To advance time series forecasting (TSF), various methods have been proposed to improve prediction accuracy, evolving from statistical techniques to data-driven deep learning architectures. Despite their effectiveness, most existing methods still adhere to a fast thinking paradigm-relying on extracting historical patterns and mapping them to future values as their core modeling philosophy, lacking an explicit thinking process that incorporates intermediate time series reasoning. Meanwhile, emerging slow-thinking LLMs (e.g., OpenAI-o1) have shown remarkable multi-step reasoning capabilities, offering an alternative way to overcome these issues. However, prompt engineering alone presents several limitations - including high computational cost, privacy risks, and limited capacity for in-depth domain-specific time series reasoning. To address these limitations, a more promising approach is to train LLMs to develop slow thinking capabilities and acquire strong time series reasoning skills. For this purpose, we propose Time-R1, a two-stage reinforcement fine-tuning framework designed to enhance multi-step reasoning ability of LLMs for time series forecasting. Specifically, the first stage conducts supervised fine-tuning for warmup adaptation, while the second stage employs reinforcement learning to improve the model's generalization ability. Particularly, we design a fine-grained multi-objective reward specifically for time series forecasting, and then introduce GRIP (group-based relative importance for policy optimization), which leverages non-uniform sampling to further encourage and optimize the model's exploration of effective reasoning paths. Experiments demonstrate that Time-R1 significantly improves forecast performance across diverse datasets.

Graph neural networks (GNN) have shown promising results for several domains such as materials science, chemistry, and the social sciences. GNN models often contain millions of parameters, and like other neural network (NN) models, are often fed only a fraction of the graphs that make up the training dataset in batches to update model parameters. The effect of batching algorithms on training time and model performance has been thoroughly explored for NNs but not yet for GNNs. We analyze two different batching algorithms for graph-based models, namely static and dynamic batching for two datasets, the QM9 dataset of small molecules and the AFLOW materials database. Our experiments show that changing the batching algorithm can provide up to a 2.7x speedup, but the fastest algorithm depends on the data, model, batch size, hardware, and number of training steps run. Experiments show that for a select number of combinations of batch size, dataset, and model, significant differences in model learning metrics are observed between static and dynamic batching algorithms.

Large language models enable inexpensive AI-generated annotations, but using them reliably for causal inference remains challenging. Naively pooling AI and human data induces bias, while existing methods such as Prediction-Powered Inference (PPI; Angelopoulos et al., 2023a) treat AI outputs as proxies of true labels -- an assumption often violated for generative model outputs in practice. We propose Generative Augmented Inference (GAI), a framework that treats AI outputs as general, potentially high-dimensional informative features for learning human labels rather than as surrogates. GAI flexibly models this relationship using nonparametric methods, enabling consistent estimation and valid inference from combined human and AI data. We establish asymptotic normality and show that, under random labeling, GAI strictly improves asymptotic efficiency over human-data-only estimation whenever AI outputs are informative for true labels. Empirical studies on real-world datasets demonstrate that GAI significantly reduces estimation error and improves confidence interval quality across diverse generative data sources relative to human-only and PPI-based estimation.

This paper introduces the Kernel Neural Operator (KNO), a provably convergent operator-learning architecture that utilizes compositions of deep kernel-based integral operators for function-space approximation of operators (maps from functions to functions). The KNO decouples the choice of kernel from the numerical integration scheme (quadrature), thereby naturally allowing for operator learning with explicitly-chosen trainable kernels on irregular geometries. On irregular domains, this allows the KNO to utilize domain-specific quadrature rules. To help ameliorate the curse of dimensionality, we also leverage an efficient dimension-wise factorization algorithm on regular domains. More importantly, the ability to explicitly specify kernels also allows the use of highly expressive, non-stationary, neural anisotropic kernels whose parameters are computed by training neural networks. We present universal approximation theorems showing that both the continuous and fully discretized KNO are universal approximators on operator learning problems. Numerical results demonstrate that on existing benchmarks the training and test accuracy of KNOs is closely comparable to or higher than that of popular neural operators while typically using an order of magnitude fewer trainable parameters, with the more expressive kernels proving important to attaining high accuracy. KNOs thus facilitate low-memory, geometrically-flexible, deep operator learning, while retaining the implementation simplicity and transparency of traditional kernel methods from both scientific computing and machine learning.

