Between-cluster variation reduced from ICC = 5.9% to 1.8% in the full model, indicating geographic differences were largely explained by individual characteristics. Age was the strongest predictor; adults aged 60-69 had nearly sevenfold higher odds than those aged 18-29 (AOR 6.92, 95% CI: 4.95-9.66). Women had lower odds than men (AOR 0.64, 95% CI: 0.52-0.79). Obesity (AOR 2.34), overweight (AOR 1.65), high cholesterol (AOR 1.40), diabetes (AOR 1.35), and single marital status (AOR 1.34) were independently significant. Western Province showed consistently lower odds than Central Province (AOR 0.48). Conclusion: Hypertension affects one in four Zambian adults, driven primarily by age, sex, obesity, dyslipidaemia, and diabetes. Geographically prioritised interventions, including community health worker-led screening programmes in Lusaka and Southern Province, would maximise population-level impact. Population-level salt reduction and alcohol policies represent cost-effective complementary strategies. Longitudinal studies with finer spatial resolution are needed to clarify causal pathways underlying observed geographic clustering and inform SDG Target 3.4 progress.

阅读与讨论 → 访问原文 →

02.

arXiv (CS.CL) 2026-06-12 DOI: arXiv:2606.09073

A Unifying Lens on Reward Uncertainty in RLHF

作者:

Ely Hahami↗Yoel Zimmermann↗Ray Zhou↗Jack Benarroch Jedlicki↗

Reinforcement learning from human feedback (RLHF) is bottlenecked by reward hacking, where the policy exploits errors in a proxy reward model (RM) and produces high RM scores without genuine quality gains. A natural mitigation is pessimism: lowering rewards in regions where the RM is uncertain. However, standard scalar RMs provide no principled notion of uncertainty. We argue that the right object is a distributional reward model $p(r\mid x,y)$. Under either a Bayesian inference or a KL-distributionally robust optimization (KL-DRO) lens, the KL-regularized RLHF objective admits a closed-form effective reward $\tilde r(x,y) = \pm\beta\log\mathbb{E}_p[e^{\pm r/\beta}]$. The pessimistic branch unifies the prior heuristics for RM ensemble aggregation: mean aggregation, worst-case optimization (WCO), and uncertainty-weighted optimization (UWO) all emerge as limits or truncations of this single expression. This also clarifies the implicit assumptions of each existing rule.

阅读与讨论 → 访问原文 →

03.

arXiv (CS.CV) 2026-06-12 DOI: arXiv:2511.04260

Proto-LeakNet: Towards Signal-Leak Aware Attribution in Synthetic Human Face Imagery

作者:

Claudio Giusti↗Luca Guarnera↗Sebastiano Battiato↗

The growing sophistication of synthetic image and deepfake generation models has turned source attribution and authenticity verification into a critical challenge for modern computer vision systems. Recent studies suggest that diffusion pipelines unintentionally imprint persistent statistical traces, known as signal-leaks, within their outputs, particularly in latent representations. Building on this observation, we propose Proto-LeakNet, a signal-leak-aware and interpretable attribution framework that integrates Closed-set classification with a density-based Open-set evaluation on the learned embeddings, enabling analysis of unseen generators without retraining. Acting in the latent domain of diffusion models, our method re-simulates partial forward diffusion to expose residual generator-specific cues. A temporal attention encoder aggregates multi-step latent features, while a feature-weighted prototype head structures the embedding space and enables transparent attribution. Trained solely on closed data and achieving a Macro AUC of 98.13\%, Proto-LeakNet learns a latent geometry that remains robust under post-processing, surpassing state-of-the-art methods, and achieves strong separability both between real images and known generators, and between known and unseen ones. The codebase is available at the following link: https://github.com/claudiunderthehood/Proto-LeakNet .

阅读与讨论 → 访问原文 →

04.

arXiv (quant-ph) 2026-06-16 DOI: arXiv:2606.14933

Analyzing Initialization Strategies for the Local Unitary Cluster Jastrow Ansatz within the Quantum-Centric Supercomputing Framework

作者:

