Background: Trustworthy AI serves as a foundational pillar for two major AI ethics conferences: AIES and FAccT. However, current research often adopts techno-centric approaches, focusing primarily on technical attributes such as reliability, robustness, and fairness, while overlooking the sociotechnical dimensions critical to understanding AI trustworthiness in real-world contexts. Objectives: This scoping review aims to examine how the AIES and FAccT communities conceptualize, measure, and validate AI trustworthiness, identifying major gaps and opportunities for advancing a holistic understanding of trustworthy AI systems. Methods: We conduct a scoping review of AIES and FAccT conference proceedings to date, systematically analyzing how trustworthiness is defined, operationalized, and applied across different research domains. Our analysis focuses on conceptualization approaches, measurement methods, verification and validation techniques, application areas, and underlying values. Results: While significant progress has been made in defining technical attributes such as transparency, accountability, and robustness, our findings reveal critical gaps. Current research often predominantly emphasizes technical precision at the expense of social and ethical considerations. The sociotechnical nature of AI systems remains less explored and trustworthiness emerges as a contested concept shaped by those with the power to define it. Conclusions: An interdisciplinary approach combining technical rigor with social, cultural, and institutional considerations is essential for advancing trustworthy AI. We propose actionable measures for the AI ethics community to adopt holistic frameworks that genuinely address the complex interplay between AI systems and society, ultimately promoting responsible technological development that benefits all stakeholders.

Subject-driven image generation models face a fundamental trade-off between identity preservation (fidelity) and prompt adherence (editability). While online reinforcement learning (RL), specifically GPRO, offers a promising solution, we find that a naive application of GRPO leads to competitive degradation, as the simple linear aggregation of rewards with static weights causes conflicting gradient signals and a misalignment with the temporal dynamics of the diffusion process. To overcome these limitations, we propose Customized-GRPO, a novel framework featuring two key innovations: (i) Synergy-Aware Reward Shaping (SARS), a non-linear mechanism that explicitly penalizes conflicted reward signals and amplifies synergistic ones, providing a sharper and more decisive gradient. (ii) Time-Aware Dynamic Weighting (TDW), which aligns the optimization pressure with the model's temporal dynamics by prioritizing prompt-following in the early, identity preservation in the later. Extensive experiments demonstrate that our method significantly outperforms naive GRPO baselines, successfully mitigating competitive degradation. Our model achieves a superior balance, generating images that both preserve key identity features and accurately adhere to complex textual prompts.

The reliable execution of high-level missions in multi-robot systems with heterogeneous agents, requires robust methods for detecting spurious behaviors. In this paper, we address the challenge of identifying spurious executions of plans specified as a Linear Temporal Logic (LTL) formula, as incorrect task sequences, violations of spatial constraints, timing inconsistencies, or deviations from intended mission semantics. To tackle this, we introduce a structured data generation framework based on the Nets-within-Nets (NWN) paradigm, which coordinates robot actions with LTL-derived global mission specifications. We further propose a Transformer-based anomaly detection pipeline that classifies robot trajectories as normal or anomalous. Experimental evaluations show that our method achieves high accuracy (91.3%) in identifying execution inefficiencies, and demonstrates robust detection capabilities for core mission violations (88.3%) and constraint-based adaptive anomalies (66.8%). An ablation experiment of the embedding and architecture was carried out, obtaining successful results where our novel proposition performs better than simpler representations.

Building effective knowledge graphs (KGs) for Retrieval-Augmented Generation (RAG) is pivotal for advancing question answering (QA) systems. However, its effectiveness is hindered by a fundamental disconnect: the knowledge graph (KG) construction process is decoupled from its downstream application, yielding suboptimal graph structures. To bridge this gap, we introduce AutoGraph-R1, the first framework to directly optimize KG construction for task performance using Reinforcement Learning (RL). AutoGraph-R1 trains an LLM constructor by framing graph generation as a policy learning problem, where the reward is derived from the graph's functional utility in a RAG pipeline. We design two novel, task-aware reward functions, one for graphs as knowledge carriers and another as knowledge indices. Across multiple QA benchmarks, AutoGraph-R1 consistently enables graph RAG methods to achieve significant performance gains over using task-agnostic baseline graphs. Our work shows it is possible to close the loop between construction and application, shifting the paradigm from building intrinsically ``good'' graphs to building demonstrably ``useful'' ones.