Although autonomous underwater vehicles promise the capability of marine ecosystem monitoring, their deployment is fundamentally limited by the difficulty of controlling vehicles under highly uncertain and non-stationary underwater dynamics. To address these challenges, we employ a data-driven reinforcement learning approach to compensate for unknown dynamics and task variations. Traditional single-task reinforcement learning has a tendency to overfit the training environment, thus, limit the long-term usefulness of the learnt policy. Hence, we propose to use a contextual multi-task reinforcement learning paradigm instead, allowing us to learn controllers that can be reused for various tasks, e.g., detecting oysters in one reef and detecting corals in another. We evaluate whether contextual multi-task reinforcement learning can efficiently learn robust and generalisable control policies for autonomous underwater reef monitoring. We train a single context-dependent policy that is able to solve multiple related monitoring tasks in a simulated reef environment in HoloOcean. In our experiments, we empirically evaluate the contextual policies regarding sample-efficiency, zero-shot generalisation to unseen tasks, and robustness to varying water currents. By utilising multi-task reinforcement learning, we aim to improve the training effectiveness, as well as the reusability of learnt policies to take a step towards more sustainable procedures in autonomous reef monitoring.

To encourage diverse exploration in reinforcement learning (RL) for large language models (LLMs) without compromising accuracy, we propose Policy Split, a novel paradigm that bifurcates the policy into normal and high-entropy modes with a high-entropy prompt. While sharing model parameters, the two modes undergo collaborative dual-mode entropy regularization tailored to distinct objectives. Specifically, the normal mode optimizes for task correctness, while the high-entropy mode incorporates a preference for exploration, and the two modes learn collaboratively. Extensive experiments demonstrate that our approach consistently outperforms established entropy-guided RL baselines across various model sizes in general and creative tasks. Further analysis reveals that Policy Split facilitates dual-mode exploration, where the high-entropy mode generates distinct behavioral patterns to the normal mode, providing unique learning signals.

Traditional neural network models for intent inference rely heavily on observable states and struggle to generalize across diverse tasks and dynamic environments. Recent advances in Vision Language Models (VLMs) and Vision Language Action (VLA) models introduce common-sense reasoning through large-scale multimodal pretraining, enabling zero-shot performance across tasks. However, these models still lack explicit mechanisms to represent and update belief, limiting their ability to reason like humans or capture the evolving human intent over long-horizon. To address this, we propose a belief-aware VLM framework that integrates retrieval-based memory and reinforcement learning. Instead of learning an explicit belief model, we approximate belief using a vector-based memory that retrieves relevant multimodal context, which is incorporated into the VLM for reasoning. We further refine decision-making using a reinforcement learning policy over the VLM latent space. We evaluate our approach on publicly available VQA datasets such as HD-EPIC and demonstrate consistent improvements over zero-shot baselines, highlighting the importance of belief-aware reasoning.

Auto-regressive models (ARMs) have established a dominant paradigm in language modeling. However, their strictly sequential sampling paradigm imposes fundamental constraints on both inference efficiency and modeling flexibility. To address these limitations, diffusion-based large language models (dLLMs) have been proposed, offering the potential for parallel sampling and flexible language modeling. Despite these advantages, current dLLMs sampling strategies rely primarily on token level information, which fails to account for global sequence structure and often yields suboptimal results. In this paper, we study the sampling order selection problem from the perspective of log-likelihood maximization. We show that this problem is NP-hard and propose an optimal sampling-rank-based approximation that makes the objective computationally tractable. We further prove that the tractable objective is optimized by sampling tokens in descending order of their attention-matrix column sums. This finding provides a principled justification for attention-guided sampling and offers a theoretically grounded alternative to greedy search. We instantiate this theoretical insight in a new training-free sampling algorithm, termed Attn-Sampler, and further propose dynamic attention thresholding for practical acceleration. Extensive experiments across multiple benchmarks validate the effectiveness of our proposed method, demonstrating that it achieves superior generation quality while enhancing the sampling parallelism.

