arXiv:2606.04254v1 Announce Type: new
Abstract: This study investigates whether behavioral and performance indicators derived from a Moodle-based learning management system are associated with university students' depression and anxiety in two undergraduate Computer Engineering courses. Using a quantitative observational design, LMS event logs, academic records, and self-reported Beck Depression Inventory-II and Beck Anxiety Inventory scores from 97 students were integrated. A broad set of behavioral and performance indicators spanning temporal engagement, session structure, deadline-related behavior, page-refresh patterns, and LMS navigation was extracted from raw event logs and analyzed using descriptive statistics, independent-samples t-tests with Benjamini-Hochberg FDR correction, effect sizes, and Spearman correlations; inventory scores were confirmed invariant by sex and academic year. Several indicators were significantly associated with depression and anxiety. Higher depression was associated with shifted temporal activity patterns, longer session durations, and shorter homework submission lead times, while higher anxiety was associated with concentrated temporal engagement and session-based differences. These findings suggest that routine LMS data can provide meaningful behavioral signals related to student well-being and may support earlier educational awareness of students who experience mental-health-related strain. At the same time, such indicators should be interpreted as contextual and non-diagnostic markers rather than as substitutes for clinical assessment.
Science Journals
arXiv:2606.05118v1 Announce Type: new
Abstract: This paper examines whether and how artificial intelligence (AI) advances scientific creativity. Drawing on scientific publications, the primary output of researchers, we analyze over one million publications from OpenAlex to investigate the relationship between AI adoption and multiple dimensions of scientific creativity, including novelty (recombinant novelty and object novelty) and impact (3-year short-run citation impact and 10-year long-run citation impact). We find that AI publications are significantly more likely to achieve top-decile creativity relative to non-AI publications, with 5.5 to 10.2 percentage point higher likelihood to rank in the top creativity decile. Critically, we uncover substantial heterogeneity across AI research modes. Tool-oriented AI research, which applies existing AI models to domain tasks, is associated with the largest gains in recombinant-based creativity, while Adaptation-oriented AI research, modifying AI models for domain-specific problems, is associated with relatively higher object-based creativity. These findings reveal that AI does not advance science through a single mechanism but through structurally distinct creative pathways that depend on how AI is incorporated into the research process. Our results contribute to ongoing debates about AI's role in science and carry direct implications for research evaluation and science policy, highlighting the need for assessment frameworks that can distinguish between recombinant and conceptual forms of creativity and that recognize how different modes of AI adoption produce fundamentally different types of scientific contribution.
arXiv:2606.04135v1 Announce Type: new
Abstract: Time series forecasting relies on historical patterns, but real-world series often exhibit non-stationarity and regime shifts that challenge fully parametric forecasters. Inspired by Retrieval-Augmented Generation (RAG), recent work augments forecasters by retrieving relevant historical segments and using them as external evidence at inference time. However, due to the intrinsic non-stationarity of real-world time series, a highly similar past segment does not necessarily imply a similar future, rendering similarity-only retrieval brittle and prone to redundancy. We propose Stationarity-Aware Retrieval-Augmented Time Series Forecasting (SARAF), a framework that adaptively balances relevance and diversity in retrieval. SARAF first forms a candidate pool via temporal similarity with time-aligned enhancement, then applies a diversity-aware selection strategy to cover heterogeneous historical regimes, with the diversification strength automatically modulated by dataset-level stationarity. Moreover, SARAF uses stationarity-aware aggregation to fuse the retrieved futures. Extensive experiments on eight real-world datasets show that SARAF achieves competitive forecasting performance and improves average accuracy and robustness over strong baselines, with particularly clear benefits under challenging non-stationary settings. Code: https://github.com/ShiqiaoZhou/SARAF.
