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

Treatment Response Optimized Clinical Decision Support AI System via Digital Twin Simulation

arXiv:2606.17405v1 Announce Type: new Abstract: Clinical decision support AI systems (CDSASs) must adapt to evolving patient conditions in real-time while adhering to strict safety constraints. We present an online adaptive framework that integrates Treatment Effect (TE) estimation to quantify clinical benefits, a patient Digital Twin (DT) to simulate treatment trajectories, and Reinforcement Learning (RL) for sequential decision-making. The AI system is initially trained on historical medical records and operates in a continuous learning loop. To ensure safety, a rule-based module monitors vital signs and blocks contraindicated treatments. Cases with strong internal model disagreement are flagged for clinician review, simulated in our experiments via a pre-trained outcome model. We validate our framework using both a synthetic clinical simulator and a real-world ovarian cancer dataset from The Cancer Genome Atlas (TCGA). In both simulated and clinical settings, our method demonstrated superior effectiveness and stability in recommending treatments compared to standard computational baselines. Furthermore, the AI system maintains low latency and requires expert consultation for only a minority of cases in our experimental validation, demonstrating its potential as a safe, clinician-supervised tool for personalized medicine that continuously improves through practical use.

02.
arXiv (CS.AI) 2026-06-11

PROJECTMEM: A Local-First, Event-Sourced Memory and Judgment Layer for AI Coding Agents

arXiv:2606.12329v1 Announce Type: new Abstract: AI coding assistants now support a growing share of software work, from quick scripts to production applications. Yet these agents remain largely stateless: each new session re-reads project files, re-derives prior decisions, and - most costly - may repeat debugging attempts that already failed. Reconstructing this context can consume an estimated 5,000-20,000 tokens per session; the bottleneck is often not model capability but missing project memory. We present projectmem, an open-source, local-first memory and judgment layer for AI coding agents. projectmem records development as an append-only, plain-text event log of typed events - issues, attempts, fixes, decisions, and notes - and deterministically projects that log into compact, AI-readable summaries served through the Model Context Protocol (MCP). Beyond storage, projectmem adds a deterministic pre-action gate that warns an agent before it repeats a previously failed fix or edits a known-fragile file. We frame this as Memory-as-Governance: memory that does not merely answer the agent but acts on its next action. The system runs fully offline with no telemetry; its immutable log also serves as a provenance trail for reproducible, auditable AI-assisted development. projectmem ships as a three-dependency Python package (14 MCP tools, 19 CLI commands, 37 automated tests) and is evaluated through a two-month self-study across 10 projects comprising 207 logged events. Source code: https://github.com/riponcm/projectmem.

03.
Nature (Science) 2026-06-12

‘Student Geng’ ignites research-integrity scandal in China after calling out senior academics<b> </b>

作者:

Video blogger’s viral accusations of data manipulation in Nature journals have sparked intense debate and speedy institutional investigations. Video blogger’s viral accusations of data manipulation in Nature journals have sparked intense debate and speedy institutional investigations.

04.
arXiv (CS.LG) 2026-06-15

A theoretical model for task routing in mixture-of-expert transformers

arXiv:2606.14398v1 Announce Type: new Abstract: Mixture-of-experts (MoE) layers enable the scaling of transformer models while keeping the inference compute fixed. While task-expert specialization has been observed in empirical studies of frontier MoE transformer models, existing theoretical work analyzes this using continuous mixture models that cannot be used to model natural language effectively. An important open question is to theoretically explain task-expert specialization in transformer MoE models using discrete models of language. To address this, we represent structured knowledge via syntactic templates and finite key-value dictionaries, and prove formally that a single-layer MoE transformer can encode knowledge by using experts that specialize in the corresponding tasks. Our construction shows how queries are routed to unique, task-specific experts whose size depends solely on the intrinsic complexity of the given task (i.e. the combined size of its syntactic templates and factual dictionary). Our construction provides a theoretical support for empirical results on localized knowledge circuits in MoE models. We support our theoretical findings with experiments evaluating model performance under varying MoE loss functions.