Sequential experimental design under expensive, gradient-free objectives is a central challenge in computational statistics: evaluation budgets are tightly constrained and information must be extracted efficiently from each observation. We propose \textbf{ALMAB-DC}, a GP-based sequential design framework combining active learning, multi-armed bandits (MAB), and distributed asynchronous computing for expensive black-box experimentation. A Gaussian process surrogate with uncertainty-aware acquisition identifies informative query points; a UCB or Thompson-sampling bandit controller allocates evaluations across parallel workers; and an asynchronous scheduler handles heterogeneous runtimes. We present cumulative regret bounds for the bandit components and characterize parallel scalability via Amdahl's Law. We validate ALMAB-DC on five benchmarks. On the two statistical experimental-design tasks, ALMAB-DC achieves lower simple regret than Equal Spacing, Random, and D-optimal designs in dose--response optimization, and in adaptive spatial field estimation matches the Greedy Max-Variance benchmark while outperforming Latin Hypercube Sampling; at $K=4$ the distributed setting reaches target performance in one-quarter of sequential wall-clock rounds. On three ML/engineering tasks (CIFAR-10 HPO, CFD drag minimization, MuJoCo RL), ALMAB-DC achieves 93.4\% CIFAR-10 accuracy (outperforming BOHB by 1.7\,pp and Optuna by 1.1\,pp), reduces airfoil drag to $C_D = 0.059$ (36.9\% below Grid Search), and improves RL return by 50\% over Grid Search. All advantages over non-ALMAB baselines are statistically significant under Bonferroni-corrected Mann--Whitney $U$ tests. Distributed execution achieves $7.5\times$ speedup at $K = 16$ agents, consistent with Amdahl's Law.

The Shapley value, and its broader family of semi-values, has received much attention in various attribution problems. A fundamental and long-standing challenge is their efficient approximation, since exact computation generally requires an exponential number of utility queries in the number of players $n$. To meet the challenges of large-scale applications, we explore the limits of efficiently approximating semi-values under a $Θ(n)$ space constraint. Building upon a vector concentration inequality, we establish a theoretical framework that enables sharper query complexities for existing unbiased randomized algorithms. Within this framework, we systematically develop a linear-space algorithm that requires $O(\frac{n}{ε^{2}}\log\frac{1}δ)$ utility queries to ensure $P(\|\hat{\boldsymbolϕ}-\boldsymbolϕ\|_{2}\geqε)\leq δ$ for all commonly used semi-values. In particular, our framework naturally bridges OFA, unbiased kernelSHAP, SHAP-IQ and the regression-adjusted approach, and definitively characterizes when paired sampling is beneficial. Moreover, our algorithm allows explicit minimization of the mean squared error $\mathbb{E}[\|\hat{\boldsymbolϕ}-\boldsymbolϕ\|_{2}^{2}]$ for each specific utility function. Accordingly, we introduce the first adaptive, linear-time, linear-space randomized algorithm, Adalina, that theoretically achieves improved mean squared error. All of our theoretical findings are experimentally validated. Our code is available at https://github.com/watml/adalina.