Grier M. Jones↗Maforikan J. Amoussou↗Maximilian O. Leach↗Hans-Arno Jacobsen↗

arXiv:2606.14933v1 Announce Type: cross Abstract: In this study, we analyze the choice of local unitary cluster Jastrow (LUCJ) ansatz initialization and sensitivity of the sample-based quantum diagonalization (SQD) algorithm within the quantum-centric supercomputing (QCSC) framework. We examine six initialization strategies, including those based on coupled-cluster singles and doubles (CCSD), M{\o}ller-Plesset second-order perturbation theory (MP2), data-driven coupled-cluster (DDCC), and trivial (zeroes and random) initializations, across twelve molecular systems and three basis sets (STO-3G, cc-pVDZ, and aug-cc-pVDZ). We find that while the mean absolute percentage errors (MAPEs) between the alternative and CCSD-initialized t2-amplitudes span many orders of magnitude, the resulting SQD energies are largely insensitive to this variation. In particular, most initializations recover energies within chemical accuracy (+/-1.6 mEh) of the CCSD reference, with convergence improving as the basis set size increases. Notably, random initialization achieves performance competitive with CCSD across all basis sets, while zeroes initialization, despite having smaller deviations from CCSD, yields the worst energy agreement. Our results highlight that the proximity to the CCSD initialization is not a reliable predictor of the quality of electronic energies. These findings establish that configuration recovery within SQD, rather than circuit initialization, is the dominant factor governing energy accuracy, and suggest that computationally cheaper initialization strategies are viable alternatives to CCSD for QCSC workflows

阅读与讨论 → 访问原文 →

05.

arXiv (CS.AI) 2026-06-11 DOI: arXiv:2606.12065

Automating Geometry-Intensive Compliance Checking in BIM: Graph-Based Semantic Reasoning Framework

作者:

Zixuan Xiao↗Pei Troh Koh↗Jun Ma↗Jack C. P. Cheng↗

arXiv:2606.12065v1 Announce Type: new Abstract: Automating compliance check for geometry-intensive regulations remains a significant technical bottleneck in Building Information Modeling (BIM), primarily due to the semantic disparity between high-level regulatory logic and structured IFC data. Existing methods, often reliant on static rule templates, struggle to traverse multi-hop reasoning chains or resolve latent spatial dependencies across multiple building entities. To address these challenges, a Spatial-Geometric Reasoning System for Building Information Modeling (SGR-BIM) is proposed as an integrative graph-driven reasoning framework. SGR-BIM dynamically constructs a cross-modal knowledge graph that aligns user intent, regulatory semantics, and BIM geometry, enabling interpretable reasoning without rigid hard-coding. Validated on 679 expert-verified queries from fire safety codes, the framework achieves 84.3% accuracy, representing an 8.6% improvement over enhanced-tool single-agent baselines. This research provides a graph-based semantic reasoning paradigm, enhancing the transparency and flexibility of automated geometric compliance check workflows in the Architecture, Engineering, and Construction (AEC) industry.

阅读与讨论 → 访问原文 →

06.

arXiv (CS.AI) 2026-06-18 DOI: arXiv:2606.19152

AdsMind: A Physics-Grounded Multi-Agent System for Self-Correcting Discovery of Adsorption Configurations on Heterogeneous Catalyst Surfaces

作者:

Zongmin Zhang↗Yuyang Lou↗Bowen Zhang↗Junwu Chen↗Ryo Kuroki↗Xuan Vu Nguyen↗Edvin Fako↗Lixue Cheng↗Philippe Schwaller↗

arXiv:2606.19152v1 Announce Type: cross Abstract: Identifying the lowest-energy surface-adsorbate configuration is critical for modeling heterogeneous catalysis, yet exhaustive exploration with ab initio calculations is computationally prohibitive. Machine-learning force fields (MLFFs) accelerate structural relaxation but leave the search over the vast configurational space a major bottleneck, and open-loop large language model (LLM) agents lack a physics-grounded feedback mechanism to correct erroneous initial guesses. We propose AdsMind (Adsorption configuration discovery with Machine intelligence and relaxation feedback), a closed-loop multi-agent framework that enables autonomous error correction through MLFF relaxation feedback. Across four LLM backends, AdsMind achieves consistently high search reliability, with success rates of 100% and 98.8% on the benchmarks AA20 and OCD-GMAE62. Relative to its single-pass (1-Shot) ablation it reduces cross-backend energy dispersion, and it uses only 4.11 and 4.67 MLFF relaxations per case, respectively – an approximately 14-fold reduction over heuristic enumeration baselines. Density functional theory (DFT) validation using VASP/PBE on six representative AA20 systems shows that the reported open-loop Adsorb-Agent outputs exhibit qualitative adsorption-energy sign errors for molecular adsorbates, whereas AdsMind preserves the correct sign in all tested cases with closer quantitative agreement. AdsMind thus delivers reliability, self-reflection, and interpretability simultaneously, supporting more DFT-informed autonomous chemistry workflows.