Training effective multilingual embedding models presents unique challenges due to the diversity of languages and task objectives. Although small multilingual models (<1 B parameters) perform well on multilingual tasks generally, they consistently lag behind larger models (>1 B) in the most prevalent use case: retrieval. This raises a critical question: Can smaller models be retrofitted specifically for retrieval tasks to enhance their performance? In this work, we investigate key factors that influence the effectiveness of multilingual embeddings, focusing on training data scale, negative sampling strategies, and data diversity. We find that while increasing the scale of training data yields initial performance gains, these improvements quickly plateau - indicating diminishing returns. Incorporating hard negatives proves essential for consistently improving retrieval accuracy. Furthermore, our analysis reveals that task diversity in the training data contributes more significantly to performance than language diversity alone. As a result, we develop a compact (approximately 300M) multilingual model that achieves retrieval performance comparable to or even surpassing current strong 7B models.

We propose and test the LLM Brain Rot Hypothesis: continual exposure to junk web text induces lasting cognitive decline in large language models (LLMs). To unveil junk effects, we designed a novel controlled experiment on real Twitter/X corpora, by constructing junk and reverse-controlled datasets via two orthogonal operationalizations: M1 (engagement degree) and M2 (semantic quality), with matched token scale and training operations across conditions. Compared to the control group, continual pre-training of 4 LLMs on the junk dataset causes non-trivial declines (Hedges' g>0.3) on reasoning, long-context understanding, safety, and inflating "dark traits" (e.g., psychopathy, narcissism). The gradual mixtures of junk and control datasets also yield dose-response cognition decay: for example, under M1, ARC-Challenge with Chain-of-Thought drops 72.1 -> 57.2 and RULER-CWE 83.7 -> 52.3 as junk ratio rises from 0% to 100%. Error forensics reveal several key insights. First, we identify thought-skipping as the primary lesion in reasoning: models increasingly truncate or skip chains. Second, partial but incomplete healing is observed: scaling instruction tuning and clean continual pre-training improve the declined cognition, yet cannot restore baseline capability, suggesting persistent representational drift rather than format mismatch. Finally, we discover that the popularity, a non-semantic metric, of a tweet is a better indicator of the Brain Rot effect than the length in M1. Together, the results provide significant, multi-perspective evidence that social effects of data could be a causal driver of LLM capability decay in continual pre-training, thereby motivating routine "cognitive health checks" for deployed and evolving LLMs.

In future intelligent transportation systems, autonomous cooperative planning (ACP), becomes a promising technique to increase the effectiveness and security of multi-vehicle interactions. However, multiple uncertainties cannot be fully addressed for existing ACP strategies, e.g. perception, planning, and communication uncertainties. To address these, a novel deep reinforcement learning-based autonomous cooperative planning (DRLACP) framework is proposed to tackle various uncertainties on cooperative motion planning schemes. Specifically, the soft actor-critic (SAC) with the implementation of gate recurrent units (GRUs) is adopted to learn the deterministic optimal time-varying actions with imperfect state information occurred by planning, communication, and perception uncertainties. In addition, the real-time actions of autonomous vehicles (AVs) are demonstrated via the Car Learning to Act (CARLA) simulation platform. Evaluation results show that the proposed DRLACP learns and performs cooperative planning effectively, which outperforms other baseline methods under different scenarios with imperfect AV state information.

Gait recognition is an important biometric for human identification at a distance, particularly under low-resolution or unconstrained environments. Current works typically focus on either 2D representations (e.g., silhouettes and skeletons) or 3D representations (e.g., meshes and SMPLs), but relying on a single modality often fails to capture the full geometric and dynamic complexity of human walking patterns. In this paper, we propose a multi-modal and multi-task framework that combines 2D temporal silhouettes with 3D SMPL features for robust gait analysis. Beyond identification, we introduce a multitask learning strategy that jointly performs gait recognition and human attribute estimation, including age, body mass index (BMI), and gender. A unified transformer is employed to effectively fuse multi-modal gait features and better learn attribute-related representations, while preserving discriminative identity cues. Extensive experiments on the large-scale BRIAR datasets, collected under challenging conditions such as long-range distances (up to 1 km) and extreme pitch angles (up to 50°), demonstrate that our approach outperforms state-of-the-art methods in gait recognition and provides accurate human attribute estimation. These results highlight the promise of multi-modal and multitask learning for advancing gait-based human understanding in real-world scenarios.