05.
medRxiv (Medicine) 2026-06-12

Microbial etiology, antibiotic susceptibility profiles, and multidrug resistance of urinary tract infections at a secondary healthcare facility in Ghana

Background: Rising antibiotic resistance challenges empirical therapies for urinary tract infections (UTIs). This study evaluated the microbial etiology, susceptibility profiles, and multidrug resistance (MDR) patterns of uropathogens among outpatients at the Berekum Holy Family Hospital, Ghana. Methods: This cross-sectional study (February to August 2021) screened 263 symptomatic outpatients. Mid-stream urine samples underwent quantitative culture, biochemical identification, and antimicrobial susceptibility testing via the Kirby-Bauer disc diffusion method following the 2021 CLSI guidelines. Results: Significant bacteriuria prevalence was 22.8% (60/263). UTIs predominated in females (78.3%, 47/60; p = 0.1501) and individuals [&ge;]45 years (33.3%, 20/60). Gram-negative rods accounted for 90.0% of isolates, primarily Escherichia coli (26.7%), Citrobacter spp. (25.0%), and Enterobacter spp. (21.7%); Staphylococcus aureus (10.0%) was the only Gram-positive pathogen. Extreme phenotypic resistance was observed against piperacillin/tazobactam (98.3%), cefotaxime (93.3%), tetracycline (88.3%), and cefoperazone (85.0%). Conversely, highest therapeutic susceptibilities were retained by amikacin (78.3%), levofloxacin (61.7%), and gentamicin (58.3%). Conclusion: The high prevalence of MDR uropathogens against advanced beta-lactamase inhibitor combinations and cephalosporins necessitates an immediate re-evaluation of regional empirical protocols. Amikacin, levofloxacin, and gentamicin remain viable options prior to culture confirmation. These findings establish a crucial phenotypic baseline to guide localized prescribing policies and regional antimicrobial resistance tracking strategies.

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

A Machine Learning Framework for Real-Time Personalized Ergonomic Pose Analysis

This paper introduces a new methodology for real-time prediction of ergonomic and non-ergonomic human poses using volumetric video data in three dimensions. Although the methodology was designed for ergonomic assessments, it can be adapted to other applications requiring real-time analysis of human posture. One aspect that makes this system stand out is its ability to analyze 3D point clouds during the assessment, enabling computation from multiple angles. This overcomes a critical limitation of cameras which provide often a fixed viewpoint, thereby restricting the data available for a thorough postural evaluation, especially when occlusions occur. The system continuously and automatically performs pose inference using the chosen perspective on the real-time streaming data; however, only the poses manually selected and labeled by the user are used to train the personalized deep learning classifier. The methodology has been refined through a case study in which RGB-D cameras captured subjects performing load-lifting tasks, enabling real-time skeletal labeling. The model was trained on this data and, following the training phase, performs inference on new streaming data in real time. This research offers a scalable and pragmatic approach for real-time ergonomic evaluation by combining state-of-the-art 3D data technologies and traditional 2D pose estimation algorithms. It addresses the increasing need for safety and health monitoring in workplace environments, marking a notable contribution to the domain.

07.
arXiv (CS.CL) 2026-06-19

Target-Side Paraphrase Augmentation for Sign Language Translation with Large Language Models

Sign language translation (SLT) remains constrained by the limited availability of paired sign-video/text corpora and by the heavy-tailed vocabularies typical of real-world datasets. We study a target-side augmentation strategy in which a large language model (LLM) generates controlled paraphrase variants of the reference spoken-language sentence while the sign input remains unchanged. Concretely, we use GPT-4o to produce semantically faithful variants of the training targets and train a Signformer-style pose-based Transformer under a two-stage schedule: pre-training on the augmented corpus followed by fine-tuning on the original references. We evaluate this strategy on three datasets that span complementary challenges: PHOENIX14T (German Sign Language), a real-world corpus with moderate lexical diversity; the Greek Sign Language Dataset with highly controlled, repetitive recordings; and LSA-T (Argentinian Sign Language), a naturalistic corpus with a large vocabulary and severe long-tail sparsity. This range allows us to characterize precisely when and why target-side augmentation is beneficial. On PHOENIX14T, augmentation improves BLEU-4 from 9.56 to 10.33, demonstrating that paraphrastic exposure helps the decoder generalize beyond memorized reference phrasing. The near-saturated GSL baseline and the extremely sparse LSA-T setting reveal the limits of the approach: in both cases, single-reference lexical overlap metrics are insufficient to capture the full picture, motivating a complementary semantic evaluation. To our knowledge, this is the first study to examine LLM-generated target-side paraphrases as an augmentation mechanism for SLT, and the first to apply an LLM-as-a-Judge evaluation protocol to SLT. This complementary evaluation reveals gains in semantic fidelity that lexical overlap metrics understate.