阅读与讨论 → 访问原文 →

07.

arXiv (CS.AI) 2026-06-11 DOI: arXiv:2606.11417

Signed Compression Progress on a Sealed Audit is Goodhart-Resistant

作者:

Ayush Mittal↗Dhruv Gupta↗

arXiv:2606.11417v1 Announce Type: cross Abstract: Compression progress is a long-standing proposal for intrinsic motivation: reward an agent when its world model becomes better at predicting or compressing experience. The folk claim is that this reward is "credible" because it is paid only for learning. We make this precise and prove it. If intrinsic reward is the signed decrease of a fixed sealed-audit loss, r_t = E(theta_{t-1}) - E(theta_t), then cumulative reward telescopes exactly to endpoint audit improvement, so no policy can push reward up indefinitely while true audit performance stagnates or degrades. For finite audit panels the same result holds with a sharp false-positive budget: cumulative empirical reward is at most true audit improvement plus 2 Delta_n(F, delta), the uniform audit deviation of the model class. This is horizon-free: adaptivity over time costs nothing once the sealed panel uniformly controls the class. The theorem also identifies the failure modes: the guarantee disappears if progress is clipped, scored on the agent's own stream, exposed to a high-capacity model on a reusable panel, or applied to a neural class that makes Delta_n vacuous. We give a Lean 4 mechanization of the structural core (telescoping, the finite-audit bound, finite Gibbs, and the entropy floor) and an experiment suite on ARC-TGI grid-transformation generators with adaptive holdout attacks. Experiments confirm the theory: finite-audit deviation scales as n^{-0.527}; signed progress resists clip-farming, stream leakage, and noisy-TV curiosity; naive reusable audits are exploitable by black-box scalar feedback, while standard release defenses keep the attack below the 2 Delta_n threshold. Signed compression progress on a sealed audit is an accounting signal of genuine improvement.

阅读与讨论 → 访问原文 →

08.

arXiv (CS.AI) 2026-06-17 DOI: arXiv:2604.10827

Know Thy Reasoner: Not All Language Models Explore Alike

作者:

Moulik Choraria↗Argyrios Gerogiannis↗Anirban Das↗Supriyo Chakraborty↗Sourya Basu↗Sambit Sahu↗Lav R. Varshney↗

arXiv:2604.10827v2 Announce Type: replace Abstract: Compute scaling for LLM reasoning trades off exploring solution approaches (breadth) against refining promising ones (depth), yet why a given trade-off works, and why it often fails to transfer across models, remains unclear. We argue that the optimal strategy depends on the model's diversity profile, the spread of probability mass across solution approaches, and that this must be characterized before any exploration strategy is adopted. We formalize this with a framework decomposing reasoning uncertainty, deriving when depth-based refinement outperforms parallel sampling, and validate it across three model families at both inference and training. Our central finding is that the diversity regime dictates the strategy: low-diversity aligned models benefit from depth-based refinement with lightweight intrinsic signals, whereas high-diversity base models are often harmed by it, and instead need breadth or stronger signals to compensate.

阅读与讨论 → 访问原文 →

09.

arXiv (CS.CL) 2026-06-19 DOI: arXiv:2606.20097

HydraHead: From Head-Level Functional Heterogeneity to Specialized Attention Hybridization

作者:

Zhentao Tan↗Wei Chen↗Jingyi Shen↗Yao Liu↗Xu Shen↗Yue Wu↗Jieping Ye↗

The quadratic complexity of attention poses a critical bottleneck for long-context processing, spurring interest in hybrid attention designs. Most open-source hybrid models adopt a layer-wise strategy. Yet, prior work has noted the inherent difficulty of integrating Linear Attention (LA) with Full Attention (FA), suggesting that the design space of attention hybridization remains underexplored. To probe this space, we conduct interpretability analysis and observe that layers exhibit block-wise functional similarity, while individual heads within the same layer display distinct functional specialization despite sharing input features. This head-level heterogeneity suggests that the head dimension provides a natural and principled granularity for fusing heterogeneous attention signals. Building on this insight, we introduce HydraHead, a novel architecture that hybridizes FA and LA along the head axis. HydraHead features two key innovations: (1) an interpretability-driven selection strategy that identifies retrieval-critical heads and preserves FA only for them, and (2) a scale-normalized fusion module that reconciles the distributional gap between FA and LA head outputs. By leveraging a three-stage transfer pipeline with parameter reuse and distillation, we achieve high-performance hybrid models with minimal training overhead. Under a unified training setup, HydraHead outperforms other hybrid designs in long-context tasks while maintaining strong general reasoning. With interpretability-driven head selection, it matches a 3:1 layer-wise hybrid's long-context performance at a 7:1 LA-to-FA ratio. Crucially, trained on only 15B tokens, HydraHead achieves over 69% improvement over the baseline at 512K context length, approaching Qwen3.5, a leading model of comparable size with a native context length of 256K. This highlights the significant scaling potential of head-level hybridization.