Autonomous navigation in unfamiliar environments requires robots to simultaneously explore, localise, and plan under uncertainty, without relying on predefined maps or extensive training. We present Active Inference MAPping and Planning (AIMAPP), a framework unifying mapping, localisation, and decision-making within a single generative model, drawing on cognitive-mapping concepts from animal navigation (topological organisation, discrete spatial representations and predictive belief updating) as design inspiration. The agent builds and updates a sparse topological map online, learns state transitions dynamically, and plans actions by minimising Expected Free Energy. This allows it to balance goal-directed and exploratory behaviours. We implemented AIMAPP as a ROS-compatible system that is sensor and robot-agnostic and integrates with diverse hardware configurations. It operates in a fully self-supervised manner, is resilient to sensor failure, continues operating under odometric drift, and supports both exploration and goal-directed navigation without any pre-training. We evaluate the system in large-scale real and simulated environments against state-of-the-art planning baselines, demonstrating its adaptability to ambiguous observations, environmental changes, and sensor noise. The model offers a modular, self-supervised solution to scalable navigation in unstructured settings. AIMAPP is available at https://github.com/decide-ugent/aimapp.

The rapid growth of AI-generated imagery has blurred the boundary between real and synthetic content, raising practical concerns for digital integrity. Vision-language models (VLMs) can provide natural language explanations, but standard one-pass classifiers often miss subtle artifacts in high-quality synthetic images and offer limited grounding in the pixels. We propose Locate-Then-Examine (LTE), a two-stage VLM-based forensic framework that first localizes suspicious regions and then re-examines these crops together with the full image to refine the real vs. AI-generated verdict and its explanation. LTE explicitly links each decision to localized visual evidence through region proposals and region-aware reasoning. To support training and evaluation, we introduce TRACE, a dataset of 20,000 real and high-quality synthetic images with region-level annotations and automatically generated forensic explanations, constructed by a VLM-based pipeline with additional consistency checks and quality control. Across TRACE and multiple external benchmarks, LTE achieves competitive accuracy and improved robustness while providing human-understandable, region-grounded explanations suitable for forensic deployment.

Integrating textual graphs into Large Language Models (LLMs) is promising for complex graph-based QA. However, a key bottleneck is retrieving informative yet compact subgraphs that fit the LLM context. Existing retrievers often struggle, relying either on shallow embedding similarity or costly interactive policies that require excessive supervision. To address these challenges, we introduce an agentic textual graph reasoning framework featuring an LLM-based retriever trained with synthetic stepwise supervision. Rather than relying on final answer rewards which often yield sparse and unstable signals, we optimize the retriever by evaluating each step against offline-extracted golden subgraphs. Our approach distills golden subgraphs via a specialized data synthesis pipeline to formulate dense rewards, facilitating a two-stage training scheme that effectively learns the interactive graph exploration policy. Based on extensive experiments on three common datasets in comparison with seven strong baselines, our approach achieves an average improvement of 15.6% in accuracy and 17.2% in F1 score. The advantage is even higher in more complicated multi-hop reasoning tasks.

Training large-scale recommendation models under a single global objective implicitly assumes homogeneity across user populations. However, real-world data are composites of heterogeneous cohorts with distinct conditional distributions. As models increase in scale and complexity and as more data is used for training, they become dominated by central distribution patterns, neglecting head and tail regions. This imbalance limits the model's learning ability and can result in inactive attention weights or dead neurons. In this paper, we reveal how the attention mechanism can play a key role in factorization machines for shared embedding selection, and propose to address this challenge by analyzing the substructures in the dataset and exposing those with strong distributional contrast through auxiliary learning. Unlike previous research, which heuristically applies weighted labels or multi-task heads to mitigate such biases, we leverage partially conflicting auxiliary labels to regularize the shared representation. This approach customizes the learning process of attention layers to preserve mutual information with minority cohorts while improving global performance. We evaluated proposed method on massive production datasets with billions of data points each for six SOTA models. Experiments show that the factorization machine is able to capture fine-grained user-ad interactions using the proposed method, achieving up to a 0.16% reduction in normalized entropy overall and delivering gains exceeding 0.30% on targeted minority cohorts.