08.
Nature Medicine 2026-06-11

Clinical Profile and Genomic Characterization of the 2026 Bundibugyo Virus Index Case in Uganda

Bundibugyo virus disease (BVD) remains a high-consequence threat in Eastern and Central Africa, where cross-border mobility, nonspecific early symptoms, and delayed recognition can obscure transmission. In this case report, we describe Uganda’s 2026 BVD index case: a male patient who traveled from the Democratic Republic of the Congo to Uganda and was admitted to a private hospital in Kampala on 11 May 2026 after more than two weeks of vomiting and diarrhea, with epigastric pain, weakness, and hiccups. He deteriorated rapidly, developing acute kidney injury, pulmonary edema, hepatic dysfunction, hypoxemia, delirium, atrial flutter, possible disseminated intravascular coagulation, and multiorgan failure, and died on 14 May. A posthumous EDTA whole-blood specimen tested at the Central Emergency Response and Surveillance Laboratory was positive for orthoebolavirus RNA and confirmed as Bundibugyo virus (BDBV) by RT-qPCR. Sequencing achieved 99% genome coverage at ≥100× depth. The 2026 BDBV genome formed a distinct lineage approximately equidistant from the 2007–2008 Butalya and 2012 Isiro variants, differing by 216–227 nucleotides (~1.2% sequence divergence). Here, we demonstrate the value of fatality surveillance, private-sector surveillance, diagnostic optimization through national specimen referral, and rapid molecular-genomic diagnostics for early detection, transmission chain interruption, and public health response coordination.

09.
arXiv (CS.CL) 2026-06-11

The Structural Attention Tax: How Retrieval Format Hijacks In-Context Learning Independent of Content

Retrieval-augmented generation (RAG) systems inject external knowledge to improve LLM outputs, yet the format of injected content – distinct from its semantic relevance – can independently distort the model's attention distribution. We identify and formalise a phenomenon we term the structural attention tax: knowledge graph (KG) triples, due to their relational delimiters and repeated slot patterns, capture 2-3x more attention per token than semantically equivalent natural-language text ($\hat{o}$(KG) $\approx$ 0.70 vs. $\hat{o}$(neutral) $\approx$ 0.25), compressing demonstration attention by up to 42% – regardless of whether the triples are relevant or noise. We develop a formal framework decomposing attention scores into semantic and structural components (Eq. 2), derive a compression bound (Proposition 1) connecting token-level format bias to demonstration attention loss, and show that the structural term governs how much attention is diverted while the semantic term governs whether this helps or hurts. This decoupling reveals two orthogonal axes for improving retrieval-augmented ICL: optimising retrieval quality (semantic axis) and reducing format-driven attention capture (structural axis). Empirically, across two model families (Mistral-7B, LLaMA-3-8B) and three QA benchmarks, we observe that source-task alignment dominates: task-matched BM25 retrieval achieves 58-62% on HotpotQA vs. ConceptNet's 25-27%, a >30 pp gap that dwarfs all gating strategies ($\leq$2 pp). We derive five structure-aware mitigation strategies from the framework, ranging from zero-cost prompt modifications to training-time regularisation; format flattening (S3) is validated by both accuracy and attention-level evidence from a verbalized-triple control, while structural dispersal (S1) yields mixed results that illuminate the challenges of format-level intervention.

