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) 2026-06-16

Adherence to Red Reflex and Vision Screening Recommendations: A Deep Dive into Primary Care Implementation Gaps

Introduction: Early childhood vision screening is critical for detecting amblyopia and other vision-threatening conditions. Despite screening recommendations during well-child visits, rates remain low. Red reflex assessment is recommended to identify serious ocular pathology, yet its use in primary care is not well described. We examined rates and drivers of vision screening in pediatric primary care. Methods: We conducted a retrospective review of electronic health records for children 3 to 5 years attending well-child visits in 2022 in one of three representative primary care clinics within a university health system. Outcomes were documented red reflex and functional vision tests. We evaluated associations with patient demographics and clinic site using multivariable logistic regression Results: Among 1,003 visits, 21.1% (n=212) had a documented red reflex assessment, and 60.8% (n=610) a functional vision test. Younger children (ages 3 and 4 vs. 5 years) had higher odds of red reflex assessment [adjusted odds ratio (aOR) 9.00 and 8.64], and lower odds of a functional vision (aOR 0.47 and 0.59) test. Females had higher odds of red reflex assessment (aOR 1.53). Other/Multiracial children had lower odds of red reflex assessment than Non-Hispanic White children (aOR 0.48). Screening rates varied significantly by clinic site Conclusions: Visual function and red reflex assessment are inconsistently performed in pediatric primary care, with particularly low rates of red reflex documentation. Screening rates varied between clinics and were affected by age. These findings highlight missed opportunities for early detection of vision-threatening conditions and identify targets for improving adherence to pediatric vision screening recommendations

02.
arXiv (quant-ph) 2026-06-15

Quantifying and detecting quantum-state texture

arXiv:2604.07257v2 Announce Type: replace Abstract: Quantum-state texture is a recently proposed quantum resource that characterizes the inhomogeneity of a quantum state's matrix element distribution in the computational basis, enriching our understanding of quantum state structure. To expand its quantification toolkit and establish detection methods, in this article, we investigate the resource theory of texture from both quantitative and detection perspectives. First, we construct a texture measure $\mathcal{T}^{GR}_{\alpha,z}(\rho)$ based on the $\alpha$-$z$ Rényi relative entropy and present some of its inherent properties. Second, we analyze the mathematical relationships between several existing texture measures, revealing connections among different quantifiers. Finally, drawing on the witness concept from other resource theories, we systematically introduce texture witnesses into the texture theory and provide examples of texture witnesses with special properties.

03.
bioRxiv (Bioinfo) 2026-06-11

PhyloZoo: a unified framework for phylogenetic network analysis in Python

作者:

Reticulate evolutionary processes (events in which lineages merge, such as hybridization, recombination, and horizontal gene transfer) are widespread across nature but cannot be represented by phylogenetic trees alone. Phylogenetic networks have therefore become an important modelling tool, yet existing software is typically tied to specific inference paradigms and provides limited support for working with multiple network representations in a unified and programmable environment. PhyloZoo is an open-source Python framework that lowers the barrier to developing practical, easy-to-use software for phylogenetic network analysis. It provides data structures and algorithms covering the main representations used in the field, together with dedicated visualization tools and robust I/O for all major phylogenetic file formats. A particular emphasis lies on semi-directed phylogenetic networks, which explicitly represent root uncertainty and have so far received limited support in existing software. By offering a shared foundation for developing interoperable tools and a combinatorial layer that supports computational proofs and theoretical exploration, PhyloZoo enables reproducible workflows for applied, methodological, and theoretical studies of reticulate evolution. Availability and implementation: PhyloZoo is implemented in Python and installable from PyPI, with source code, documentation, and examples available at https://github.com/nholtgrefe/phylozoo.

04.
arXiv (quant-ph) 2026-06-16

Synthesizing Arbitrary Non-Hermitian Hamiltonian with Stochastic Floquet Engineering

arXiv:2606.15664v1 Announce Type: new Abstract: The conventional Floquet engineering scheme synthesizes a given target Hamiltonian with a deterministic temporal periodic driving field. In this work, we introduce the stochastic Floquet engineering scheme that can synthesize an arbitrary non-Hermitian target Hamiltonian using a time-periodic driving field with noisy amplitude. Our method is rooted in the Hermitian dynamics taking noise as a valuable quantum resource with no need for loss or gain in prior. We apply our method to engineer a cavity Hamiltonian with dissipative coupling between Fock states, and to prepare a given quantum state from a generally arbitrary quantum state. The stochastic Floqut engineering also provides a way to generate non-unitary quantum gates, which take advantage in certain tasks compared to unitary quantum computing, without the need for ancillae or state-dependent updating.