arXiv:2602.08142v2 Announce Type: replace
Abstract: Machine learning applications require fast and reliable per-sample uncertainty estimation. A common approach is to use predictive distributions from Bayesian or approximation methods and additively decompose uncertainty into aleatoric (i.e., data-related) and epistemic (i.e., model-related) components. However, additive decomposition has recently been questioned, with evidence that it breaks down when using finite-ensemble sampling and/or mismatched predictive distributions. This paper introduces Variance-Gated Ensembles (VGE), an intuitive, differentiable framework that injects epistemic sensitivity via a signal-to-noise gate computed from ensemble statistics. VGE provides: (i) a Variance-Gated Margin Uncertainty (VGMU) score that couples decision margins with ensemble predictive variance; and (ii) a Variance-Gated Normalization (VGN) layer that generalizes the variance-gated uncertainty mechanism to training via per-class, learnable normalization of ensemble member probabilities. We derive closed-form vector-Jacobian products enabling end-to-end training through ensemble sample mean and variance. VGE matches or exceeds state-of-the-art information-theoretic baselines while remaining computationally efficient. As a result, VGE provides a practical and scalable approach to epistemic-aware uncertainty estimation in ensemble models.
arXiv:2602.06911v2 Announce Type: replace
Abstract: As increasingly capable open-weight large language models (LLMs) are deployed, improving their tamper resistance against unsafe modifications, whether accidental or intentional, becomes critical to minimize risks. However, there is no standard approach to evaluate tamper resistance. Varied datasets, metrics, and tampering configurations make it difficult to compare safety, utility, and robustness across different models and defenses. To address this, we introduce TamperBench, the first unified framework to systematically evaluate the tamper resistance of LLMs. TamperBench (i) curates a repository of state-of-the-art weight-space fine-tuning attacks, latent-space representation attacks, and alignment-stage defenses; (ii) enables realistic adversarial evaluation through systematic hyperparameter sweeps per attack-model pair; and (iii) provides both safety and utility evaluations. We use TamperBench to evaluate 21 open-weight LLMs, including defense-augmented variants, across nine tampering threats using standardized safety and capability metrics with hyperparameter sweeps per model-attack pair. The results provide insights including effects of post-training on tamper resistance, that jailbreak-tuning is typically the most severe attack, and that current alignment-stage defenses largely fail to withstand attack sweeps. Code is available at https://github.com/criticalml-uw/TamperBench.
arXiv:2606.04189v1 Announce Type: new
Abstract: Aspect-Based Sentiment Analysis (ABSA) requires high-quality datasets to train reliable models. However, existing annotation tools treat output as flat files, leaving researchers to manually consolidate multi-annotator data, reconstruct relational structures, and compute reliability metrics through custom scripts. This paper introduces ACAT (Aspect-based sentiment analysis Collaborative Annotation Tool), a web-based platform natively supporting four ABSA workflows: (1) Aspect-Category Sentiment Analysis, (2) Clause-Level Segmentation, (3) Aspect-Term Sentiment Analysis with character-level position tracking, and (4) Aspect Sentiment Triplet Extraction with dual span offset preservation. Its core contribution is an automated Extract, Transform, Load (ETL) pipeline that aligns collaborative annotations and computes Inter-Annotator Agreement (IAA) metrics directly at export, yielding training-ready datasets. In a preliminary validation on 1,002 restaurant reviews with two annotators of differing expertise, ACAT achieves a median annotation time of 31.58 seconds and a raw IAA ranging from 0.78 to 0.86 across all tasks.
arXiv:2602.06310v2 Announce Type: replace
Abstract: With the rapid rise of AI coding agents, the fundamental premise of what it means to be a software engineer is in question. In this vision paper, we examine what it means for an AI agent to be considered a software engineer and then critically think about what makes such an agent trustworthy. Grounded in established definitions of SE (SE) and informed by recent research on agentic AI systems, we conceptualise AI software engineers as participants in human-AI SE teams composed of human software engineers and AI agents, and we distinguish trustworthiness as a key property of these systems and actors rather than a subjective human attitude. Extending on historical perspectives and emerging visions, we identify key dimensions that contribute to the trustworthiness of AI software engineers, spanning technical quality, transparency and accountability, epistemic humility, and societal and ethical alignment. Beyond defining these dimensions, we address a critical but underexplored challenge: how trustworthiness can be operationalised in practice. We therefore introduce the notion of evidence-centric inspection, arguing that developers should evaluate selective signals and justifications of trustworthiness rather than raw outputs, and we outline implications for rethinking verification, validation, and code review in human-AI SE teams.