10.
arXiv (CS.LG) 2026-06-15

LapidaryEngine: Fully Conversational Crystal Generation

arXiv:2606.14215v1 Announce Type: new Abstract: The emergence of Large Language Models (LLMs) has inspired the vision of generating bespoke crystal materials directly from natural-language instructions, enabling users to design materials through intuitive, conversational interaction. Existing text-to-crystal generative models represent important early steps toward this goal, but they suffer from two critical limitations: (i) restricted input formats that require highly structured descriptions (e.g., chemical formulas), and (ii) one-directional generation, where models can map text to crystal but cannot perform the inverse. These limitations prevent fully conversational workflows and hinder alignment with users' inherently ambiguous and evolving desiderata. We address these challenges with LapidaryEngine, the first model to support fully conversational crystal generation. LapidaryEngine accepts free-form natural-language requests and performs iterative refinement and editing in a dialogue-like manner. The key innovation is a pivot representation, a third, intermediate form that enables bidirectional translation between text and crystal structures despite the absence of direct paired datasets. Leveraging this pivot allows robust interpretation of user feedback and precise structural control. We demonstrate LapidaryEngine across diverse tasks, including insulator discovery, stability optimization, compositional modification, and structural editing, showcasing its ability to align generated materials with user intent in an interactive manner.

11.
arXiv (CS.CL) 2026-06-15

LLMs Contain Multitudes: How Deployment Context Reshapes Model-Level Preferences and Values

Large language models (LLMs) are increasingly characterised in recent evaluation work as having stable, model-level preference and value systems. However, accompanying robustness checks are limited to incidental prompt perturbations such as syntax variation and option reordering. This leaves open whether the measured properties survive when the surrounding task context changes, as it does in most real deployments. We test this directly across two established pairwise paradigms: ranking country preferences and eliciting utility judgements. In both, we make the deployment context – the high-level task the model is performing while making concrete value-dependent choices – our controlled variable, varied across framings such as writing a Reddit post or a news article. Across five LLMs and over 1.2M pairwise decisions, deployment context produces variation far larger than prompt paraphrasing and temperature controls. In country preference rankings over 15 countries, context induces widespread, statistically significant rank shifts; the aggregate Global North favouritism reported in prior work is itself context-dependent, with each model's bias shifting systematically across contexts. In utility elicitation over 50 outcomes, broad cross-category ordering is preserved, but fine-grained rankings within domains vary substantially, and cardinal exchange rates between outcomes (e.g. how many lives in one region equal one in another) shift by a factor of 2.47 at the median. Reported model-level preferences and utilities are therefore better understood as context-conditioned measurements than fixed model-level properties: safety guarantees obtained under one framing provide limited assurance in another.

12.
bioRxiv (Bioinfo) 2026-06-10

HOMED enables hierarchical and multimodal optimization of DNA methylation deconvolution across tissues

Cellular heterogeneity is a major confounder in bulk DNA methylation data for epigenome-wide association studies. Existing reference-based DNAm deconvolution methods often ignore hierarchies among related cell types and may generalize poorly across datasets due to limited variability in reference profiles. We developed HOMED (Hierarchically Optimized Methylation Deconvolution), a framework that integrates cell-lineage hierarchies, single-cell RNA sequencing-guided deconvolution, and paired bulk RNA-seq/DNAm data for CpG signature optimization. Across simulated and real peripheral blood mononuclear cell, lung, and placental datasets, HOMED consistently yielded the highest PCCs and lowest RMSEs, outperforming existing scRNA-seq-guided DNAm deconvolution methods, improving accuracy, resolution, and cross-tissue generalizability.

13.
arXiv (CS.CV) 2026-06-16

An Ensemble Deep Learning Approach for Reliable and Scalable Lemon Leaf Disease Classification

Early detection of plant diseases is crucial to plants and for the farmers. Plant diseases reduce fruit yield and quality, and plants are more susceptible to other stresses when they are infected. The lemon leaf disease dataset contains 1354 images. The dataset has 9 classes. Among the 9 classes only one class is for healthy leaf, and the other 8 classes are leaf diseases. The dataset was split into training (70%), testing (15%) and validation (15%) sets after comprehensive preprocessing. Two pretrained models (InceptionV3 and MobileNetV2) were applied and then combined these models using an ensemble technique to boost robustness. Ensemble models showed a promising performance of 99.27% accuracy. Adversarial Training is applied to improve models' ability and ensure reliable predictions under noisy data. Grad-CAM visualization highlights the important regions of leaf images that validate the model prediction with confidence level.