Standard representational similarity methods align each layer of a network to its best match in another independently, producing asymmetric results, lacking a global alignment score, and struggling with networks of different depths. These limitations arise from ignoring global activation structure and restricting mappings to rigid one-to-one layer correspondences. We propose Multi-Level Optimal Transport (MOT), a unified framework that jointly infers soft, globally consistent layer-to-layer couplings and neuron-level transport plans. MOT allows source neurons to distribute mass across multiple target layers while minimizing total transport cost under marginal constraints. This yields both a single alignment score for the entire network comparison and a soft transport plan that naturally handles depth mismatches through mass distribution. We evaluate MOT on vision models, large language models, and human visual cortex recordings. Across all domains, MOT matches or surpasses standard pairwise matching in alignment quality. Moreover, it reveals smooth, fine-grained hierarchical correspondences: early layers map to early layers, deeper layers maintain relative positions, and depth mismatches are resolved by distributing representations across multiple layers. These structured patterns emerge naturally from global optimization without being imposed, yet are absent in greedy layer-wise methods. MOT thus enables richer, more interpretable comparisons between representations, particularly when networks differ in architecture or depth. We further extend our method to a three-level MOT framework, providing a proof-of-concept alignment of two networks across their training trajectories and demonstrating that MOT uncovers checkpoint-wise correspondences missed by greedy layer-wise matching.

When people query Vision-Language Models (VLMs) but cannot see the accompanying visual context (e.g. for blind and low-vision users), augmenting VLM predictions with natural language explanations can signal which model predictions are reliable. However, prior work has found that explanations can easily convince users that inaccurate VLM predictions are correct. To remedy undesirable overreliance on VLM predictions, we propose evaluating two complementary qualities of VLM-generated explanations via two quality scoring functions. We propose Visual Fidelity, which captures how faithful an explanation is to the visual context, and Contrastiveness, which captures how well the explanation identifies visual details that distinguish the model's prediction from plausible alternatives. On the A-OKVQA, VizWiz, and MMMU-Pro tasks, these quality scoring functions are better calibrated with model correctness than existing explanation qualities. We conduct a user study in which participants have to decide whether a VLM prediction is accurate without viewing its visual context. We observe that showing our quality scores alongside VLM explanations improves participants' accuracy at predicting VLM correctness by 11.1%, including a 15.4% reduction in the rate of falsely believing incorrect predictions. These findings highlight the utility of explanation quality scores in fostering appropriate reliance on VLM predictions.

Logical reasoning is a fundamental capability of large language models. However, existing studies often overlook the interaction between logical complexity and semantic complexity, leading to systems that struggle with abstract propositions, ambiguous contexts, and conflicting stances that are central to human reasoning. We propose LogicAgent, a semiotic-square-guided framework that jointly addresses these two axes of difficulty. The semiotic square provides a principled structure for multi-perspective semantic analysis, and LogicAgent integrates automated deduction with reflective verification to manage logical complexity across deeper reasoning chains. To support evaluation under these conditions, we introduce RepublicQA, a benchmark that couples semantic complexity with logical depth. RepublicQA reaches college-level semantic difficulty (FKGL 11.94), contains philosophically grounded abstract propositions with systematically constructed contrary and contradictory forms, and offers a semantically rich setting for assessing logical reasoning in large language models. Experiments show that LogicAgent achieves state-of-the-art performance on RepublicQA with a 6.25 percent average improvement over strong baselines, and generalizes effectively to mainstream logical reasoning benchmarks including ProntoQA, ProofWriter, FOLIO, and ProverQA, achieving an additional 7.05 percent average gain. These results demonstrate the effectiveness of semiotic-grounded multi-perspective reasoning in enhancing logical performance. Code is available at https://github.com/AI4SS/Logic-Agent.