阅读与讨论 → 访问原文 →

10.

arXiv (CS.AI) 2026-06-16 DOI: arXiv:2606.16276

SpecAlign: Efficient Specification-Grounded Alignment of Large Language Models via Synthetic Data

作者:

Wenjie Wang↗Yue Huang↗Zhengqing Yuan↗Han Bao↗Shiyi Du↗Yuchen Ma↗Yue Zhao↗Yanfang Ye↗Xiangliang Zhang↗

arXiv:2606.16276v1 Announce Type: new Abstract: As large language models (LLMs) are increasingly deployed in real-world applications, alignment is no longer governed by a single universal notion of safety or helpfulness, but instead by provider- or application-specific model specifications. These specifications are typically long, structured, and frequently updated, yet existing alignment pipelines lack a systematic mechanism to operationalize them as training signals. In this paper, we propose specification-grounded alignment, a new alignment paradigm that treats provider-authored model specifications as the primary alignment target rather than abstract principles or static benchmarks. To instantiate this paradigm, we introduce SpecAlign, a framework that synthesizes alignment data directly from specification documents. SpecAlign combines structured rule annotation, controllable specification instantiation, and multi-agent adversarial data synthesis to generate fine-grained, boundary-aware preference pairs that capture both compliant behaviors and meaningful specification violations. Experiments across multiple model specifications and backbone models demonstrate that training with SpecAlign consistently improves rule compliance while preserving general capabilities and avoiding over-conservative behavior. These results suggest that grounding alignment in explicit model specifications enables rapid, precise, and scalable adaptation of LLM behavior to evolving policy requirements.

阅读与讨论 → 访问原文 →

11.

bioRxiv (Bioinfo) 2026-06-21 DOI: HASH:23e38677966470f8e02bd28fe5117889

GENATATORs: ab initio Gene Annotation With DNA Language Models

作者:

Shmelev↗Shadskiy↗Kuratov↗Burtsev↗Kardymon↗Fishman↗

Inference of gene structure and location from genome sequences - known as de novo gene annotation - is a fundamental task in biological research. However, sequence grammar encoding gene structure is complex and poorly understood, often requiring costly transcriptomic data for accurate gene annotation. In this work, we benchmark current solutions and develop new methods of gene annotation. We show that pretrained DNA language model (DNA LM) embeddings do not capture the features necessary for precise gene segmentation, and that task-specific fine-tuning remains essential. We comprehensively evaluate the impact of model architecture, training strategy, receptive field size, dataset composition, and data augmentations on gene segmentation performance. We revisit standard evaluation protocols, showing that commonly used per-token and per-sequence metrics fail to capture the challenges of real-world gene annotation. We introduce and theoretically justify new biologically grounded metrics, along with benchmarking datasets that better capture annotation quality. We show that fine-tuned DNA LMs outperform existing annotation tools, generalizing across species separated by hundreds of millions of years from those seen during training, and providing segmentation of previously intractable non-coding transcripts and untranslated regions of protein-coding genes. Our results thus provide a foundation for new biological applications centered on accurate gene annotation.

阅读与讨论 → 访问原文 →

12.

arXiv (CS.LG) 2026-06-16 DOI: arXiv:2602.24012

InfoNCE Induces Gaussian Distribution

作者:

Roy Betser↗Eyal Gofer↗Meir Yossef Levi↗Guy Gilboa↗

arXiv:2602.24012v2 Announce Type: replace Abstract: Contrastive learning has become a cornerstone of modern representation learning, allowing training with massive unlabeled data for both task-specific and general (foundation) models. A prototypical loss in contrastive training is InfoNCE and its variants. In this work, we show that the InfoNCE objective induces Gaussian structure in representations that emerge from contrastive training. We establish this result in two complementary regimes. First, we show that under certain alignment and concentration assumptions, projections of the high-dimensional representation asymptotically approach a multivariate Gaussian distribution. Next, under less strict assumptions, we show that adding a small asymptotically vanishing regularization term that promotes low feature norm and high feature entropy leads to similar asymptotic results. We support our analysis with experiments on synthetic and CIFAR-10 datasets across multiple encoder architectures and sizes, demonstrating consistent Gaussian behavior. This perspective provides a principled explanation for commonly observed Gaussianity in contrastive representations. The resulting Gaussian model enables principled analytical treatment of learned representations and is expected to support a wide range of applications in contrastive learning.