14.
arXiv (quant-ph) 2026-06-17

Cumulant expansion approach to the decay dynamics of interacting Mössbauer nuclei after strong impulsive excitation

arXiv:2510.00970v2 Announce Type: replace Abstract: Recent progress in accelerator-based x-ray sources brings higher excitation of ensembles of Mössbauer nuclei closer to experimental feasibility. Yet, a theoretical modeling of the decay dynamics of the interacting nuclear ensemble after the impulsive excitation is still an open challenge. Here, we derive a set of nonlinear equations which is capable of efficiently modeling large nuclear ensembles for arbitrary degrees of excitation. As key signature for higher excitation, we identify a non-linear time-evolution of the nuclear dipole phase, which can be tuned via the scattering geometry, and interferometrically be measured. Furthermore, we identify interesting finite-size effects in the nuclear dynamics of small ensembles. Our results provide important guidance for future experiments aiming at the non-linear excitation of nuclei. We further envision the exploration of finite size-effects in Mössbauer spectroscopy with highest spatial resolution, i.e., small sample volumes.

15.
arXiv (CS.CV) 2026-06-16

DiverseDiT: Towards Diverse Representation Learning in Diffusion Transformers

Recent breakthroughs in Diffusion Transformers (DiTs) have revolutionized the field of visual synthesis due to their superior scalability. To facilitate DiTs' capability of capturing meaningful internal representations, recent works such as REPA incorporate external pretrained encoders for representation alignment. However, the underlying mechanisms governing representation learning within DiTs are not well understood. To this end, we first systematically investigate the representation dynamics of DiTs. Through analyzing the evolution and influence of internal representations under various settings, we reveal that representation diversity across blocks is a crucial factor for effective learning. Based on this key insight, we propose DiverseDiT, a novel framework that explicitly promotes representation diversity. DiverseDiT incorporates long residual connections to diversify input representations across blocks and a representation diversity loss to encourage blocks to learn distinct features. Extensive experiments on ImageNet 256x256 and 512x512 demonstrate that our DiverseDiT yields consistent performance gains and convergence acceleration when applied to different backbones with various sizes, even when tested on the challenging one-step generation setting. Furthermore, we show that DiverseDiT is complementary to existing representation learning techniques, leading to further performance gains. Our work provides valuable insights into the representation learning dynamics of DiTs and offers a practical approach for enhancing their performance.

16.
arXiv (CS.CL) 2026-06-11

Models That Know How Evaluations Are Designed Score Safer

The validity of AI safety evaluations depends on models behaving consistently across controlled and deployment settings. Prior work has identified test-time contextual cues, such as hypothetical scenarios, as a source of verbalized evaluation awareness and subsequent behavioral shift. In this paper, we investigate a potential explanation of this phenomenon: evaluation meta-knowledge, defined as parametric knowledge about the structural traits that characterize evaluations. Similar to dataset contamination, where benchmark exposure leads to higher performance through memorization, we hypothesize that models trained on texts describing evaluation practices may implicitly learn to recognize and respond to evaluation-like contexts, for instance, through exposure to scientific articles or social media posts about AI benchmarking. To test this, we fine-tune models on synthetic documents describing evaluation traits such as verifiable structures or moral dilemmas. Evaluating this fine-tuned model on six safety benchmarks, we find that it is significantly safer than the base model and control model. This behavioral shift persists even when restricting the analysis to responses lacking explicit verbalization of evaluation awareness. Our results demonstrate that evaluation meta-knowledge may inflate safety benchmark performance, introducing a novel confounder that is independent of explicit memorization or verbalized evaluation awareness, thus, challenging to detect. These findings have important implications for the design and interpretation of AI safety evaluations. Our code and models are available at https://github.com/compass-group-tue/arxiv2026_evaluation_meta_knowledge.