05.
arXiv (CS.AI) 2026-06-15

FedRot-LoRA: Mitigating Rotational Misalignment in Federated LoRA

arXiv:2602.23638v3 Announce Type: replace-cross Abstract: Federated LoRA provides a communication-efficient mechanism for fine-tuning large language models on decentralized data. In practice, however, a discrepancy between the factor-wise averaging used to preserve low rank and the mathematically correct aggregation of local updates can cause significant aggregation error and unstable training. We argue that a major source of this problem is rotational misalignment, arising from the rotational invariance of low-rank factorizations – semantically equivalent updates can be represented in different latent subspaces across clients since $(B_i R_i)(R_i^\top A_i) = B_i A_i$. When such misaligned factors are averaged directly, they interfere destructively and degrade the global update. To address this issue, we propose FedRot-LoRA, a federated LoRA framework that aligns client updates via orthogonal transformations prior to aggregation. This alignment preserves the semantic update while reducing cross-client subspace mismatch, without increasing communication cost or restricting model expressivity. We provide a convergence analysis that examines the aggregation error induced by factor-wise averaging and shows how rotational alignment yields a tighter upper bound on this error. Extensive experiments on natural language understanding and generative tasks demonstrate that FedRot-LoRA consistently outperforms existing federated LoRA baselines across a range of heterogeneity levels and LoRA ranks.

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

Dual Dimensionality for Local and Global Attention

Decoder-only Transformers compute attention over the KV cache of preceding tokens. Keys (and Values) are typically represented with the same dimensionality, regardless of its distance from the prediction target. In natural language, however, the next word is most strongly influenced by the immediately preceding tokens. We hypothesize that local and distant tokens impose asymmetric demands on representational capacity: local tokens are more critical for predicting immediate outputs and thus require richer representations, whereas distant tokens primarily serve as long-range memory, for which lower-dimensional representations may suffice. We formalize this idea as Distance-Adaptive Representation (DAR), implemented in a controlled setting that preserves full-dimensional representations within a local context window while assigning reduced-dimensional representations (e.g. 1/4 of the original dimensionality) to tokens beyond that window. Across multiple pretraining scales (70M to 410M parameters), as well as continued supervised fine-tuning on a 1B-scale model, this approach closely matches the performance of full-dimensional baselines. In contrast, uniformly reducing dimensionality across all token positions leads to worse performance. These results challenge the common assumption that key and value dimensionality should be uniform across token positions. Our findings suggest a new direction for designing attention architectures that adaptively allocate representational capacity across sequences, enabling further reductions in KV cache during inference.

07.
arXiv (quant-ph) 2026-06-19

Propagating Collective Spin-valley Modes in Twisted WSe2

arXiv:2507.18770v2 Announce Type: replace-cross Abstract: The emergence of neutral collective modes is a hallmark of correlated quantum phases but is often challenging to probe experimentally. In two-dimensional flatband systems, charge responses have been intensively investigated yet neutral excitations remain largely unexplored. In particular, intervalley coherent state (IVC) features a neutral Goldstone mode due to spontaneously broken valley U(1) symmetry. While IVC state has been proposed as a unifying theme across graphene and semiconductor based systems, its defining feature, the neutral Goldstone mode, remains elusive in experiment. Here we investigate space and time resolved transport of neutral modes in twisted WSe2 moire superlattices through a novel ultrafast imaging technique. We uncover two new propagating collective modes with very different velocities, which emerge near the van Hove singularity (VHS) in both intermediate (3.5 to 4 degree) and large (around 5 degree) angle twisted WSe2. The fast-propagating mode has a large speed of about 3 km/s and is consistent with a Goldstone mode for an IVC state, while the slow-moving mode is likely a gapped amplitude mode. They can be understood as the spin-valley analogues of collective modes of a superfluid, whose propagation is imaged for the first time in a condensed matter system. Our study demonstrates a powerful new approach for probing charge-neutral modes in quantum materials and offers key insights into the interplay between charge and spin-valley physics in moire superlattices.