arXiv:2602.05725v3 Announce Type: replace
Abstract: Muon updates matrix parameters via the matrix sign of the gradient and has shown strong empirical gains, yet its dynamics and scaling behavior remain unclear in theory. We study Muon in a linear associative memory model with softmax retrieval and a hierarchical frequency spectrum over query-answer pairs, with and without label noise. In this setting, we show that Gradient Descent (GD) learns frequency components at highly imbalanced rates, leading to slow convergence bottlenecked by low-frequency components. In contrast, the Muon optimizer mitigates this imbalance, leading to faster and more uniform progress. Specifically, in the noiseless case, Muon achieves an exponential speedup over GD; in the noisy case with a power-law frequency spectrum, we derive Muon's scaling law and demonstrate its superior scaling efficiency over GD. Furthermore, we show that Muon can be interpreted as an implicit matrix preconditioner arising from adaptive task alignment and block-symmetric gradient structure. In contrast, the preconditioner with coordinate-wise sign operator could match Muon under oracle access to unknown task representations, which is infeasible for SignGD in practice. Experiments on synthetic long-tail classification and LLaMA-style pre-training corroborate the theory.
arXiv:2604.04970v2 Announce Type: replace
Abstract: Atomic-resolution imaging on molten metal/solid interfaces at temperatures above 200 {\deg}C was achieved using a high-temperature, high-speed atomic force microscope (AFM) equipped with a qPlus sensor. A tip-scanning high-speed Quadpod scanner for a large mass load of qPlus sensor (2.3 g) was developed to enhance thermal drift tolerance by high-speed scanning and thermal insulation from the heated specimen. This scanner has dominant resonant frequencies of 7.05 kHz (lateral) / 29.7 kHz (vertical) without a load. In addition, the Hybrid-loop frequency demodulation technique for low-resonant-frequency ($f_0$) sensors with a wider bandwidth than conventional phase-locked loop was also established, providing a demodulation bandwidth of $B_{\Delta f_\mathrm{inst}}\sim 0.26 f_0$ without exceeding the theoretical noise of the input deflection signal. Combining these techniques enabled atomic-resolution imaging on the molten $\mathrm{Ga/PtGa_x}$ interface at $\sim$210 {\deg}C. The topographic images obtained at $\sim$210 {\deg}C showed a relatively low-symmetry surface with an oblique lattice with a superstructure, which differed from the primitive rectangular lattice observed in the non-heated sample left at room temperature for 96 h. This demonstrates that the developed high-temperature, high-speed AFM techniques for qPlus sensors enable visualization of non-aqueous liquid/solid interfaces above 200 {\deg}C at atomic resolution, which has various potential applications, such as injection modeling, soldering, and the fabrication of liquid-metal-based catalysts.
arXiv:2604.04974v3 Announce Type: replace
Abstract: Video is a scalable observation of physical dynamics: it captures how objects move, how contact unfolds, and how scenes evolve under interaction -- all without requiring robot action labels. Yet translating this temporal structure into reliable robotic control remains an open challenge, because video lacks action supervision and differs from robot experience in embodiment, viewpoint, and physical constraints. This survey reviews methods that exploit non-action-annotated temporal video to learn control interfaces for robotic manipulation. We introduce an interface-centric taxonomy organized by where the video-to-control interface is constructed and what control properties it enables, identifying three families: direct video-action policies, which keep the interface implicit; latent-action methods, which route temporal structure through a compact learned intermediate; and explicit visual interfaces, which predict interpretable targets for downstream control. For each family, we analyze control-integration properties -- how the loop is closed, what can be verified before execution, and where failures enter. A cross-family synthesis reveals that the most pressing open challenges center on the robotics integration layer -- the mechanisms that connect video-derived predictions to dependable robot behavior -- and we outline research directions toward closing this gap.
arXiv:2606.04248v1 Announce Type: new
Abstract: Decentralized rigid formation flocking requires a swarm of autonomous agents to maintain a predetermined geometric configuration while moving, relying solely on local sensing and communication. However, existing decentralized control methods struggle to maintain strict inter-agent distance constraints in cluttered environments, often suffering from local minima deadlocks, high frequency control oscillations, or limited flexibility during obstacle navigation, resulting in low success rate. To address these limitations, we propose Rigid Swarm Control (RSC), a decentralized control framework for large-scale rigid formation flocking. To escape local minima via robust long-term planning while ensuring short-term safety, RSC integrates finite-horizon trajectory predictions with a reactive artificial potential field (APF) safety controller within a hybrid architecture. Furthermore, to accelerate formation reassembly after obstacle traversal without interrupting task execution, RSC introduces an online leader-follower reconfiguration mechanism based on stable role exchange. Extensive evaluations in challenging cluttered environments with 25 UAVs demonstrate that RSC reliably unifies rigid formation maintenance, obstacle avoidance, and target tracking. Under strict success criteria - collision-free operation with a maximum relative edge-length error below 10%, RSC achieves an 83% success rate, significantly outperforming existing heuristic and learning-based baselines that fall below 5%.