Existing accountability frameworks for AI systems, legal, ethical, and regulatory, rest on a shared assumption: for any consequential outcome, at least one identifiable person had enough involvement and foresight to bear meaningful responsibility. This paper proves that agentic AI systems violate this assumption not as an engineering limitation but as a mathematical necessity once autonomy exceeds a computable threshold. We introduce Human-Agent Collectives, a formalisation of joint human-AI systems where agents are modelled as state-policy tuples within a shared structural causal model. Autonomy is characterised through a four-dimensional information-theoretic profile (epistemic, executive, evaluative, social); collective behaviour through interaction graphs and joint action spaces. We axiomatise legitimate accountability through four minimal properties: Attributability (responsibility requires causal contribution), Foreseeability Bound (responsibility cannot exceed predictive capacity), Non-Vacuity (at least one agent bears non-trivial responsibility), and Completeness (all responsibility must be fully allocated). Our central result, the Accountability Incompleteness Theorem, proves that for any collective whose compound autonomy exceeds the Accountability Horizon and whose interaction graph contains a human-AI feedback cycle, no framework can satisfy all four properties simultaneously. The impossibility is structural: transparency, audits, and oversight cannot resolve it without reducing autonomy. Below the threshold, legitimate frameworks exist, establishing a sharp phase transition. Experiments on 3,000 synthetic collectives confirm all predictions with zero violations. This is the first impossibility result in AI governance, establishing a formal boundary below which current paradigms remain valid and above which distributed accountability mechanisms become necessary.

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.

Multiple-choice questions (MCQs) are widely used to evaluate large language models (LLMs). However, LLMs remain vulnerable to the presence of plausible distractors. This often diverts attention toward irrelevant choices, resulting in unstable oscillation between correct and incorrect answers. In this paper, we propose Inclusion-of-Thoughts (IoT), a progressive self-filtering strategy that is designed to mitigate this cognitive load (i.e., instability of model preferences under the presence of distractors) and enable the model to focus more effectively on plausible answers. Our method operates to reconstruct the MCQ using only plausible option choices, providing a controlled setting for examining comparative judgements and therefore the stability of the model's internal reasoning under perturbation. By explicitly documenting this filtering process, IoT also enhances the transparency and interpretability of the model's decision-making. Extensive empirical evaluation demonstrates that IoT substantially boosts chain-of-thought performance across a range of arithmetic, commonsense reasoning, and educational benchmarks with minimal computational overhead.

Vision-centric retrieval for VQA requires retrieving images to supply missing visual cues and integrating them into the reasoning process. However, selecting the right images and integrating them effectively into the model's reasoning remains challenging.To address this challenge, we propose R3G, a modular Reasoning-Retrieval-Reranking framework.It first produces a brief reasoning plan that specifies the required visual cues, then adopts a two-stage strategy, with coarse retrieval followed by fine-grained reranking, to select evidence images.On MRAG-Bench, R3G improves accuracy across six MLLM backbones and nine sub-scenarios, achieving state-of-the-art overall performance. Ablations show that sufficiency-aware reranking and reasoning steps are complementary, helping the model both choose the right images and use them well. We release code and data at https://github.com/czh24/R3G.

Large language models (LLMs) exhibiting test-time scaling behavior, such as extended reasoning traces and self-verification, have demonstrated remarkable performance on complex, long-term reasoning tasks. However, the robustness of these reasoning behaviors remains underexplored. To investigate this, we conduct a systematic evaluation of multiple reasoning models across three scenarios: (1) problems augmented with lengthy, irrelevant context; (2) multi-turn conversational settings with independent tasks; and (3) problems presented as a subtask within a complex task. We observe an interesting phenomenon: reasoning models tend to produce much shorter reasoning traces (up to 65%) for the same problem under different context conditions compared to the traces produced when the problem is presented in isolation. A finer-grained analysis reveals that this compression is associated with a decrease in self-verification and uncertainty management behaviors, such as double-checking. While this behavioral shift does not compromise performance on straightforward problems, it might affect performance on more challenging tasks. Additionally, we show that targeted supervised fine-tuning partially mitigates the adverse effects of irrelevant context. We hope our findings draw additional attention to both the robustness of reasoning models and the problem of context management for LLMs and LLM-based agents.