08.
PLOS Computational Biology 2026-06-18

Mechanisms underlying spontaneous and evoked calcium responses in oligodendrocyte precursor cells: A modeling investigation

作者:

by Martin Lardy, Leqi Wang, Claire Guerrier, Veronica T. Cheli, Pablo M. Paez, Anmar Khadra Calcium (Ca2+) signaling has emerged as a central regulator of activity-dependent myelination in oligodendrocytes. These Ca2+ signals encompass both the stimulus-independent spontaneous Ca2+ local transients (SCaLTs) generated intrinsically in a voltage-independent manner or facilitated by the membrane voltage, as well as evoked responses triggered by ATP and glutamate release. To investigate the regulatory mechanisms underlying this combined spiking activity, we developed a stochastic spatiotemporal flux-balance model of Ca2+ transients in oligodendrocyte precursor cells (OPCs). The model incorporates all the relevant fluxes in these cells and integrates membrane voltage dynamics with a Ca2+-induced Ca2+-release (CICR) mechanism using parameters fitted to Ca2+ fluorescence recordings. The model reproduced the intrinsic and voltage-facilitated SCaLTs in OPCs in the absence of purinergic and glutamatergic receptors, and captured the three distinct patterns of evoked Ca2+ responses induced by prolonged ATP and glutamate stimulations identified using machine classifier. The model highlighted the role of ATP and glutamate in generating these clusters, and showed that the fast dynamics of CICR is key to producing these evoked responses. Further analysis of the model also revealed that voltage-gated L- and T-type Ca2+ channels slightly increase the frequency of SCaLTs, while stimulation with ATP and glutamate, using randomly distributed pulses mimicking in vivo conditions, leads to an increase in both the amplitudes of Ca2+ spikes (i.e., the combination of SCaLTs and evoked responses) and the prevalence of wide spikes, especially upon glutamate stimulation. Bifurcation analysis of the deterministic version of the model, in the absence of diffusion, demonstrated that ATP and glutamate stimulation can shift the system into an oscillatory regime, thereby increasing the deterministic component of SCaLT dynamics. This study thus offers a comprehensive representation of OPC Ca2+ transients linking recorded in vitro behaviors to in vivo dynamics.

09.
arXiv (quant-ph) 2026-06-11

Controlled ion-ion interactions and cavity-enhanced emission of a coherent dinuclear Eu$^{3+}$ complex

arXiv:2606.11947v1 Announce Type: new Abstract: Molecular rare-earth-ion complexes offer unique opportunities for quantum technologies by combining the intrinsic coherence properties of rare-earth ions with chemically tunable molecular environments. A crucial capability is the realization of multi-qubit architectures with defined qubit couplings to enable two-qubit quantum gates. Here, we investigate the optical coherence properties and excitation-induced interactions of two Eu$^{3+}$-based molecular complexes, comparing a mononuclear reference system with a dinuclear analogue in which two Eu$^{3+}$ ions are positioned at a well-defined intramolecular distance of about 7 Angstrom. Using cryogenic ensemble spectroscopy, including spectral hole burning, free-induction decay, and photon echo measurements at temperatures down to 100 mK, we demonstrate long optical coherence times $T_{2,o}$ of up to 9 $\mu$s. As a key step toward scalable multi-qubit architectures, a control-target sequence was implemented to probe conditional ion-ion interactions, revealing a stronger interaction-induced dephasing in the dinuclear complex. Finally, we show the integration of the dinuclear complex into a fiber-based optical microcavity, and observe an 380-fold emission enhancement of the $\mathrm{}^5\mathrm{D}_0\rightarrow\mathrm{}^7\mathrm{F}_0$ transition. Together, these results position molecular rare-earth complexes as versatile and chemically tunable building blocks for scalable quantum technologies.