阅读与讨论 → 访问原文 →

13.

arXiv (CS.LG) 2026-06-15 DOI: arXiv:2606.13982

Adaptive Nucleus Truncation for Long-Form Reasoning

作者:

Ousmane Amadou Dia↗

arXiv:2606.13982v1 Announce Type: cross Abstract: Sampling plays an important role in long-form language-model reasoning. Over thousands of decoding steps, small changes in the candidate token set can compound into different reasoning trajectories, stability profiles, and final answers. Existing truncation methods such as top-$p$, min-$p$, and fixed top-$n\sigma$ sampling improve over unrestricted sampling, but they rely on fixed thresholds that cannot adapt to changes in entropy, task difficulty, training stage, or generation budget. We introduce Adaptive Nucleus Truncation Sampling (ANTS), which extends top-\(n\sigma\) sampling from a fixed decoding rule into an adaptive rollout-control mechanism for long-form generation. ANTS selects standardized neighborhoods around the maximum logit before temperature scaling, adapts the truncation width using an entropy-conditioned controller, and retains a no-truncation fallback arm to stabilize training when truncation becomes unsafe. On a 33B-total / 4B-active sparse Mixture-of-Experts reasoning model, ANTS improves average performance over percentage-based benchmarks by +1.9, +3.8, and +5.2 points at 8K, 16K, and 32K generation budgets, respectively. The strongest gains appear on instruction following and mathematical reasoning, with IFBench improving by more than 10 points at 32K and AIME 2025 improving by 7 points. Code generation reveals an important budget interaction. On Codeforces, ANTS trails the baseline at 8K, but reverses this gap and substantially improves ELO at 16K and 32K. These results suggest that sampler design should be treated not just as a decoding hyperparameter, but as part of how we stabilize and scale long-budget reasoning.

阅读与讨论 → 访问原文 →

14.

arXiv (CS.LG) 2026-06-16 DOI: arXiv:2606.15978

Scalar-Stepsize Nonuniform Monte Carlo Optimistic Policy Iteration: A Certified Counterexample

作者:

Yuanlong Chen↗

arXiv:2606.15978v1 Announce Type: new Abstract: Tsitsiklis proved convergence of Monte Carlo optimistic policy iteration under a uniform update structure and identified nonuniform update frequencies as a delicate obstruction. We give a certified negative answer for the natural scalar-stepsize, unnormalized asynchronous state-value recursion with fixed nonuniform state-selection probabilities. In a three-state, two-action discounted MDP, the nonuniform update frequencies induce a diagonally scaled greedy-policy mean field with a certified nonconstant attracting hybrid periodic orbit. With a bounded unbiased geometric-horizon estimator and Robbins–Monro stepsizes, the original stochastic recursion remains trapped near the cycle with positive probability and therefore fails to converge. The example pinpoints a geometric obstruction: uniform sampling gives radial residual contraction, whereas scalar nonuniform sampling anisotropically distorts the residual dynamics and can generate switched attracting cycles.

阅读与讨论 → 访问原文 →

15.

arXiv (CS.AI) 2026-06-11 DOI: arXiv:2304.13905

LSTM based IoT Device Identification

作者:

Kahraman Kostas↗

arXiv:2304.13905v2 Announce Type: replace-cross Abstract: While the use of the Internet of Things is becoming more and more popular, many security vulnerabilities are emerging with the large number of devices being introduced to the market. In this environment, IoT device identification methods provide a preventive security measure as an important factor in identifying these devices and detecting the vulnerabilities they suffer from. In this study, we present an end-to-end machine learning pipeline that identifies IoT devices in the Aalto university dataset (IoT devices captures) using Long Short-Term Memory (LSTM) networks. Raw network packet captures (PCAP) are processed into 25 engineered features, which are then arranged as sliding-window time-series sequences. We systematically evaluate sequence lengths from 2 to 20, reporting that performance improves approximately linearly up to length 6 and thereafter in a wave-like pattern, reaching its peak at length 18. On the final held-out test set with the optimal configuration, the model achieves an accuracy of 79.85% and a macro-averaged F1-score of 75.70% across 27 device classes.

阅读与讨论 → 访问原文 →

16.

arXiv (CS.CL) 2026-06-11 DOI: arXiv:2606.11316

Schützen: Evaluating LLM Safety in Bulgarian and German Contexts

作者:

Kiril Georgiev↗Yuxia Wang↗Dimitar Iliyanov Dimitrov↗Preslav Nakov↗Ivan Koychev↗

Large language models are increasingly deployed across professional domains, bringing hard-to-predict risks, including the generation of harmful or disrespectful content. Although substantial progress has been made in developing safety evaluation datasets, existing resources remain overwhelmingly English- and Chinese-centric. This limitation is particularly pronounced when evaluating languages that operate within shared sociocultural, legal, and ethical contexts. To address this gap, we introduce Sch\"{u}tzen: a German–Bulgarian safety dataset designed to assess model answerability under risk, covering both a low-resource language (Bulgarian) and a high-resource language (German). Experiments with multilingual and language-specific LLMs reveal pronounced cross-language differences in safety behavior, highlighting the necessity of tailored, region-specific evaluation resources to support the responsible deployment of LLMs in Germany and Bulgaria. Datasets and code are available at https://github.com/xnlp-lab/Schutzen. Warning: this paper contains examples that may be offensive, harmful, or biased.