arXiv:2606.04246v1 Announce Type: new
Abstract: Automatic generation of RTL code for digital hardware designs remains challenging due to long-horizon reasoning, multi-step dependencies, and strict correctness constraints in Verilog and VHDL. We present StepPRM-RTL, a novel framework that combines stepwise trajectory modeling, process-reward modeling (PRM), and retrieval-augmented fine-tuning (RAFT) to enhance both the functional correctness and reasoning fidelity of LLM-based RTL code generation. StepPRM-RTL constructs stepwise reasoning trajectories from canonical solutions, where each step contains a rationale and incremental code modification. A Process Reward Model (PRM) evaluates intermediate steps, providing dense feedback that guides reinforcement-style updates during RAFT fine-tuning. Monte Carlo Tree Search (MCTS) explores alternative reasoning paths, enriching the training dataset with high-quality trajectories. This integration of stepwise and outcome-aware rewards allows the model to learn both how and why to construct correct RTL, improving long-horizon reasoning beyond standard supervised or outcome-based training. Experimental evaluation on benchmark Verilog and VHDL datasets demonstrates that StepPRM-RTL outperforms the best prior methods by over 10\% in functional correctness and reasoning fidelity metrics. Ablation studies confirm that the combination of PRM-guided rewards and stepwise trajectory exploration is key to its performance. StepPRM-RTL generalizes across RTL languages and provides a scalable framework for high-fidelity, interpretable code generation, establishing a new standard for LLM-assisted hardware design automation.
arXiv:2606.04125v1 Announce Type: new
Abstract: The Poisson Nernst Planck PNP system constitutes a canonical stiff coupled PDE problem where the charge density prefactor produces extreme coefficient ratios and the electric double layer imposes sharp boundary layers. Physics informed neural networks PINNs are appealing here because they require no mesh and differentiate through the physics automatically. Spectral bias and multi task loss imbalance however have limited their accuracy on stiff PNP systems. We present the first systematic data free benchmark of eleven PINN configurations organised into four strategy groups on a physically parametrised one dimensional PNP model for a lithium symmetric cell implemented within NVIDIA PhysicsNeMo Sym and validated against a finite volume method FVM reference. Root mean square errors RMSE span across architectures. The balanced residual decay rate BRDR scheme matches Neural Tangent Kernel NTK performance for concentration fields while reducing mean wall clock time making it the preferable strategy under compute constraints. Loss landscape geometry corroborates the RMSE ranking. We release an open source PhysicsNeMo Sym implementation for reuse on stiff coupled PDE problems in computational mechanics.
arXiv:2606.04243v1 Announce Type: new
Abstract: We study the integral equation $\int_0^m \eta\rho(\eta)/(C-\eta)\,d\eta = 1$ with $C>m$, where $\rho$ is a $C^1$ probability density on $[0,M]$ vanishing polynomially at $\eta=M$. Setting $\mathcal{I}^+(m) := \lim_{C \downarrow m}\int_0^m \eta\rho(\eta)/(C-\eta)\,d\eta$ and $\Omega := \{m \in (0,M) : \mathcal{I}^+(m) > 1\}$, the equation determines $C$ implicitly as a function of $m$ on $\Omega$, and our object of study is the dimensionless ratio $\beta(m) := C(m)/m$.
Writing $h(\eta) := \eta\rho(\eta)$, our main theorem establishes openness of $\Omega$, $C^1$-smoothness of $\beta$, a sign formula identifying $\beta'(m)$ with a positively-weighted integral of $dh/d\ln\eta$, transfer of monotonicity from $h$ to $\beta$, and existence of an interior critical point of $\beta$ when $h$ is unimodal and two technical hypotheses hold.