10.
arXiv (CS.LG) 2026-06-12

Strategic PAC Learnability via Geometric Definability

arXiv:2605.13426v3 Announce Type: replace Abstract: Strategic classification studies learning settings in which individuals can modify their features, at a cost, in order to influence the classifier's decision. A central question is how the sample complexity of the induced (strategic) hypothesis class depends on the complexities of the underlying hypothesis class and the cost structure governing feasible manipulations. Prior work has shown that in several natural settings, such as linear classifiers with norm costs, the induced complexity can be controlled. We begin by showing that such guarantees fail in general - even in simple cases: there exist hypothesis classes of VC dimension $1$ on the real line such that, even under the simplest interval neighborhoods, the induced class has infinite VC dimension. Thus, strategic behavior can turn an easy learning problem into a non-learnable one. To overcome this, we introduce structure via a geometric definability assumption: both the hypothesis class and the cost-induced neighborhood relation can be defined by first-order formulas over $\mathbb{R}_{\mathtt{exp}}$. Intuitively, this means that hypotheses and costs can be described using arithmetic operations, exponentiation, logarithms, and comparisons. This captures a broad range of natural classes and cost functions, including $\ell_p$ distances, Wasserstein distance, and information-theoretic divergences. Under this assumption, we prove that learnability is preserved, with sample complexity controlled by the complexity of the defining formulas.

11.
Nature Medicine 2026-06-09

Adjuvanted inactivated rabies virus-vectored Lassa virus vaccine in healthy adults: a phase 1 trial

Lassa fever causes substantial morbidity and mortality in West Africa, and no licensed vaccine is available. We evaluated LASSARAB, an inactivated rabies virus-vectored Lassa virus (Josiah strain) glycoprotein complex vaccine. We conducted a randomized, controlled, dose-escalation phase 1 trial. Participants (total n = 54) received two intramuscular doses of LASSARAB containing 700 (n = 15), 1,400 (n = 15) or 2,800 (n = 14) relative units of antigen formulated with the TLR-4 agonist 3D-6-acyl PHAD-SE adjuvant, or licensed rabies vaccine control (n = 10), administered 28 days apart. This protocol-defined interim analysis reports the primary safety evaluation and secondary immunogenicity assessments through day 61. There were no prespecified hypotheses or formal power calculations. All primary safety end points demonstrated an acceptable safety profile. After dose 1, local solicited adverse events occurred in 86.7–100.0% of LASSARAB groups and 80% of controls; systemic events in 33.3–71.4% and 60.0% of controls. After dose 2, local solicited adverse events occurred in 66.7–86.7% of LASSARAB groups and 55.6% of controls; systemic events in 53.3–71.4% of LASSARAB groups and 55.6% of controls. Events were predominantly mild and self-limited. Unsolicited adverse events occurred in 28.6–60.0% of LASSARAB groups and 20.0% of controls. No serious adverse event, immune-mediated condition or sensorineural hearing loss occurred. Safety laboratory abnormalities occurred in 13.3–66.7% of LASSARAB groups and 30.0% of controls (14 mild, 6 moderate and none severe). After two doses, Lassa virus GPC IgG ELISA seroconversion (≥fourfold rise) was achieved in 100.0% (44 of 44) of LASSARAB recipients and 0.0% (0 of 10) of controls. Rabies glycoprotein IgG ELISA seroconversion (≥fourfold rise) and neutralizing antibody by rapid fluorescent focus inhibition test (RFFIT) seroprotection (≥0.5 IU ml−1) were also 100% across all groups, including controls. LASSARAB + 3D-6-acyl phosphorylated hexaacyl disaccharide (PHAD)-SE demonstrated a favorable safety profile and immunogenicity against Lassa and rabies viruses. The per-protocol final study report will include safety and durability through day 394. ClinicalTrials.gov identifier NCT06546709 . An interim report of a first-in-human phase 1 trial found an adjuvanted, combination inactivated rabies-vectored, Lassa fever vaccine (LASSARAB + 3D-6-acyl PHAD-SE) to be safe and induced immunogenicity to both Lassa and rabies viruses in healthy participants.

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

Actionable Activation Directions for Detecting and Mitigating Emergent Misalignment Across Language Model Families

Fine-tuning language models on insecure code induces emergent misalignment with poorly understood internal structure. We investigate whether this misalignment corresponds to a causally actionable activation-space direction shared across architectures. Across four instruction-tuned model families (Qwen2.5-1.5B, Gemma-2-2B, Llama-3.2-1B, Ministral-3-3B) finetuned identically, a difference-in-means direction achieves 99.6% separation of aligned and misaligned activations at each model's final layer. Causal steering by subtracting this direction reduces code spillover by 21-51 points, while a secure-code control confirms content specificity. Cross-architecture transfer via ridge regression maps yields large behavioral suppression (up to 46 points) but fails specificity controls as random and orthogonal directions perform comparably. We identify a two-tier specificity structure: within-model directions are causally specific and actionable; cross-model directions are causally real but non-specific. An asymmetric transfer topology emerges, with Gemma and Qwen acting as geometric donors and Llama as a receiver. These findings define the limits of linear cross-architecture correction and recommend within-model probing for auditing.