17.
arXiv (CS.LG) 2026-06-12

Rubric-Guided Self-Distillation: Post-Training Without Rubric Verifiers

arXiv:2606.12507v1 Announce Type: new Abstract: Rubrics have emerged as an alternative to RLVR in open-ended domains where a single ground-truth final answer is not available. Existing rubric-based training methods rely on an LLM verifier that scores each rollout against rubrics. This introduces substantial training-time overhead, exposes optimization to verifier-specific biases, and reduces rubric feedback to a sparse end-of-trajectory signal. We propose Rubric-Guided Self-Distillation (RGSD), a verifier-free training method in which the base policy, conditioned on the rubric, serves as the teacher for the unconditioned student. RGSD distills the rubric-conditioned teacher distribution into the student token-by-token, replacing sparse trajectory-level rewards with dense per-token learning signals and removing the LLM judge from the training loop entirely. Across Qwen-2.5 (3B, 7B) and Qwen3-Thinking (4B, 8B) models on medical and science domains, RGSD achieves rubric satisfaction comparable to judge-based GRPO while using one on-policy rollout per prompt and no training-time verifier calls. Ablations show that raw rubrics provide a stronger teacher enrichment signal than self-generated reference responses, while a stronger GRPO judge can outperform RGSD in some settings, positioning RGSD as a complementary verifier-free alternative when verifier cost or reliability is the bottleneck.

18.
medRxiv (Medicine) 2026-06-15

Bidirectional associations between cannabis use, oddball performance, and P3 event-related potential

Importance: Cannabis use remains prevalent in youth despite concerns regarding its potential impact on cognitive function. Unraveling whether the association between cannabis use and cognition is partially due to preexisting differences or primarily related to use is vital to understanding underlying mechanisms. Objective: To estimate the longitudinal association between cannabis initiation and cognitive trajectories, indexed by task performance and P3 event-related potential (ERP), and to estimate whether baseline cognition is associated with cannabis initiation. Design: Data were analyzed from the ongoing longitudinal Collaborative Study on the Genetics of Alcoholism (COGA) cohort, which was followed up approximately every 2-5 years from 2004 to 2025. Setting: 6 sites across the United States. Participants: Adolescent and young adult offspring of past COGA participants and control families who reported on their cannabis use and who had Visual Oddball (VOP) performance and P3 ERP data (N=4814; 52.4% female, 68.4% white) were grouped based on the timing of cognitive data collection relative to cannabis initiation into Pre-onset (n=2,449; [&ge;]1 assessment) and Post-onset (n=998; [&ge;]3 assessments) subsamples. Main Outcomes and Measures: VOP measures include performance accuracy (%), reaction times (ms), and P3 amplitude (V) and latency (ms) during target trials. Cannabis measures included lifetime use of cannabis (i.e., ever used) and age at first use. Results: High P3 amplitude, and prolonged P3 latency and reaction time were associated with a reduced hazard of cannabis initiation (All Hazards Ratio, [H.R.s]< 0.91, p's

19.
arXiv (CS.CL) 2026-06-12

SkillChain: Closing the Loop on Skill Evolution for Image-Based E-Commerce AI Assistants

Image-based AI assistants are now deployed at production scale on e-commerce platforms, where a single uploaded image can trigger fundamentally different user intents: product search, style recommendation, visual encyclopedia, or utility tool calls, each demanding its own response format, tool invocation, and domain knowledge. Without per-intent behavioral constraints, LLM-based systems conflate these heterogeneous modes and fall short of domain quality standards, while the breadth and dynamism of the intent space render manual engineering infeasible. To address this, we present SkillChain, which closes the production feedback loop on Skill evolution, automating the lifecycle of Skills through three stages: Skill Creator for bootstrapping from task specs and trajectories, Route Optimizer for routing alignment, and Body Refiner for iterative Skill Body refinement via dual-path LLM-Judge evaluation. Deployed on a production-scale e-commerce image assistant, SkillChain substantially improves aggregate response quality, with the strongest gains on structural compliance and content quality; a one-week online A/B experiment further confirms significant gains in user engagement, content consumption, and long-term retention.