Numerically, $\beta$ has a single critical point in seven log-concave test densities (mostly Beta-type), in support of a separate uniqueness conjecture. A bimodal density that violates both unimodality and log-concavity exhibits three critical points; this shows that dropping the two hypotheses jointly admits multiple critical points, but does not separate their roles.
arXiv:2602.05657v2 Announce Type: replace
Abstract: The study of tail behaviour of SGD-induced processes has been attracting a lot of interest, due to offering strong guarantees with respect to individual runs of an algorithm. While many works provide high-probability guarantees, quantifying the error rate for a fixed probability threshold, there is a lack of work directly studying the probability of failure, i.e., quantifying the tail decay rate for a fixed error threshold. Moreover, existing results are of finite-time nature, limiting their ability to capture the true long-term tail decay which is more informative for modern learning models, typically trained for millions of iterations. Our work closes these gaps, by studying the long-term tail decay of SGD-based methods through the lens of large deviations theory, establishing several strong results in the process. First, we provide an upper bound on the tails of the gradient norm-squared of the best iterate produced by (vanilla) SGD, for non-convex costs and bounded noise, with long-term decay at rate $e^{-t/\log(t)}$. Next, we relax the noise assumption by considering clipped SGD (c-SGD) under heavy-tailed noise with bounded moment of order $p \in (1,2]$, showing an upper bound with long-term decay at rate $e^{-t^{\beta_p}/\log(t)}$, where $\beta_p = \frac{4(p-1)}{3p-2}$ for $p \in (1,2)$ and $e^{-t/\log^2(t)}$ for $p = 2$. Finally, we provide lower bounds on the tail decay, at rate $e^{-t}$, showing that our rates for both SGD and c-SGD are tight, up to poly-logarithmic factors. Notably, our results demonstrate an order of magnitude faster long-term tail decay compared to existing work based on finite-time bounds, which show rates $e^{-\sqrt{t}}$ and $e^{-t^{\beta_p/2}}$, $p \in (1,2]$, for SGD and c-SGD, respectively. As such, we uncover regimes where the tails decay much faster than previously known, providing stronger long-term guarantees for individual runs.
arXiv:2602.04613v2 Announce Type: replace
Abstract: Mechanistic Interpretability (MI) seeks to explain how neural networks implement their capabilities, but the scale of Large Language Models (LLMs) has limited prior MI work in Machine Translation (MT) to word-level analyses. We study sentence-level MT from a mechanistic perspective by analyzing attention heads to understand how LLMs internally encode and distribute translation functions. We decompose MT into two subtasks: producing text in the target language (i.e. target language identification) and preserving the input sentence's meaning (i.e. sentence equivalence). Across three families of open-source models and 20 translation directions, we find that distinct, sparse sets of attention heads specialize in each subtask. Based on this insight, we construct subtask-specific steering vectors and show that modifying just 1% of the relevant heads enables instruction-free MT performance comparable to instruction-based prompting, while ablating these heads selectively disrupts their corresponding translation functions.
arXiv:2602.03920v2 Announce Type: replace
Abstract: Robot Foundation Models (RFMs) represent a promising approach to developing general-purpose home robots. Given the broad capabilities of RFMs, users will inevitably ask an RFM-based robot to perform tasks that the RFM was not trained or evaluated on. In these cases, it is crucial that users understand the risks associated with attempting novel tasks due to the relatively high cost of failure. Furthermore, an informed user who understands an RFM's capabilities will know what situations and tasks the robot can handle. In this paper, we study how non-roboticists interpret performance information from RFM evaluations. These evaluations typically report task success rate (TSR) as the primary performance metric. While TSR is intuitive to experts, it is necessary to validate whether novices also use this information as intended. Toward this end, we conducted a study in which users saw real evaluation data, including TSR, failure case descriptions, and videos from multiple published RFM research projects. The results highlight that non-experts not only use TSR in a manner consistent with expert expectations but also highly value other information types, such as failure cases that are not often reported in RFM evaluations. Furthermore, we find that users want access to both real data from previous evaluations of the RFM and estimates from the robot about how well it will do on a novel task.