Estimation of heterogeneous long-term treatment effects (HLTEs) is relevant for personalized decision-making in marketing, economics, and medicine, where short-term observational datasets are often combined with long-term observational datasets. However, HLTE estimation is challenging due to limited overlap in treatment assignments or in long-term outcomes for certain subpopulations, which can lead to unstable HLTE estimates with large finite-sample variance. To address this challenge, we introduce the LT-O-learners (Long-Term Orthogonal Learners), a set of novel orthogonal learners for HLTE estimation in the canonical HLTE setting with surrogacy. The key idea of our LT-O-learners is to retarget the loss via custom overlap weights that downweight low-overlap samples. We show that the retargeted loss recovers the true HLTE pointwise and satisfies Neyman-orthogonality. We further prove two key theoretical results: (i) The nuisance error enters the error bound only through higher-order terms, which means our learners are robust to nuisance estimation error. (ii) Under a linear function class, the retargeting effectively controls the asymptotic variance of the HLTE estimator via the overlap weights in low-overlap regimes. We conduct experiments on synthetic and real-world datasets to confirm the theoretical properties of our LT-O-learners, particularly robustness in low-overlap regimes. To our knowledge, ours are the first orthogonal learners for HLTE estimation robust to low overlap in long-term settings.

Understanding emotions in natural language is inherently a multi-dimensional reasoning problem, where multiple affective signals interact through context, interpersonal relations, and situational cues. However, most existing emotion understanding benchmarks rely on short texts and predefined emotion labels, reducing this process to independent label prediction and ignoring the structured dependencies among emotions. To address this limitation, we introduce Emotional Scenarios (EmoScene), a theory-grounded benchmark of 4,731 contextrich scenarios annotated with an 8-dimensional emotion vector derived from Plutchik's basic emotions. Motivated by the observation that emotions rarely occur independently, we further propose an entanglement-aware Bayesian inference framework that incorporates emotion co-occurrence statistics to perform joint posterior inference over the emotion vector. This lightweight post-processing does not require any parameter updates and improves the structural consistency of predictions, and yields overall gains of 2.24% Lexical Accuracy without any additional cost. EmoScene therefore provides a challenging benchmark for studying multi-dimensional emotion understanding and the limitations of current language models.

Modern Text-to-Image (T2I) diffusion models have achieved remarkable semantic alignment, yet they often suffer from a significant lack of variety, converging on a narrow set of visual solutions for any given prompt. This typicality bias presents a challenge for creative applications that require a wide range of generative outcomes. We identify a fundamental trade-off in current approaches to diversity: modifying model inputs requires costly optimization to incorporate feedback from the generative path. In contrast, acting on spatially-committed intermediate latents tends to disrupt the forming visual structure, leading to artifacts. In this work, we propose to apply repulsion in the Contextual Space as a novel framework for achieving rich diversity in Diffusion Transformers. By intervening in the multimodal attention channels, we apply on-the-fly repulsion during the transformer's forward pass, injecting the intervention between blocks where text conditioning is enriched with emergent image structure. This allows for redirecting the guidance trajectory after it is structurally informed but before the composition is fixed. Our results demonstrate that repulsion in the Contextual Space produces significantly richer diversity without sacrificing visual fidelity or semantic adherence. Furthermore, our method is uniquely efficient, imposing a small computational overhead while remaining effective even in modern "Turbo" and distilled models where traditional trajectory-based interventions typically fail.

Large Vision-Language Models (LVLMs) often produce plausible but unreliable outputs, making robust uncertainty estimation essential. Recent work on semantic uncertainty estimates relies on external models to cluster multiple sampled responses and measure their semantic consistency. However, these clustering methods are often fragile, highly sensitive to minor phrasing variations, and can incorrectly group or separate semantically similar answers, leading to unreliable uncertainty estimates. We propose Semantic Gaussian Process Uncertainty (SGPU), a Bayesian framework that quantifies semantic uncertainty by analyzing the geometric structure of answer embeddings, avoiding brittle clustering. SGPU maps generated answers into a dense semantic space, computes the Gram matrix of their embeddings, and summarizes their semantic configuration via the eigenspectrum. This spectral representation is then fed into a Gaussian Process Classifier that learns to map patterns of semantic consistency to predictive uncertainty, and that can be applied in both black-box and white-box settings. Across six LLMs and LVLMs on eight datasets spanning VQA, image classification, and textual QA, SGPU consistently achieves state-of-the-art calibration (ECE) and discriminative (AUROC, AUARC) performance. We further show that SGPU transfers across models and modalities, indicating that its spectral representation captures general patterns of semantic uncertainty.