13.
arXiv (CS.AI) 2026-06-17

How Inference Compute Shapes Frontier LLM Evaluation

arXiv:2606.17930v1 Announce Type: new Abstract: AI evaluations are shifting toward harder tasks that benefit from longer trajectories involving tool use and iterative problem solving. As a result, performance is increasingly sensitive to the amount and allocation of compute available at test time ("inference compute"). Yet many evaluations still report performance at a single restrictive budget, meaning that low scores may reflect the evaluation setup rather than the model's underlying capability. To test this, we evaluate up to 12 frontier language models on seven challenging benchmarks spanning software engineering, mathematics, medicine, and cybersecurity. We use a controlled setup combining three simple inference-scaling interventions: larger token budgets, context compaction, and repeated submission attempts, guided either by the model itself or by minimal correctness feedback. We find three main results. First, larger token budgets substantially improve performance on benchmarks across multiple domains, including cybersecurity, FrontierMath, Humanity's Last Exam, and TerminalBench. Second, fixed-budget evaluations can increasingly understate frontier capability as models advance. Newer models reach higher performance at large budgets, where they unlock harder tasks and solve them more reliably. Third, benchmarks differ in which inference-scaling methods help most: repeated submission broadly improves performance, but the value of larger token budgets, external feedback, and parallel attempts varies by benchmark. Overall, our results show that benchmark scores are protocol-dependent. We therefore argue that evaluations should report capability as a function of inference-time compute, specify protocol choices explicitly, and compare model generations over a large shared compute range at matched budgets, especially in safety- or policy-relevant settings.

14.
arXiv (CS.CL) 2026-06-15

Learning to Hear Hesitation: Continual Learning for Disfluency-Aware ASR

Despite advances in large-scale Automatic Speech Recognition (ASR), disfluent speech remains challenging, as state-of-the-art systems are often optimized to omit disfluencies, leading to information loss and hallucinations. Prior work has focused on verbatim transcription and the integration of disfluency markers, but adapting models on limited datasets can lead to catastrophic forgetting of general-domain knowledge. We address this gap by leveraging continual learning (CL) with explicit disfluency tokens. We first introduce these tokens into a pretrained ASR model to establish stable token mechanisms, and then continue training on additional datasets with varying disfluency distributions. Through a detailed analysis of model dynamics during training, we identify a trade-off between marker learning and ASR performance, and a consistent cross-attention head mechanism shared across CL methods.

15.
arXiv (CS.AI) 2026-06-18

Sparsity Curse: Understanding RLVR Model Parameter Space from Model Merging

arXiv:2606.18521v1 Announce Type: cross Abstract: Reinforcement Learning with Verifiable Reward (RLVR) has emerged as a powerful post-training paradigm that surpasses Supervised Fine-Tuning (SFT) in eliciting reasoning intelligence and resisting catastrophic forgetting. Recent studies further reveal that RLVR induces highly sparse and off-principal parameter updates compared to SFT. This naturally raises the question: does such sparsity make RLVR models more amenable to model merging? If so, model merging would offer a scalable, training-free path to aggregate diverse reasoning capabilities from independently trained RLVR models. Surprisingly, we find the opposite, uncovering a sparsity curse: the sparse RLVR updates are spread farther apart in parameter space, forming near-orthogonal shortcuts that make aggregation inherently fragile. This is likely rooted in the stochasticity of RL optimization and the diversity of emergent reasoning patterns. Unlike SFT models that converge to shared, flat basins and merge naturally, RLVR models suffer severe degradation under standard merging methods. Through systematic empirical analysis of the update geometry, we characterize the mechanisms behind this failure and propose Sensitivity-aware Resolving Merging (SAR-Merging), a merging recipe tailored for the unique structure of RLVR parameter spaces. SAR-Merging resolves conflicts in overlapping update regions via Fisher Information-based sensitivity arbitration, followed by magnitude-aware sparsification and rescaling to preserve fragile reasoning pathways. Experiments on mathematical and coding benchmarks demonstrate that SAR-Merging substantially outperforms existing merging methods on RLVR models, enabling both single-task enhancement and multi-capability fusion.