Reasoning capability is essential to ensure the factual correctness of the responses of transformer-based Large Language Models (LLMs), and robust reasoning about transitive relations is instrumental in many settings, such as causal inference. Hence, it is essential to investigate the capability of transformers in the task of inferring transitive relations (e.g., knowing A causes B and B causes C, then A causes C). The task of inferring transitive relations is equivalent to the task of connectivity in directed graphs (e.g., knowing there is a path from A to B, and there is a path from B to C, then there is a path from A to C). Past research focused on whether transformers can learn to infer transitivity from in-context examples provided in the input prompt. However, transformers' capability to infer transitive relations from training examples and how scaling affects the ability is unexplored. In this study, we seek to answer this question by generating directed graphs to train transformer models of varying sizes and evaluate their ability to infer transitive relations for various graph sizes. Our findings suggest that transformers are capable of learning connectivity on "grid-like'' directed graphs where each node can be embedded in a low-dimensional subspace, and connectivity is easily inferable from the embeddings of the nodes. We find that the dimensionality of the underlying grid graph is a strong predictor of transformers' ability to learn the connectivity task, where higher-dimensional grid graphs pose a greater challenge than low-dimensional grid graphs. In addition, we observe that increasing the model scale leads to increasingly better generalization to infer connectivity over grid graphs. However, if the graph is not a grid graph and contains many disconnected components, transformers struggle to learn the connectivity task, especially when the number of components is large.

WARNING: This paper contains examples of offensive materials. To address the proliferation of toxic content on social media, we introduce SMARTER, we introduce SMARTER, a data-efficient two-stage framework for explainable content moderation using Large Language Models (LLMs). In Stage 1, we leverage LLMs' own outputs to generate synthetic explanations for both correct and incorrect labels, enabling alignment via preference optimization with minimal human supervision. In Stage 2, we refine explanation quality through cross-model training, allowing weaker models to align stylistically and semantically with stronger ones. Experiments on three benchmark tasks -- HateXplain, Latent Hate, and Implicit Hate -- demonstrate that SMARTER enables LLMs to achieve up to a 13% macro-F1 improvement over standard few-shot baselines while using only a fraction of the full training data. Our framework offers a scalable strategy for low-resource settings by harnessing LLMs' self-improving capabilities for both classification and explanation.

Accurate 3D reconstruction in degraded imaging conditions remains a key challenge in photogrammetry and neural rendering. In underwater environments, spatially varying visibility caused by scattering, attenuation, and sparse observations leads to highly non-uniform information quality. Existing 3D Gaussian Splatting (3DGS) methods typically optimize primitives based on photometric signals alone, resulting in imbalanced representation, with overfitting in well-observed regions and insufficient reconstruction in degraded areas. In this paper, we propose SWAGSplatting (Semantic-guided Water-scene Augmented Gaussian Splatting), a multimodal framework that integrates semantic priors into 3DGS for robust, high-fidelity underwater reconstruction. Each Gaussian primitive is augmented with a learnable semantic feature, supervised by CLIP-based embeddings derived from region-level cues. A semantic consistency loss is introduced to align geometric reconstruction with high-level semantics, improving structural coherence and preserving salient object boundaries under challenging conditions. Furthermore, we propose an adaptive Gaussian primitive reallocation strategy that redistributes representation capacity based on both primitive importance and reconstruction error, mitigating the imbalance introduced by conventional densification. This enables more effective modeling of low-visibility regions without increasing computational cost. Extensive experiments on real-world datasets, including SeaThru-NeRF, Submerged3D, and S-UW, demonstrate that the proposed method consistently outperforms state-of-the-art approaches in terms of average PSNR, SSIM, and LPIPS. The results validate the effectiveness of integrating semantic priors for high-fidelity underwater scene reconstruction. Code is available at https://github.com/theflash987/SWAGSplatting.

Knowledge-intensive visual question answering (VQA) requires external knowledge beyond image content, demanding precise visual grounding and coherent integration of visual and textual information. Although multimodal retrieval-augmented generation has achieved notable advances by incorporating external knowledge bases, existing approaches largely adopt single-pass frameworks that often fail to acquire sufficient knowledge and lack mechanisms to revise misdirected reasoning. We propose PMSR (Progressive Multimodal Search and Reasoning), a framework that progressively constructs a structured reasoning trajectory to enhance both knowledge acquisition and synthesis. PMSR uses dual-scope queries conditioned on both the latest record and the trajectory to retrieve diverse knowledge from heterogeneous knowledge bases. The retrieved evidence is then synthesized into compact records via compositional reasoning. This design facilitates controlled iterative refinement, which supports more stable reasoning trajectories with reduced error propagation. Extensive experiments across six diverse benchmarks (Encyclopedic-VQA, InfoSeek, MMSearch, LiveVQA, FVQA, and OK-VQA) demonstrate that PMSR consistently improves both retrieval recall and end-to-end answer accuracy.