Symbolic regression aims to find symbolic expressions that describe datasets. Due to its inherent interpretability, symbolic regression (SR) is a powerful paradigm for scientific discovery. Recent advances have expanded SR to describe related phenomena using a single expression with varying sets of parameters, thereby introducing a single categorical variable. To illustrate, this enables the search for a single expression describing temperaturedependent viscosity across multiple fluids, while simultaneously identifying a distinct set of fluid-specific parameters. We expand upon prior efforts by considering multiple categorical variables and introducing intermediate levels of parameter sharing. Rather than parameters being either entirely universal or entirely unique, some parameters can also be shared across specific categories while remaining distinct for others. This allows for separating universal effects (shared parameters), category-specific trends (partially-shared parameters), and category interactions (non-shared parameters). We test the limits of this setup in terms of reducing data requirements and transfer learning using a synthetic, fitting-only example. Furthermore, we apply the method to an astrophysics dataset also used in a previous single-category study. In comparison, we achieve similar fit quality with significantly fewer parameters while extracting additional information about the problem.

We present T$^\star$, a simple TraceRL-based training curriculum for progressive block-size scaling in masked diffusion language models (MDMs). Starting from an AR-initialized small-block MDM, T$^\star$ transitions smoothly to larger blocks, enabling higher-parallelism decoding with minimal performance degradation on math reasoning benchmarks. Moreover, further analysis suggests that T$^\star$ may actually converge to an alternative decoding schedule that achieves comparable performance.

While Large Language Models (LLMs) are increasingly used as primary sources of information, their potential for political bias may impact their objectivity. Existing benchmarks of LLM social bias primarily evaluate demographic stereotypes, and when political bias is measured, it is done so at a coarse level, overlooking the values that shape sociopolitical reasoning. We introduce PoliticsBench, a multi-stage roleplay benchmark for evaluating fine-grained value expression in LLMs. Across twenty evolving scenarios, models articulate tradeoffs, take positions, and make decisions under competing pressures. Across eight prominent LLMs, we show that scenario-based prompting elicits broader and more strongly expressed value profiles than direct political questions, with peak interaction stages increasing the number of strongly activated value dimensions by approximately $0.75$ (out of 10 total dimensions), a statistically significant increase relative to baseline prompting ($p < 0.05$). In addition, commitment to a stance increases over the course of interaction, rising by approximately $1.4$ points on a $[0,5]$ scale from initial to decision stages. While responses become less robust to scenario paraphrasing in later interaction stages, inter-judge agreement remains relatively stable. Our results suggest that evaluating LLM political behavior requires moving beyond static prompts toward longer interactive settings that capture how values are applied in context.

If autoresearch is itself a form of research, then autoresearch can be applied to research itself. We present Bilevel Autoresearch, a bilevel framework in which an outer autoresearch loop improves an inner autoresearch loop by reading its code and traces, identifying bottlenecks, and generating injectable Python search mechanisms at runtime. The inner loop optimizes task performance; the outer loop optimizes how the inner loop searches. Both loops use the same LLM, so improvements come from the bilevel architecture rather than a stronger meta-level model, although the outer loop consumes additional inference and wall-clock budget. On Karpathy's GPT pretraining benchmark, the meta-autoresearch outer loop achieves a 5x improvement over the standard inner loop alone (-0.045 vs. -0.009 val_bpb), while parameter-level adjustment without mechanism change yields no reliable gain. The outer loop instantiates mechanisms from adjacent search domains, including combinatorial optimization, multi-armed bandits, and design of experiments, without human specification of the final mechanism design. Trace analysis suggests that these mechanisms break deterministic search patterns and force exploration of directions the LLM's priors avoid. The experiments demonstrate, on this benchmark, a first bilevel step: an outer loop improves the search behavior of an inner loop. Code is the mechanism carrier in this implementation, but skills, prompts, workflows, evaluators, domain principles, world-model assumptions, and memory schemas can also encode mechanisms that shape future agent behavior. This suggests a path toward recursive bootstrapping, where mechanisms discovered for the inner loop can be fed back to improve the meta-level loop itself.