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

Unsupervised Diffusion Solver for Combinatorial Optimization via Combinatorial Adjoint Matching

arXiv:2605.30920v2 Announce Type: replace Abstract: Diffusion-based neural solvers have shown strong promise for combinatorial optimization (CO), but existing methods typically rely on supervised training with large collections of near-optimal solutions. In this work, we extend adjoint-based trajectory optimization methods to discrete combinatorial domains. We formulate diffusion-based CO as a stochastic control problem over Continuous-Time Markov Chains and introduce discrete adjoint dynamics for propagating optimization signals through discrete generative trajectories. Building on this formulation, we propose Combinatorial Adjoint Matching (CAM), an unsupervised training framework for discrete diffusion solvers with structured and low-variance trajectory-level optimization signals. Empirically, CAM consistently outperforms existing unsupervised diffusion baselines and achieves performance competitive with strong supervised diffusion solvers and even traditional solvers across diverse combinatorial optimization problems. Our code is available at https://github.com/Shengyu-Feng/CAM.

17.
arXiv (CS.AI) 2026-06-11

Robust Instruction Compliance in Cooperative Multi-Agent Reinforcement Learning

arXiv:2605.12655v3 Announce Type: replace Abstract: Multi-agent reinforcement learning (MARL) in real-world use cases may need to adapt to external natural language instructions that interrupt ongoing behavior and conflict with long-horizon objectives. However, conditioning rewards on instructions introduces a fundamental failure mode as Bellman updates couple value estimates across instruction contexts, leading to inconsistent values when instructions interrupt macro-actions. We propose Macro-Action Value Correction for Instruction Compliance (MAVIC), which corrects Bellman backups at instruction boundaries by correcting the incoming instruction objective and restoring the continuation value under the current objective. Unlike reward shaping, MAVIC modifies the bootstrapping target itself, enabling consistent value estimation under stochastic instruction switching within a unified policy. We provide theoretical analysis and an actor-critic implementation, and show that MAVIC achieves high instruction compliance while preserving base task performance in increasingly complex cooperative multi-agent environments.

18.
Nature Biotechnology 2026-06-19

Optimized R2 retroelement complexes for DNA insertion into plant genomes

Traditional approaches for DNA insertion into plant genomes using Agrobacterium tumefaciens result in random integration. Newer genetic engineering methods based on nucleases, prime editors, transposases and recombinases extend capabilities but remain constrained with low efficiencies, off-target integration or limited payload size. Here we adapt the avian Taeniopygia guttata R2 protein (R2Tg) for targeted DNA insertion into plant genomes by engineering R2Tg expression cassettes and RNA payloads carrying intron-disrupted reporters, with optimized ribosomal DNA homology arms and untranslated regions. In Arabidopsis thaliana protoplasts, Nicotiana benthamiana leaves and Solanum lycopersicum seedlings, our R2Tg editor system achieves targeted insertion of full-length payloads ranging from 2.2 kb to 5 kb. In Nicotiana benthamiana leaves, integration occurs, on average, at 1 copy per genome, which is 30 times more efficient than that achieved by Cas9 homology-directed repair. This work establishes an R2Tg ribonucleoprotein platform for targeted DNA insertion into plant genomes, using a multicopy genomic safe-harbor site to enable efficient addition of multikilobase genes. R2 retrotransposons are used to integrate DNA into plant and crop 25S ribosomal DNA sites.

19.
arXiv (CS.CL) 2026-06-11

Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling

Reinforcement learning (RL) has become a key component in modern large language models, yet the rollout stage remains the key bottleneck in RL training pipelines. Although Multi-Token Prediction (MTP) offers a natural solution to accelerate rollouts through speculative decoding, many studies have observed that MTP acceptance rates degrade significantly during RL training, leading to limited speedup performance. To address this bottleneck, we present Bebop, a systematic study of MTP in LLM post-training, and offer practical recipes to integrate MTP into large-scale RL pipelines. First, we reveal that the MTP acceptance rate is fundamentally bounded by the fluctuation of model entropy, which demonstrates a clear negative linear relationship with the rise of entropy in the RL stage. Second, we show that probabilistic rejection sampling largely alleviates the disturbance introduced by entropy in RL compared to greedy draft sampling. We further identify that the conventional MTP training objectives (cross-entropy or KL) are suboptimal in such settings, and therefore we propose a novel end-to-end TV loss that directly optimizes multi-step rejection sampling acceptance rate, yielding ~10% acceptance rate improvements, achieving up to 95% acceptance rates and up to 25% extra inference throughput gains across mathematical reasoning, code generation, and agentic tasks. Third, we test various online MTP training strategies during RL and show that pre-RL MTP training with e2e TV loss and rejection sampling achieves a consistent acceptance rate and speedup throughout the entire RL, eliminating the need for costly online MTP updating. We provide extensive experiments and analysis that validate our findings. Experimental results show our method achieves up to 1.8x end-to-end acceleration in async RL training of Qwen3.5, Qwen3.6, and Qwen3.7 models.