Post-Training Quantization (PTQ) is a critical strategy for efficient Large Language Models (LLMs) deployment. However, existing scaling laws primarily focus on general performance, overlooking crucial fine-grained factors and how quantization differentially impacts diverse knowledge capabilities. To address this, we establish Task-Stratified Knowledge Scaling Laws. By stratifying capabilities into memorization, application, and reasoning, we develop a framework that unifies model size, bit-width, and fine-grained factors: group size and calibration set size. Validated on 293 diverse PTQ configurations, our framework demonstrates strong fit and cross-architecture consistency. It reveals distinct sensitivities across knowledge capabilities: reasoning is precision-critical, application is scale-responsive, and memorization is calibration-sensitive. We highlight that in low-bit scenarios, optimizing these fine-grained factors is essential for preventing performance collapse. These findings provide an empirically-backed foundation for designing knowledge-aware quantization strategies.

Retrieval-augmented generation (RAG) has become a widely recognized paradigm to combine parametric memory with non-parametric memories. An RAG model consists of two serial connecting components (retriever and generator). A major challenge in end-to-end optimization of the RAG model is that marginalization over relevant passages (modeled as discrete latent variables) from a knowledge base is required. Traditional top-K marginalization and variational RAG (VRAG) suffer from biased or high-variance gradient estimates. In this paper, we propose and develop joint stochastic approximation (JSA) based end-to-end training of RAG, which is referred to as JSA-RAG. The JSA algorithm is a stochastic extension of the EM (expectation-maximization) algorithm and is particularly powerful in estimating discrete latent variable models. Extensive experiments are conducted on five datasets for two tasks (open-domain question answering, knowledge-grounded dialogs) and show that JSA-RAG significantly outperforms both vanilla RAG and VRAG. Further analysis shows the efficacy of JSA-RAG from the perspectives of generation, retrieval, and low-variance gradient estimate.

In safety-critical applications data-driven models must not only be accurate but also provide reliable uncertainty estimates. This property, commonly referred to as calibration, is essential for risk-aware decision-making. In regression a wide variety of calibration metrics and recalibration methods have emerged. However, these metrics differ significantly in their definitions, assumptions and scales, making it difficult to interpret and compare results across studies. Moreover, most recalibration methods have been evaluated using only a small subset of metrics, leaving it unclear whether improvements generalize across different notions of calibration. In this work, we systematically extract and categorize regression calibration metrics from the literature and benchmark these metrics independently of specific modelling methods or recalibration approaches. Through controlled experiments with real-world, synthetic and artificially miscalibrated data, we demonstrate that calibration metrics frequently produce conflicting results. Our analysis reveals substantial inconsistencies: many metrics disagree in their evaluation of the same recalibration result, and some even indicate contradictory conclusions. This inconsistency is particularly concerning as it potentially allows cherry-picking of metrics to create misleading impressions of success. We identify the Expected Normalized Calibration Error (ENCE) and the Coverage Width-based Criterion (CWC) as the most dependable metrics in our tests. Our findings highlight the critical role of metric selection in calibration research.

Humanoids operating in real-world workspaces must frequently execute task-driven, short-range movements to SE(2) target poses. To be practical, these transitions must be fast, robust, and energy efficient. While learning-based locomotion has made significant progress, most existing methods optimize for velocity-tracking rather than direct pose reaching, resulting in inefficient, marching-style behavior when applied to short-range tasks. In this work, we develop a reinforcement learning approach that directly optimizes humanoid locomotion for SE(2) targets. Central to this approach is a new constellation-based reward function that encourages natural and efficient target-oriented movement. To evaluate performance, we introduce a benchmarking framework that measures energy consumption, time-to-target, and footstep count on a distribution of SE(2) goals. Our results show that the proposed approach consistently outperforms standard methods and enables successful transfer from simulation to hardware, highlighting the importance of targeted reward design for practical short-range humanoid locomotion.