阅读与讨论 → 访问原文 →

17.

arXiv (CS.AI) 2026-06-24 DOI: arXiv:2606.24626

SAFARI: Scaling Long Horizon Agentic Fault Attribution via Active Investigation

作者:

Chenyang Zhu↗Jiayu Yao↗Kushal Chawla↗Youbing Yin↗Nathan Wolfe↗Pengshan Cai↗Jingyu Wu↗Spencer Hong↗Sangwoo Cho↗Shi-Xiong Zhang↗Daben Liu↗Sambit Sahu↗…

arXiv:2606.24626v1 Announce Type: new Abstract: As autonomous agents tackle increasingly complex multi-step, multi-agent tasks, their execution trajectories have scaled beyond the constraints of even the largest context windows. Current methods for effectively diagnosing agent failures load the full trajectory into an LLM's context window, which suffers from attention dilution and fails when agentic traces inevitably exceed context limits. To address this, we introduce SAFARI (Scaling long-horizon Agentic Fault AttRibution via active Investigation), a framework that replaces linear context loading with a tool-augmented diagnostic loop. By equipping LLMs with a specialized toolbox to read and search trajectory segments alongside a persistent Short-Term Memory (STM) for cross-turn reasoning, SAFARI effectively decouples diagnostic accuracy from architectural context limits. Our experiments demonstrate that SAFARI outperforms state-of-the-art results by 20% on the Who&When dataset within a 1M token budget, and by 19% on TRAIL GAIA subset on a 25K token budget. Most significantly, SAFARI maintains a 0.58 precision even when the target fault resides 5x beyond the model's native context window, a scenario where traditional evaluators fail entirely.

阅读与讨论 → 访问原文 →

18.

medRxiv (Medicine) 2026-06-22 DOI: HASH:5516a3b44e114a4522e4b1f234c3f04c

Survival differences and artemisinin resistance in severe malaria among HIV coinfected patients: data from Mozambique

作者:

Noormahomed↗E. V↗de Sousa↗I. M↗McArthur↗Chambal↗Akrami↗Cowell↗Buene↗T. P↗Schooley↗Ajay↗…

Abstract Background Malaria remains a significant cause of morbidity and mortality, especially in sub-Saharan Africa, where rates of HIV coinfection are high. This study aimed to determine whether Plasmodium falciparum malaria treatment outcomes and rates of antimalarial resistance markers differ according to HIV serostatus in Mozambique. Methodology We conducted an observational study of non-pregnant adults, with and without HIV coinfection, admitted to the Hospital Central de Maputo for treatment of severe malaria. Plasmodium falciparum DNA was extracted from whole blood and sequenced to identify single-nucleotide polymorphisms. Statistical analyses to compare clinical outcomes and rates of nonsynonymous mutations in genes associated with drug resistance were performed in R version 4.2. Results We recruited 149 study participants aged between 18-62 years, 72 (48.3%) were female, and 59 (39.6%) were infected with HIV. Comparing clinical outcomes, we found a significant difference in anemia (hemoglobin

阅读与讨论 → 访问原文 →

19.

arXiv (CS.LG) 2026-06-12 DOI: arXiv:2606.12623

Estimating Individualized Treatment Effects in Acute Ischemic Stroke with Causal Transformation Models (TRAM-DAG): A Multi-Centre Observational Study with External RCT Validation

作者:

Oliver D\"urr↗Lisa Herzog↗Pascal B\"uhler↗Susanne Wegener↗Beate Sick↗

arXiv:2606.12623v1 Announce Type: cross Abstract: Personalized medicine in acute ischemic stroke requires moving beyond average treatment effects (ATE) to individualized treatment effect (ITE) estimates to support treatment decisions. In acute ischemic stroke, mechanical thrombectomy has been shown to be more effective on average than lysis in randomized controlled trials (RCTs), such as the MR CLEAN study. We aim to identify which individual patients benefit most from mechanical thrombectomy compared to lysis. The outcome of interest is the modified Rankin Scale (mRS) at three months, an ordinal measure of functional disability (0: no symptoms, 6: death). We demonstrate that causal transformation models on directed acyclic graphs (TRAM-DAG) can be used for ITE estimation after being fitted on observational MAGIC multi-center stroke patient data. To ensure comparability with the MR CLEAN population, which we use for validation, we train the TRAM-DAG on a MAGIC sub-population with NIHSS at admission >= 6, corresponding to one inclusion criterion of MR CLEAN. The fitted model is then used to estimate ITEs for stroke patients in the MR CLEAN population. While these ITE estimates cannot be confirmed experimentally, we show that their average is consistent with the trial's reported ATE. Furthermore, the ITE estimates correctly rank trial patients by their observed frequency of a good outcome (mRS at three months