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

ACC: Compiling Agent Trajectories for Long-Context Training

Recent development of agents has renewed demand for long-context reasoning capacity of LLMs. However, training LLMs for this capacity requires costly long-document curation or heuristic context synthesis. We observe that agents produce massive trajectories when solving problems, invoking tools and receiving environment observations across many turns. The evidence needed to answer the original question is thus scattered throughout these turns, requiring integration of distant context segments. Nevertheless, standard agent SFT masks tool responses and only trains turn-level tool selection, creating a supervision blind spot where these scattered signals go unused. We propose Agent Context Compilation (ACC), which converts trajectories from search, software engineering, and database querying agents into long-context QA pairs that combine the original question with tool responses and environment observations gathered across multiple turns, training the model to answer directly without tool use. This makes the dependencies between the question and the evidence explicit, enabling direct supervision of long-context reasoning over distant segments without additional annotation. ACC is a simple but effective approach that can be combined with any existing long-context extension or training method, providing scalable supervised fine-tuning data. We validate ACC on long-range dependency modeling tasks through MRCR and GraphWalks, challenging benchmarks requiring cross-turn coreference resolution and graph traversal over extended contexts. Training Qwen3-30B-A3B with ACC achieves 68.3 on MRCR (+18.1) and 77.5 on GraphWalks (+7.6), results comparable to Qwen3-235B-A22B, while preserving general capabilities on GPQA, MMLU-Pro, AIME, and IFEval. Further mechanism analysis reveals that the ACC-trained model exhibits task-adaptive attention restructuring and expert specialization.

21.
arXiv (CS.CL) 2026-06-16

Lect\=uraAgents: A Multi-Agent Framework for Adaptive Personalized AI-Assisted Learning and Embodied Teaching

Effective personalized AI-assisted learning demands systems that can not only generate accurate learner-specific educational materials, but also dynamically adapt their instruction to diverse learners. However, existing educational agents have primarily focused on lecture content automation and simulations, which often fall short of modelling multimodal and embodied instructional methods tailored for the individual learner. To this end, we propose Lect\=uraAgents - a multi-agent framework that enables personalized learning through end-to-end adaptive embodied teaching. At its core, Lect\=uraAgents mirrors a professor-student relationship, in which a ProfessorAgent leads a collaborative team of specialized subordinate agents through research, planning, review, and embodied delivery of lecture contents that adapt to a learner's needs. The framework offers three main contributions: (1) a hierarchical multi-agent architecture for end-to-end personalized learning; (2) an adaptive embodied teaching mechanism, wherein the ProfessorAgent executes visible and pedagogically motivated teaching actions (e.g., handwrite, highlight, underline, etc.) over contents in a teaching environment; and (3) a Teaching Action-Speech Alignment (TASA) algorithm that employs salience-based heuristics and temporal semantic segmentation to generate coherent teaching action sequences aligned with learner profiles. We evaluate Lect\=uraAgents on diverse courses at high school, undergraduate, and graduate levels using sample-specific rubric-based analysis; with generated lecture materials and teaching actions assessed and validated by expert educators. Experimental results show consistent gains in lecture content quality, embodied teaching quality, assessment, and personalization over existing approaches, positioning Lect\=uraAgents as a pedagogically well-grounded framework for personalized learning at scale.