Large-scale Vision-Language Models (VLMs) have transformed general-purpose visual recognition through strong zero-shot capabilities. However, their performance degrades significantly in niche, safety-critical domains such as industrial spill detection, where hazardous events are rare, sensitive, and difficult to annotate. This scarcity -- driven by privacy concerns, data sensitivity, and the infrequency of real incidents -- renders conventional fine-tuning of detectors infeasible for most industrial settings. We address this challenge by introducing a scalable framework centered on a high-quality synthetic data generation pipeline. We demonstrate that this synthetic corpus enables effective Parameter-Efficient Fine-Tuning (PEFT) of VLMs and substantially boosts the performance of state-of-the-art object detectors such as YOLO and DETR. Notably, in the absence of synthetic data (SynSpill dataset), VLMs still generalize better to unseen spill scenarios than these detectors. When SynSpill is used, both VLMs and detectors achieve marked improvements, with their performance becoming comparable. Our results underscore that high-fidelity synthetic data is a powerful means to bridge the domain gap in safety-critical applications. The combination of synthetic generation and lightweight adaptation offers a cost-effective, scalable pathway for deploying vision systems in industrial environments where real data is scarce/impractical to obtain. Project Page: https://synspill.vercel.app

As large language models (LLMs) become increasingly popular, there is a growing need to predict which out of a set of LLMs will yield a successful answer to a given query at low cost. This problem promises to become even more relevant as LLM agents are asked to solve an increasing variety of "agentic'' AI tasks. Such tasks are often broken into smaller subtasks, each of which can then be executed by a LLM expected to perform well on that specific subtask. For example, to extract a diagnosis from medical records, one can first select an LLM to summarize the record, select another to validate the summary, and then select a possibly different LLM to extract the diagnosis from the summarized record. Unlike existing LLM selection or routing algorithms, this setting requires selecting a sequence of LLMs, with the output of each LLM feeding into the next and potentially influencing its success. Thus, unlike single LLM selection, the quality of each subtask's output directly affects the inputs, and hence the cost and success rate, of downstream LLMs, creating complex performance dependencies that must be learned during selection. We propose a neural contextual bandit-based algorithm that trains neural networks to guide LLM selections for the different subtasks, without requiring historical LLM performance data. We prove that our proposed Sequential Bandits algorithm achieves a sublinear regret in the number of tasks, and we experimentally validate its superior performance compared to other LLM selection algorithms on two real datasets.

Current Vision Language Models (VLMs) demonstrate a critical gap between surface-level recognition and deep narrative reasoning when processing sequential visual storytelling. Through a comprehensive investigation of manga narrative understanding, we reveal that while recent large multimodal models excel at individual panel interpretation, they systematically fail at temporal causality and cross-panel cohesion, core requirements for coherent story comprehension. We introduce a novel evaluation framework that combines fine-grained multimodal annotation, cross-modal embedding analysis, and retrieval-augmented assessment to systematically characterize these limitations. Our methodology includes (i) a rigorous annotation protocol linking visual elements to narrative structure through aligned light novel text, (ii) comprehensive evaluation across multiple reasoning paradigms, including direct inference and retrieval-augmented generation, and (iii) cross-modal similarity analysis revealing fundamental misalignments in current VLMs' joint representations. Applying this framework to Re:Zero manga across 11 chapters with 308 annotated panels, we conduct the first systematic study of long-form narrative understanding in VLMs through three core evaluation axes: generative storytelling, contextual dialogue grounding, and temporal reasoning. Our findings demonstrate that current models lack genuine story-level intelligence, struggling particularly with non-linear narratives, character consistency, and causal inference across extended sequences. This work establishes both the foundation and practical methodology for evaluating narrative intelligence, while providing actionable insights into the capability of deep sequential understanding of Discrete Visual Narratives beyond basic recognition in Multimodal Models. Project Page: https://re-verse.vercel.app

As a class of generative artificial intelligence frameworks inspired by statistical physics, diffusion models have shown extraordinary performance in synthesizing complicated data distributions through a denoising process gradually guided by score functions. Real-life data, like images, is often spatially structured in low-dimensional spaces. However, ordinary diffusion models ignore this local structure and learn spatially global score functions, which are often computationally expensive. In this work, motivated by recent advances in non-equilibrium statistical physics, we develop a generic framework for defining phases of data distributions and use it to analyze the locality requirements of denoisers in diffusion models. We define two distributions as belonging to the same data distribution phase if they can be mutually connected via spatially local operations such as local denoisers, along the same evolution path as the diffusion. We demonstrate that the reverse denoising process consists of an early trivial phase and a late data phase, sandwiching a rapid phase transition where local denoisers must fail. We further demonstrate that the performance of local denoisers is closely tied to spatial Markovianity, which provides an operational criterion for diagnosing such phase transitions. We validate this criterion through numerical experiments on real-world datasets. Our work suggests guidance for simpler and more efficient architectures of diffusion models: far from the phase transition point, we can use small local neural networks to compute the score function; global neural networks are only necessary around the narrow time interval of phase transitions. This result also opens up new directions for studying phases of data distributions, the broader science of generative artificial intelligence, and guiding the design of neural networks inspired by physics concepts.