阅读与讨论 → 访问原文 →

20.

arXiv (CS.CL) 2026-06-12 DOI: arXiv:2510.16928

ChiKhaPo: A Large-Scale Multilingual Benchmark for Evaluating Lexical Comprehension and Generation in Large Language Models

作者:

Emily Chang↗Niyati Bafna↗

Existing benchmarks for large language models (LLMs) are largely restricted to high- or mid-resource languages, and often evaluate performance on higher-order tasks in reasoning and generation. However, plenty of evidence points to the fact that LLMs lack basic linguistic competence in the vast majority of the world's 3800+ written languages. We introduce ChiKhaPo, consisting of 8 subtasks of varying difficulty designed to evaluate the lexical comprehension and generation abilities of generative models. ChiKhaPo draws on existing lexicons, monolingual data, and bitext, and provides coverage for 2700+ languages for 2 subtasks, surpassing any existing benchmark in terms of language coverage. We further show that 6 SOTA models struggle on our benchmark, and discuss the factors contributing to performance scores, including language family, language resourcedness, task, and comprehension versus generation directions. With ChiKhaPo, we hope to enable and encourage the massively multilingual benchmarking of LLMs.

阅读与讨论 → 访问原文 →

21.

arXiv (CS.AI) 2026-06-17 DOI: arXiv:2606.17454

Dissecting model behavior through agent trajectories

作者:

Gaurav Gupta↗Vatshank Chaturvedi↗Jun Huan↗Anoop Deoras↗

arXiv:2606.17454v1 Announce Type: new Abstract: AI agent performance is not just a modeling problem, it is fundamentally a systems problem. The advanced capabilities of models are realized through agent harnesses. Therefore, a gap between model assumptions and harness behavior can easily prevent the model's full capabilities from translating into agent performance. We formalize this as the `intent-execution' gap: the mismatch between what the model intends and what the harness executes, and vice versa. We argue that minimizing this intent-execution gap is as important as other aspects of harness design such as tools and execution loops. To illustrate the impact of this harness-model alignment, we develop a simple and customizable harness called `Simple Strands Agent' (SSA). SSA aims to find the bulk of common patterns which generalize across different model families (such as Claude, Gemini, GPT, Grok, Qwen), as well as a small number of model-specific preferences. We make two contributions: (i) we $reproduce or improve on the pass@1$ performance reported by diverse model-provider families on popular agentic benchmarks (SWE-Pro, SWE-Verified and Terminal-Bench-2), and (ii) building on an $analysis of 138k trajectories generated by SSA$, we look beyond the $\texttt{pass@1}$ numbers which tend to be relatively even across frontier models. By representing agent trajectories in code state-spaces, we observe model-level differences in problem-solving behavior. Finer-grained metrics such as edit frequency, testing activity, and phase-transitions reveal how individual models allocate effort across different stages of autonomous problem solving.

阅读与讨论 → 访问原文 →

22.

arXiv (CS.LG) 2026-06-16 DOI: arXiv:2606.16517

How Post-Training Shapes Biological Reasoning Models

作者:

Lukas Fesser↗Hanlin Zhang↗Michelle M. Li↗Eric Wang↗Bryan Perozzi↗Shekoofeh Azizi↗Sham M. Kakade↗Marinka Zitnik↗

arXiv:2606.16517v1 Announce Type: new Abstract: Scientific reasoning models for biology combine language models with foundation models trained on multimodal biological data, including DNA, RNA, and proteins. These models are built through post-training, yet how each stage shapes reasoning and generalization remains poorly understood. We study when post-training improves performance and when it induces over-specialization. Across genomics, transcriptomics, and proteins, we train and evaluate more than 100 biological reasoning models under controlled variation in backbone, continued pre-training (CPT), supervised fine-tuning (SFT), and reinforcement learning (RL), measuring both in-domain (ID) and out-of-domain (OOD) performance. We find that each post-training stage reshapes generalization in a distinct way rather than contributing uniform gains. CPT improves downstream performance by aligning models with biological language. SFT consistently increases ID performance but causes OOD performance to peak early and decline as models fit the training distribution. RL, when applied to strong SFT checkpoints with aligned rewards, improves OOD performance and partially recovers generalization. These results show that biological reasoning does not improve monotonically with additional supervision or compute. Instead, performance depends on how training stages are composed. Under fixed post-training budgets, the strongest ID-OOD trade-off comes from brief SFT, larger RL allocations, and asymmetric adaptation capacity across stages.