20.
arXiv (CS.CL) 2026-06-16

Few-Shot Biomedical Relation Extraction with Large Language Models: A Viable Alternative to Supervised Learning?

Biomedical relation extraction (BioRE) is a key step in transforming biomedical literature into structured knowledge. However, most existing approaches rely on supervised models trained on costly annotated datasets, limiting their scalability and adaptability across relation types and domains. We investigate few-shot BioRE using prompt-based learning with large language models (LLMs) and compare two task formulations: pairwise classification, which predicts relations for individual entity pairs, and joint generation, which extracts multiple relations in a single model call. Experiments on the BioREDirect dataset reveal a clear precision-recall trade-off. Pairwise classification achieves higher recall, whereas joint generation is more precise and computationally efficient. The best-performing model achieves a micro-F1 score of 0.44, substantially outperforming previous few-shot results (0.34) while remaining below the supervised baseline (0.56). Much of this gap is attributable to a single ambiguously defined relation type. When evaluated using macro-F1, which better captures performance across relation types in an imbalanced setting, prompt-based approaches outperform the supervised baseline (0.45 vs. 0.38), particularly on rare relation types. These findings highlight the potential of LLMs for BioRE in low-resource settings and underscore the importance of well-defined relation schemas.

21.
arXiv (CS.AI) 2026-06-17

Beyond the Sampled Token: Preserving Candidate Support in RLVR

arXiv:2510.14807v3 Announce Type: replace Abstract: We revisit exploration collapse in reinforcement learning with verifiable rewards (RLVR), from the perspective of the candidate distribution for next-token prediction. We formally show that as probability concentrates on the top-$1$ candidate, the expected number of distinct responses collapses to one regardless of the sampling budget $K$. This theoretical implication is further verified by our empirical tracking of top-$N$ candidate probabilities during training, where the top-$1$ candidate progressively dominates while plausible alternatives are suppressed. These findings suggest a key desideratum for effective exploration: preserving non-negligible probability mass on the top-$N$ candidates. To this end, we propose Candidate-aware Support Preservation (CaSP), with two complementary designs. Specifically, CaSP redistributes positive gradients among top-$N$ candidates for correct responses, and applies a stronger penalty to the top-$1$ candidate for incorrect responses. Unlike many exploration-oriented methods that improve pass@$K$ at the cost of pass@1, CaSP improves pass@$K$ across the full $K$ spectrum. These gains generalize to 6 math, 2 logical-reasoning, and 2 coding benchmarks, and scales to 32B-parameter models and sampling budgets up to $K=1024$, positioning it as a principled, candidate-level approach for RLVR exploration.

22.
arXiv (CS.AI) 2026-06-12

A Quantitative Experimental Repeated Measures Study of Training Dynamics in a Small Llama Style Language Model Under a Compute-Aware Token Budget

作者:

arXiv:2606.13370v1 Announce Type: new Abstract: This study examines training dynamics in a small Llama-style language model trained under a fixed, compute-constrained token budget. Rather than evaluating efficiency solely through endpoint performance, the study uses a quantitative experimental repeated measures design to analyze how validation loss, validation perplexity, rolling volatility, backslide behavior, spike behavior, and between-seed variability change across token-based training intervals. Six independent training runs were conducted on a 4.26-million-parameter model using the TinyStories corpus, CPU-based full-precision training, and a target budget of approximately 20 million cumulative training tokens. Metrics were collected across 21 intervals, producing 126 seed-by-interval observations. Repeated measures ANOVA showed statistically significant interval effects for validation loss, validation perplexity, and rolling volatility. Descriptive trajectories revealed rapid early improvement followed by non-monotonic degradation during later training intervals. Mean validation loss decreased from 8.3552 at initialization to 2.7996 near 4 million tokens, but increased to 3.9010 by the final checkpoint. Validation perplexity followed the same pattern, falling sharply early in training before rising later. Derived telemetry further showed recurrent validation-loss backslides and no interval-summary evidence of a stable phase under the predefined criteria. These findings suggest that compute-aware language model evaluation should examine training trajectories rather than endpoint metrics alone. In constrained compute settings, additional token exposure may increase computational cost without producing proportional generalization gains, and interval-level telemetry can reveal instability, regression, and diminishing returns that final metrics may obscure.