Diffusion policies excel at robotic manipulation by naturally modeling multimodal action distributions in high-dimensional spaces. Nevertheless, diffusion policies suffer from diffusion representation collapse: semantically similar observations are mapped to indistinguishable features, ultimately impairing their ability to handle subtle but critical variations required for complex robotic manipulation. To address this problem, we propose D2PPO (Diffusion Policy Policy Optimization with Dispersive Loss). D2PPO introduces dispersive loss regularization that combats representation collapse by treating all hidden representations within each batch as negative pairs. D2PPO compels the network to learn discriminative representations of similar observations, thereby enabling the policy to identify subtle yet crucial differences necessary for precise manipulation. In evaluation, we find that early-layer regularization benefits simple tasks, while late-layer regularization sharply enhances performance on complex manipulation tasks. On RoboMimic benchmarks, D2PPO achieves an average improvement of 22.7% in pre-training and 26.1% after fine-tuning, setting new SOTA results. In comparison with SOTA, results of real-world experiments on a Franka Emika Panda robot show the excitingly high success rate of our method. The superiority of our method is especially evident in complex tasks. Project page: https://guowei-zou.github.io/d2ppo/

Long-term time series forecasting (LTSF) underpins critical applications from energy management to weather prediction, yet achieving reliable multi-step-ahead accuracy remains challenging. Existing LTSF approaches, dominated by MLP- and Transformer-based architectures, either rely on simple linear mappings or introduce increasingly complex hand-crafted inductive biases, raising the question of whether a more expressive and principled nonlinear core could offer a better alternative. Therefore, we investigate whether Kolmogorov-Arnold Networks (KANs), a recently proposed model featuring adaptive basis functions capable of granular modulation of nonlinearities, can improve LTSF performance, and under which design choices they are most effective. Specifically, we propose KANMixer, a minimal KAN-centered architecture consisting of a multi-scale pooling frontend, a KAN-based temporal mixing backbone, and prediction heads. By avoiding heavy auxiliary modules, KANMixer enables a clear assessment of KAN components in LTSF. Across 28 benchmark-horizon settings against nine baselines, KANMixer achieves the best MSE in 16 settings and the best MAE in 11. Furthermore, extensive ablations on three representative datasets show that KAN effectiveness depends strongly on the choice of edge function; B-spline bases outperform Fourier and Wavelet alternatives; the prediction head contributes most to the gains; moderate depth is preferred over deeper unstable stacks; and decomposition priors help MLP but harm KAN. Beyond practical guidance for integrating KAN into LTSF, these results reveal an underexplored dependency between structural priors and backbone nonlinearity: design choices that benefit MLP can degrade KAN.

General AI Agents are increasingly recognized as foundational frameworks for the next generation of artificial intelligence, enabling complex reasoning, web interaction, coding, and autonomous research capabilities. However, current agent systems are either closed-source or heavily reliant on a variety of paid APIs and proprietary tools, limiting accessibility and reproducibility for the research community. In this work, we present \textbf{Cognitive Kernel-Pro}, a fully open-source and (to the maximum extent) free multi-module agent framework designed to democratize the development and evaluation of advanced AI agents. Within Cognitive Kernel-Pro, we systematically investigate the curation of high-quality training data for Agent Foundation Models, focusing on the construction of queries, trajectories, and verifiable answers across four key domains: web, file, code, and general reasoning. Furthermore, we explore novel strategies for agent test-time reflection and voting to enhance agent robustness and performance. We evaluate Cognitive Kernel-Pro on GAIA, achieving state-of-the-art results among open-source and free agents. Notably, our 8B-parameter open-source model surpasses previous leading systems such as WebDancer and WebSailor, establishing a new performance standard for accessible, high-capability AI agents. Code is available at https://github.com/Tencent/CognitiveKernel-Pro