阅读与讨论 → 访问原文 →

23.

arXiv (math.PR) 2026-06-17 DOI: arXiv:2606.17792

The Loss of Tension in an Infinite Membrane with Holes of Decaying Spatial Density

作者:

Lukas Early↗Stanislav Volkov↗

arXiv:2606.17792v1 Announce Type: new Abstract: What is the effect of randomly removing material from an infinite stretched membrane? Under what conditions can the membrane still sustain tension? This problem was introduced by Robert Connelly in connection with applications of rigidity theory in the natural sciences, and was later studied in M. V. Menshikov, K. A. Rybnikov, and S. E. Volkov, "The loss of tension in an infinite membrane with holes distributed according to a Poisson law" (2002); a discrete version was also considered in Robert Connelly, Konstantin Rybnikov, and Stanislav Volkov, "Percolation and the Loss of Tension in an Infinite Triangular Lattice" (2001). We study a mathematical framework based on a non-homogeneous Poisson point process whose intensity $\lambda$ tends to zero at infinity. The hole shapes are i.i.d.\ and independent of their locations. We show that if the intensity does not decay too quickly, then tension is still lost throughout the whole plane, as in the homogeneous model studied in 2002. Conversely, we give sufficient conditions under which complete loss of tension does not occur. Thus, both destruction and non-destruction regimes are possible even when the intensity tends to zero, indicating a phase transition in the model. The processes studied here are closely related to bootstrap percolation.

阅读与讨论 → 访问原文 →

24.

arXiv (CS.CV) 2026-06-16 DOI: arXiv:2606.16658

Vision-Language Models as Zero-Annotation Oracles in Histopathology

作者:

Vishal Jain↗Giorgio Buzzanca↗Sarah Cechnicka↗Maarten Naesens↗Priyanka Koshy↗Tri Nguyen↗Jesper Kers↗Candice Roufosse↗Bernhard Kainz↗

Foreground segmentation is the critical first step of every computational pathology pipeline, yet existing methods rely on hand-tuned heuristics or supervised models that overfit to narrow stain and scanner distributions, failing silently on specialised stains such as Jones silver or Elastica van Gieson. We propose a coarse-to-fine approach that recasts foreground segmentation as a visual perception task and leverages general-purpose vision-language models (VLMs) as zero-annotation oracles. Our key insight is that tissue-versus-background discrimination is a natural-image recognition problem, not a histopathological one, so VLMs trained on internet-scale corpora generalise where domain-specific models cannot. We introduce Leica-75, a benchmark of 75 renal transplant whole-slide images spanning three stain families. On Leica-75, our method achieves the highest segmentation quality on out-of-distribution stains (Dice 0.858 +/- 0.027 on Jones, 0.853 +/- 0.041 on EVG) with 7x lower cross-stain variance than the best supervised baseline, while remaining competitive on in-distribution H&E. Few-shot prompting with automatically curated exemplars (Auto-context) rescues hard cases on Stress-32 (n=32), a curated stress-test subset (Dice 0.470 to 0.819 for the 2B model). VLM-based annotation review matches human expert consensus (kappa=0.989 for blur detection; mean precision/recall grading accuracy 0.708 vs. human 0.646 for segmentation mask review). The resulting pseudo-labels are used to distil lightweight student models that are as performant as the teacher model while running for a fraction of the cost. Our framework provides a principled, scalable solution to a persistent infrastructure bottleneck in digital pathology.

阅读与讨论 → 访问原文 →

25.

arXiv (math.PR) 2026-06-16 DOI: arXiv:2606.14812

Joint convergence in Wiener chaos via transport hierarchy and Malliavin covariances

作者:

Ciprian A Tudor↗

arXiv:2606.14812v1 Announce Type: new Abstract: We study the joint convergence in distribution of a sequence $X_N = I_p(f_N)$ of multiple Wiener–Itô integrals of order $p\geq 2$ that converges to a Gaussian limit $Z\sim N(0,\sigma^2)$, together with another sequence $Y_N = I_q(g_N)$ converging in law. The central finding is that the joint convergence of $(X_N, Y_N)$ is completely governed by the asymptotic behavior of the iterated Malliavin covariances $Y_{r+1,N} = \langle DX_N, DY_{r,N}\rangle_H$, $r\geq 0$: joint convergence holds as soon as these covariances converge jointly with $Y_N$, and the structure of the limiting distribution is then explicitly determined by their limits. Moreover, the convergence of the Malliavin covariances is necessary for joint convergence, as shown by a counterexample. When $q

阅读与讨论 → 访问原文 →