Video Corpus Moment Retrieval (VCMR) aims to retrieve both the correct video and its temporal segment corresponding to a natural-language query, a task that is especially challenging for multi-verb queries where temporal action ordering is critical. Existing approaches often rely solely on text or static images and struggle to capture implicit motion dynamics, leading to retrieval errors and temporal misalignment. We propose GenSpan, a generation-calibrated VCMR framework that constructs short auxiliary videos from LLM-selected subtitle cues and decomposed sub-events, using these as temporal priors rather than direct retrieval targets. A token selector filters candidate-video features aligned with generated motion, and a bidirectional state-space model efficiently predicts video-moment tuples. Experiments on TVR and ActivityNet-Captions demonstrate that GenSpan improves corpus-level retrieval and moment localization, particularly for complex multi-action queries, while reducing computational cost compared to state-of-the-art multimodal baselines.

Spatial Transcriptomics (ST) profiles thousands of gene expression values at discrete spots with precise coordinates on tissue sections, preserving spatial context essential for clinical and pathological studies. With rising sequencing throughput and advancing platforms, the expanding data volumes motivate large-scale ST pretraining. However, the fundamental unit for pretraining, i.e., what constitutes a single training sample, remains ill-posed. Existing choices fall into two camps: (1) treating each spot as an independent sample, which discards spatial dependencies and collapses ST into single-cell transcriptomics; and (2) treating an entire slide as a single sample, which produces prohibitively large inputs and drastically fewer training examples, undermining effective pretraining. To address this gap, we propose treating spatial transcriptomics as croppable images. Specifically, we define a multi-channel image representation with fixed spatial size by cropping patches from raw slides, thereby preserving spatial context while substantially increasing the number of training samples. Along the channel dimension, we define gene subset selection rules to control input dimensionality and improve pretraining stability. Extensive experiments show that the proposed image-like dataset construction for ST pretraining consistently improves downstream performance, outperforming conventional pretraining schemes. Ablation studies verify that both spatial patching and channel design are necessary, establishing a unified, practical paradigm for organizing ST data and enabling large-scale pretraining.

To alleviate the heavy burden of paper screening, researchers increasingly rely on existing AI agents, such as AI reviewers or DeepResearch, for paper evaluation and novelty assessment. However, lacking specialized mechanisms for processing scholarly literature, their analyses often produce superficial results with noticeable deficiencies in quality. To bridge this gap, we introduce MemoNoveltyAgent, a multi-agent system designed to generate comprehensive and faithful novelty reports. Beyond retrieving concrete prior-paper evidence via RAG, our system incorporates a high-level abstract memory constructed from large-scale scholarly corpora. This memory organizes research into hierarchical trees to distill field-specific evolutionary trajectories, thereby providing a broader historical context. Furthermore, we decompose papers into discrete novelty points for fine-grained analysis and retrieval, while employing a self-validation mechanism to improve report faithfulness. Finally, to address the evaluation challenges of such open-ended generation tasks, we propose a RAG-augmented checklist evaluation method that enables reliable and evidence-grounded assessments. Extensive experiments demonstrate that MemoNoveltyAgent outperforms GPT-5 DeepResearch by 13.69%. Code and demo are available at https://github.com/SStan1/MemoNoveltyAgent