arXiv:2602.03542v2 Announce Type: replace
Abstract: Large language models (LLMs) are trained and tested extensively on symbolic representations such as code and graphs, yet real-world user tasks are often specified in natural language. To what extent can LLMs generalize across these representations? Here, we approach this question by studying isomorphic tasks involving procedures represented in code, graphs, and natural language (e.g., scheduling steps in planning). We find that training LLMs with popular post-training methods on graphs or code data alone does not reliably generalize to corresponding natural language tasks, while training solely on natural language can lead to inefficient performance gains. To address this gap, we propose a two-stage reinforcement learning curriculum that first trains on symbolic, then natural language data. The curriculum substantially improves model performance across model families and tasks. Remarkably, a 1.5B Qwen model trained by our method can closely match zero-shot GPT-4o in naturalistic planning. Finally, our analysis suggests that successful cross-representation generalization can be interpreted as a form of generative analogy, which our curriculum effectively encourages. The dataset and code used in this paper can be found \href{https://github.com/fangru-lin/procedure_generalization_llm}{here}.
arXiv:2606.04832v1 Announce Type: new
Abstract: The Belt & Road Initiative (BRI) is a concerted effort from Ministries under the People's Republic of China (PRC) to diplomatically and economically impose its will upon other nations. El Salvador is a US partner and a beneficiary of foreign investment under the BRI. Recent changes to Salvadoran law do not address the implied risks to the nation's supply-chain and cyber infrastructure. This work addresses the gap by exploring previously limited analysis on BRI activities, its intersection with Salvadoran law, and the national security risks introduced by supply-chain reliance from the BRI. This exploratory study examined a portion of the William & Mary AidData dataset filtered on El Salvador, social media posts, news articles, white papers, and law published by the Legislative Assembly of El Salvador. The analysis suggests that the BRI poses a national security and supply-chain risk to El Salvador through influential-subterfuge, loss of digital sovereignty, which contradicts the State Cybersecurity Agency (ACE) and existing Salvadoran laws. This study provides a foundational understanding and regional context for future research.
arXiv:2606.04196v1 Announce Type: new
Abstract: We describe a design pattern and concrete implementation for embedding distributed approximate nearest neighbor indexes inside the Apache Iceberg table format, using the Puffin sidecar file as the storage container and the snapshot summary as the binding mechanism. Modern analytical query engines increasingly adopt a compute disaggregated architecture: executors are stateless, scale elastically, and read all data from object storage. Adding vector similarity search to such an engine traditionally requires a dedicated index storage layer with its own lifecycle, consistency model, and operational surface breaking the disaggregation in variant. We show that the Puffin format, originally introduced portable level statistics and deletion vectors, is sufficient to carry full Vamana graphs at billion vector scale, and that linking these blobs through the existing statistics file snapshot summary property reduces ANN index management to standard Iceberg snapshot operations. We present a binary layout for sharded graph indexes inside Puffin, a coordinator executor protocol for distributed index build, probe, and incremental refresh, the integration into the existing optimistic-concurrency commit path of an Iceberg REST catalog, and a tiered probe strategy that places small centroid indexes on the coordinator and large DiskANN graphs on executor SSDs. The pattern inherits atomicity, time travel, multi engine read ability, and orphan file garbage collection from the table format at zero implementation cost. We discuss the recall/latency trade-offs introduced by the independent-shard design and quantify projected query and build performance for tables up to 109 vectors. Our implementation extends FlockDB, a distributed MPP engine built on DuckDB.
arXiv:2606.04197v1 Announce Type: new
Abstract: How much should an LLM agent remember, and how should multi-agent systems be connected when trying to reach consensus? We show these two design choices interact in a way that flips the sign of memory's effect on coordination. Across 432 simulation runs of a networked Naming Game on eight fixed 16-agent topologies, we vary memory depth and network structure. Longer memory slows the time to reach steady state in decentralized networks but accelerates it in centralized ones; the same parameter pushes the system in opposite directions depending on topology. Critically, "faster settling" in centralized networks means locking in to a fragmented plateau more quickly, not reaching system-wide consensus, which can be used to generate diverging opinions. We further document a memory-mediated speed-unity trade-off: centralized networks consistently preserve more competing conventions than decentralized networks, but their settling speed depends sharply on memory. At the agent level, within-network analyses show that high-betweenness bridges suffer a brokerage penalty while agents in locally clustered neighborhoods achieve higher coordination success. Finally, in search of analytically tractable generative mechanisms, we find that agents' choices are well captured by Fictitious Play, indicating belief-based rather than reward-based adaptation. The practical implication: memory depth and communication topology should be co-designed, not optimized in isolation.