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

Stepwise Token Selection for Efficient Multimodal Large Language Models

In multimodal large language models (MLLMs), inference cost is largely dominated by the visual token prefix rather than the language backbone, making token reduction a key factor for improving efficiency. Existing approaches typically assign independent importance scores to visual tokens and retain a fixed number of top-ranked tokens, implicitly assuming token independence and a uniform compression ratio across inputs. In this work, we reformulate visual token pruning as a sequential decision-making process. Specifically, we introduce a pointer-style selection mechanism that iteratively chooses informative tokens, conditioning each decision on previously selected ones, and dynamically determines when to stop via a learned termination action. This enables joint optimization of both the selected subset and its size. To enable end-to-end training under standard language modeling objectives, we design a differentiable relaxation based on a variance-preserving noise interpolation scheme, allowing gradients to propagate through the discrete selection process. Extensive experiments on LLaVA-v1.5-7B and Qwen2.5-VL-7B demonstrate that our approach consistently outperforms fixed-ratio baselines across different compression levels. Under aggressive pruning that removes 88.9% of visual tokens, our method preserves 94.6% of the original accuracy while achieving a 1.88x speed-up in prefill latency.

24.
arXiv (CS.CL) 2026-06-16

SING: Synthetic Intention Graph for Scalable Active Tool Discovery in LLM Agents

Large language model (LLM) agents increasingly rely on agent harnesses that manage context, tools, and multi-turn execution, making tools a central interface for acting in realistic digital environments. As harness-connected tool ecosystems expand to hundreds or thousands of APIs, services, and task-specific skills, exhaustive tool schema injection becomes costly and imposes a closed-world assumption that limits agents to a predefined static inventory. Retrieval-augmented tool selection offers a natural alternative, but existing one-shot retrieval methods often fail to align isolated tool descriptions with the agent's true task intention, especially in long-horizon tasks where required capabilities emerge through decomposition, observations, and newly induced subgoals. We propose SING, an intention-aware active tool discovery framework that builds an intention-tool graph linking user intentions, tool capabilities, and tool collaboration patterns, and dynamically retrieves tools according to evolving task states. Using a unified corpus of 7,471 tools, we evaluate SING on three real-world tool-use benchmarks. SING improves Global Recall@5 by up to 59.8% and downstream success rate by up to 28.9% over baselines, while reducing full-corpus tool-schema exposure by 99.8%, demonstrating that intention-aware graph structure enables more accurate and context-efficient tool discovery in large-scale agentic ecosystems.

25.
arXiv (CS.AI) 2026-06-18

Structured Representation Learning with Locally Linear Embeddings and Adaptive Feature Fusion

arXiv:2606.18469v1 Announce Type: cross Abstract: Neuroscientific research has revealed that the brain encodes complex behaviors by leveraging structured, low-dimensional manifolds and dynamically fusing multiple sources of information through adaptive gating mechanisms. Inspired by these principles, we propose a novel reinforcement learning (RL) framework that encourages the disentanglement of dynamics-specific and reward-specific features, drawing direct parallels to how neural circuits separate and integrate information for efficient decision-making. Our approach leverages locally linear embeddings (LLEs) to capture the intrinsic, locally linear structure inherent in many environments, mirroring the local smoothness observed in neural population activity, while concurrently deriving reward-specific features through the standard RL objective. An attention mechanism, analogous to cortical gating, adaptively fuses these complementary representations on a per-state basis. Experimental results on benchmark tasks demonstrate that our method, grounded in neuroscientific principles, improves learning efficiency and overall performance compared to conventional RL approaches, highlighting the benefits of explicitly modeling local state structures and adaptive feature selection as observed in biological systems.