As generative AI advances, global governance frameworks increasingly mandate verifiable content provenance. However, existing watermarking techniques face a critical policy-to-technology disconnect: sampling-based methods require computationally prohibitive inversion, while fine-tuning approaches are tethered to specific model checkpoints, hindering standardized, cross-model oversight. To bridge this gap, we introduce DiffMark, a plug-and-play multi-bit watermarking framework. DiffMark embeds a persistent, learned perturbation into every denoising step of a frozen diffusion model, accumulating a recoverable signal in the final latent space. To enable efficient training through the frozen network, we utilize Latent Consistency Models (LCMs) as a differentiable training bridge. DiffMark achieves 64-bit extraction in a single 16.4 ms forward pass, which is a $45\times$ speed-up over inversion baselines. By enabling per-image key flexibility and cross-architecture transferability without retraining, DiffMark provides the practical, scalable technical tooling necessary to operationalize user accountability and enforce emerging AI governance mandates.

Data science plays a critical role in transforming complex data into actionable insights across numerous domains. Recent developments in large language models (LLMs) and artificial intelligence (AI) agents have significantly automated data science workflow. However, it remains unclear to what extent AI agents can match the performance of human experts on domain-specific data science tasks, and in which aspects human expertise continues to provide advantages. We introduce AgentDS, a benchmark and competition designed to evaluate both AI agents and human-AI collaboration performance in domain-specific data science. AgentDS consists of 17 challenges across six industries: commerce, food production, healthcare, insurance, manufacturing, and retail banking. We conducted an open competition involving 29 teams and 80 participants, enabling systematic comparison between human-AI collaborative approaches and AI-only baselines. Our results show that current AI agents struggle with domain-specific reasoning. AI-only baselines perform below the top quartile of competition participants, while the strongest solutions arise from human-AI collaboration. These findings challenge the narrative of complete automation by AI and underscore the enduring importance of human expertise in data science, while illuminating directions for the next generation of AI. Visit the AgentDS website here: https://agentds.org/ and open source datasets here: https://huggingface.co/datasets/lainmn/AgentDS .

Text-guided image editors can now manipulate authentic medical scans with high fidelity, enabling lesion implantation/removal that threatens clinical trust and safety. Existing defenses are inadequate for healthcare. Medical detectors are largely black-box, while MLLM-based explainers are typically post-hoc, lack medical expertise, and may hallucinate evidence on ambiguous cases. We present MedForge, a data-and-method solution for pre-hoc, evidence-grounded medical forgery detection. We introduce MedForge-90K, a large-scale benchmark of realistic lesion edits across 19 pathologies with expert-guided reasoning supervision via doctor inspection guidelines and gold edit locations. Building on it, MedForge-Reasoner performs localize-then-analyze reasoning, predicting suspicious regions before producing a verdict, and is further aligned with Forgery-aware GSPO to strengthen grounding and reduce hallucinations. Experiments demonstrate state-of-the-art detection accuracy and trustworthy, expert-aligned explanations.

Large language models (LLMs) with chain-of-thought reasoning achieve state-of-the-art performance across complex problem-solving tasks, but their verbose reasoning traces and large context requirements make them impractical for edge deployment. These challenges include high token generation costs, large KV-cache footprints, and inefficiencies when distilling reasoning capabilities into smaller models for mobile devices. Existing approaches often rely on distilling reasoning traces from larger models into smaller models, which are verbose and stylistically redundant, undesirable for on-device inference. In this work, we propose a lightweight approach to enable reasoning in small LLMs using LoRA adapters combined with supervised fine-tuning. We further introduce budget forcing via reinforcement learning on these adapters, significantly reducing response length with minimal accuracy loss. To address memory-bound decoding, we exploit parallel test-time scaling, improving accuracy at minor latency increase. Finally, we present a dynamic adapter-switching mechanism that activates reasoning only when needed and a KV-cache sharing strategy during prompt encoding, reducing time-to-first-token for on-device inference. Experiments on Qwen2.5-7B demonstrate that our method achieves efficient, accurate reasoning under strict resource constraints, making LLM reasoning practical for mobile scenarios. Videos demonstrating our solution running on mobile devices are available on our project page.