arXiv:2602.02405v2 Announce Type: replace
Abstract: Improving the reasoning capabilities of large language models (LLMs) typically relies either on the model's ability to sample a correct solution to be reinforced or the existence of a stronger model able to solve the problem. However, many difficult problems remain intractable for even current frontier models, preventing the extraction of valid training signals. A promising alternative is to leverage high-quality expert human solutions, yet naive imitation of this data fails because it is fundamentally out-of-distribution: expert solutions are typically didactic, containing implicit reasoning gaps intended for human readers rather than computational models. Furthermore, high-quality expert solutions are expensive, necessitating generalizable, sample-efficient training methods. We propose Distribution Aligned Imitation Learning (DAIL), a two-step self-distillation method that bridges the distributional gap by first transforming expert solutions into detailed, in-distribution reasoning traces and then applying a contrastive objective to focus learning on expert insights and methodologies. We find that DAIL can leverage fewer than 1000 high-quality expert solutions to achieve up to 31% pass@128 gains on Qwen2.5-Instruct and Qwen3, double reasoning efficiency, and enable out-of-domain generalization.
arXiv:2602.01672v2 Announce Type: replace
Abstract: Search-augmented reasoning agents interleave multi-step reasoning with external retrieval, but uncontrolled retrieval can introduce redundant evidence, saturate the context, and destabilize reinforcement learning (RL). Existing outcome-based RL methods provide only sparse terminal rewards, offering limited guidance for intermediate information-acquisition decisions. We propose DeepControl, an adaptive information-control framework based on information utility, a state-dependent estimate of the marginal value of retrieved evidence. The framework regulates information acquisition along two axes: extent, i.e., whether retrieval should continue, and resolution, i.e., how much retrieved detail should be exposed. It implements these controls through retrieval-continuation guidance, hierarchical granularity control, and an annealed control-forcing scheme. This enables the policy to internalize effective acquisition behavior during training and operate without external control at test time. Across seven benchmarks, DeepControl consistently outperforms strong RL and retrieval baselines without explicit information control; compared with Search-R1, it improves average performance by +9.4 and +8.6 points on Qwen2.5-7B and Qwen2.5-3B, respectively. Additional analyses show improved search effectiveness, training stability, and evidence utilization.
arXiv:2606.02597v1 Announce Type: new
Abstract: The development of brain-computer interfaces (BCIs) based on electroencephalograms (EEGs) has advanced significantly mainly to machine learning. Although the majority of earlier research has been on increasing classification accuracy, relatively little focus has been placed on security and robustness. According to recent research, EEG-based BCIs are susceptible to adversarial attacks, which can cause misdiagnosis due to minute, well-crafted disturbances. Evaluating model robustness against such perturbations is therefore critical for ensuring reliable deployment. In this study, we propose a lightweight custom Convolutional Neural Network (CNN) architecture to investigate adversarial robustness in EEG-based BCIs. The suggested method is assessed using two EEG datasets and contrasted with three novel CNN models tailored to EEG, namely EEGNet, DeepConvNet, and SleepEEGNet, under gradient-based adversarial attack scenarios. According to experimental findings, the suggested model continuously performs better in classification under adversarial perturbations compared to baseline models, indicating improved robustness. These findings highlight the potential of lightweight architectures for enhancing the reliability of EEG-based BCI systems under adversarial conditions.
arXiv:2606.02884v1 Announce Type: new
Abstract: Reward guidance algorithms steer a learned generative process toward the reward-tilted measure at inference time. While empirically powerful, these methods are prone to reward hacking: the guided model over-optimizes the reward at the cost of fidelity to the learned distribution. Prior work has attributed this to the complexity of neural reward functions or implicit biases in diffusion training, but its fundamental origins remain poorly understood. We show that reward hacking arises from an approximation made in most practical implementations of reward-guided diffusion -- finite-particle plug-in estimation of the Doob h-function -- even in the simplest non-trivial settings of Gaussian and Gaussian mixture targets with quadratic rewards. In closed form, we isolate two distinct failure modes of the plug-in estimator: it leads to reward hacking within each mode and it cannot select high-reward modes. We propose a closed-form reward damping schedule that corrects the within-mode bias with no additional compute, and clarify the role of best-of-n sampling in compensating for the mode selection failure. Experiments on Gaussian mixture targets, a 2D checkerboard, and FLUX.1 text-to-image generation confirm that our theoretical insights carry over to practical settings.