22.
arXiv (CS.CV) 2026-06-16

CLAD: Constrained Latent Action Diffusion for Vision-Language Procedure Planning

We propose CLAD, a Constrained Latent Action Diffusion model for vision-language procedure planning in instructional videos. Procedure planning is the challenging task of predicting intermediate actions given a visual observation of a start and a goal state. However, future interactive AI systems must also be able to plan procedures using multi-modal input, e.g., where visual observations are augmented with language descriptions. To tackle this vision-language procedure planning task, our method uses a Variational Autoencoder (VAE) to learn the latent representation of actions and observations as constraints and integrate them into the diffusion process. This approach exploits that the latent space of diffusion models already has semantics that can be used. We use the latent constraints to steer the diffusion model to better generate actions. We report extensive experiments on the popular CrossTask, Coin, and NIV datasets and show that our method outperforms state-of-the-art methods by a large margin. By evaluating ablated versions of our method, we further show that the proposed integration of the action and observation representations learnt in the VAE latent space is key to these performance improvements.

23.
arXiv (CS.AI) 2026-06-19

Tri-Info: Generalizable, Interpretable Failure Prediction for VLA Models via Information Theory

arXiv:2606.19998v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models are increasingly deployed across diverse tasks, yet they remain black boxes whose physical interactions can cause irreversible harm, making generalizable and interpretable failure detection essential. We observe that successful and failed rollouts carry systematically different information-theoretic signatures. Building on this, we formalize VLA control as a closed-loop information pipeline and derive the Triple Information-theoretic (Tri-Info) signals that capture whether actions remain diverse, temporally consistent, and coupled to state transitions. Across six VLA models and three benchmark environments, Tri-Info matches the strongest baselines in-domain. Moreover, Tri-Info transfers across architectures, environments, and the sim-to-real gap without retraining, reaching 83\% accuracy on real-world tasks where prior detectors collapse to chance. This establishes Tri-Info as a simple yet powerful method that not only detects failures with strong cross-domain generalization, but also delivers interpretable diagnostics of the underlying failure modes.

24.
arXiv (CS.CL) 2026-06-19

Scalable Training of Spatially Grounded 2D Vision-Language Models for Radiology

We study how to train visually grounded vision-language models (VLMs) for radiology without manual spatial annotations. We introduce RefRad2D, a large-scale bilingual (German/English) dataset of 1.2M CT and MR image-text pairs derived from clinical practice, with task-specific VQA and spatial grounding subsets generated automatically via LLM-based curation and automated segmentation. Trained on this data, our model RadGrounder jointly performs report generation, visual question answering, and spatial grounding via bounding-box detection or segmentation. On external VQA benchmarks (Slake, VQA-RAD), RadGrounder achieves competitive results with specialized medical VLMs. Adding our clinical data to the training mixture improves open-ended VQA over fine-tuning on the downstream datasets alone, showing the transferability of our dataset. Crucially, adding grounding supervision does not degrade language quality, enabling spatially verifiable outputs at no cost to VQA performance.

25.
arXiv (CS.AI) 2026-06-11

Human-Guided Agentic AI for Multimodal Clinical Prediction: Lessons from the AgentDS Healthcare Benchmark

arXiv:2602.19502v2 Announce Type: replace Abstract: Agentic AI systems are increasingly capable of autonomous data science workflows, yet clinical prediction tasks demand domain expertise that purely automated approaches struggle to provide. We investigate how human guidance of agentic AI can improve multimodal clinical prediction, presenting our approach to all three AgentDS Healthcare benchmark challenges: 30-day hospital readmission prediction (Macro-F1 = 0.8986), emergency department cost forecasting (MAE = $465.13), and discharge readiness assessment (Macro-F1 = 0.7939). Across these tasks, human analysts directed the agentic workflow at key decision points, multimodal feature engineering from clinical notes, scanned PDF billing receipts, and time-series vital signs; task-appropriate model selection; and clinically informed validation strategies. Our approach ranked 5th overall in the healthcare domain, with a 3rd-place finish on the discharge readiness task. Ablation studies reveal that human-guided decisions compounded to a cumulative gain of +0.065 F1 over automated baselines, with multimodal feature extraction contributing the largest single improvement (+0.041 F1). We distill three generalizable lessons: (1) domain-informed feature engineering at each pipeline stage yields compounding gains that outperform extensive automated search; (2) multimodal data integration requires task-specific human judgment that no single extraction strategy generalizes across clinical text, PDFs, and time-series; and (3) deliberate ensemble diversity with clinically motivated model configurations outperforms random hyperparameter search. These findings offer practical guidance for teams deploying agentic AI in healthcare settings where interpretability, reproducibility, and clinical validity are essential.