Multi-agent large language model (LLM) systems often fail to reliably outperform a single strong model equipped with best-of-N sampling. We argue that a core source of this instability is ill-posed equilibrium selection: current systems specify what information agents share, but not which coordination convention should be selected. We formalize a broad class of such systems as discounted incomplete-information Markov games and show that two common pathologies, oscillation between competing conventions and drift across them, can both induce unstable learning and linear Bayesian regret. To obtain a well-posed target, we introduce the Heterogeneous Quantal Response Equilibrium (HQRE), an entropy-regularized equilibrium concept with agent- and state-dependent temperatures. Under a monotonicity condition, HQRE is unique, admits linearly convergent mirror updates, and yields bounded Bayesian regret; the same condition yields rollout-measurable stability diagnostics. We instantiate this objective in two algorithms: DICE-PC, which coordinates frozen models through prompt-control actions, and DICE-FT, which performs parameter-efficient mirror fine-tuning. Across eleven benchmarks in four domains, DICE improves accuracy-cost trade-offs over strong within-class baselines; on reasoning and planning tasks, DICE-PC improves by 4.3 percentage points on average and DICE-FT by 8.5 points.

Graph neural networks (GNNs) are widely deployed on relational data, yet they can leak sensitive or proprietary information about the training graph adjacency, e.g., social ties, transactions, and interactions. This work studies graph reconstruction attacks (GRA), a form of model inversion that reconstructs the training adjacency from a trained GNN, given different levels of attacker-side information. We first provide a systematic characterization of when and why adjacency becomes recoverable through features, labels, embeddings, and predictions, with leakage modulated by graph homophily, heterophily, and the model's inductive bias. Motivated by these findings, we view GNN inference through a Markov chain approximation lens, treating the layered forward computation as a chain of topology-dependent representations. Building on this view, we develop complementary attack and defense methods. On the attack side, we propose MC-GRA (+), which reconstructs the adjacency by optimizing a surrogate adjacency whose GNN-induced representations align with those of the target model at each layer. On the defense side, we propose MC-GPB (+), which suppresses adjacency-dependent information throughout the representation chain while aiming to preserve classification accuracy under a privacy-utility trade-off. Experiments across homophilic/heterophilic graph benchmarks and GNNs show that our attacks improve reconstruction fidelity over prior methods, while our defenses reduce reconstruction success with only minor accuracy loss.

Humans exhibit remarkable motor agility, enabling a wide range of dynamic skills such as running and jumping, which highlights the great potential of humanoid robots for athletic locomotion. Among athletic sports, long rope skipping requires two rope turners to cooperatively swing the rope while adapting to a player under different jumping rhythms, making it a meaningful yet challenging task for humanoid robots. Although existing methods for humanoid sports have achieved success in single-agent and interaction-free settings, such as running, dancing, and parkour, task scenarios that require precise coordination among multiple participants remain largely unexplored. To this end, we propose Marope, a multi-agent reinforcement learning (MARL) framework for cooperative long rope skipping with multiple humanoid robots. Specifically, Marope adopts a hierarchical reinforcement learning framework for policy training. At the lower level, it learns decentralized rope manipulation policies through MARL, while at the upper level, a centralized scheduling policy is trained to coordinate the execution of the lower-level policies. To improve generalization across different player behavioral styles, Marope further incorporates diverse jumping policies into cooperative game training. We evaluate our approach on Unitree G1 humanoid robots in both simulation and real-world settings. Experimental results demonstrate that Marope outperforms various baselines, achieving more efficient and stable rope manipulation as well as more robust and adaptable cooperation with varied players.

Humanoid behavior foundation models aim to acquire reusable whole-body control policies from broad human motion priors, enabling a single controller to produce diverse and expressive behaviors. However, existing motion-centric foundation policies largely assume that the reference motion is already physically compatible with the robot's surroundings. This assumption breaks when the demonstrator, operator, and robot inhabit different environments: a human motion may specify the intended behavior, but not the footholds, clearance, body height, or contact timing required by the robot's local terrain. We introduce \emph{Perceptive Behavior Foundation Model} (Perceptive BFM), a terrain-aware humanoid control framework that grounds human motion priors in robot-centric perception. The model preserves raw kinematic motion references as the behavioral interface, while using local terrain observations to adapt contacts, posture, and timing. To provide scalable terrain supervision, we develop \emph{terrain-conformal reference synthesis} (TCRS), which converts locomotion-oriented human motion clips into terrain-consistent references through contact-aware foothold construction, foot-geometry-aware swing optimization, support-aware root reconstruction, collision repair, and multi-point inverse kinematics. We then train a blind adapted-reference teacher and transfer its terrain-conformal behavior to a deployed raw-reference student through target-frame action alignment. The student is an identity-gated Transformer tracker whose terrain features enter through residual pathways initialized to preserve the motion-tracking prior and trained to produce local corrections only when needed.

Egocentric RGB-D videos offer a natural source of human dexterous manipulation demonstrations, but existing data is difficult to use for robot learning because object pose, geometry, and contact information are often missing or require pre-scanned object assets. We present EgoAERO, the first framework that learns dexterous manipulation from a single egocentric RGB-D human demonstration without object assets. EgoAERO reconstructs contact-consistent hand-object trajectories through asset-free object tracking and reconstruction, ego motion compensation, and adaptive contact optimization, then converts them into robot policies using two-stage residual learning. We further introduce an online quality assessment mechanism and construct EgoDex-R, a large-scale egocentric dataset with 4.3M RGB-D frames for dexterous policy learning. Simulation and real-world experiments show that EgoAERO enables single-demonstration dexterous manipulation and achieves downstream performance close to CAD-based reconstructions on HOI4D.

Sign language models are predominantly trained with gloss-sequence or text supervision, thereby under-modeling non-lexical and productive constructions. One comparatively tractable instance is spatial indexing: pointing gestures that assign discourse entities to spatial loci for subsequent co-reference, which lexicon-centric objectives largely fail to capture. We present a targeted evaluation of indexing in Sign Language Recognition, showing that despite comprising 10-15% of signing content, indexing is poorly recovered. We introduce a framework for training and evaluating indexing experts, establishing a baseline for index-aware sign language modeling. Our approach decomposes spatial reference resolution into index detection and discourse entity linking. The resulting mention representations enable automatic annotation and non-lexical structure modeling, and serve as an auxiliary indexing expert that augments a frozen SLR model at inference time.

Financial transaction processing requires extracting structured merchant information from noisy, abbreviated bank transaction strings at scale. Our current production system, a LoRA-fine-tuned LLaMA 3.1-8B, achieves 96.95% F1 on this task, but deploying 8-billion-parameter models imposes prohibitive memory, latency, and cost constraints. To identify more efficient alternatives, we conduct a deployment-focused study of 24 model variants spanning four model families: Gemma 3 (270M, 1B, 4B), Qwen 3.5 (0.8B, 2B, 4B), Aya (3.35B), and LLaMA 3.1-8B, systematically evaluating accuracy, inference throughput, training cost, and hardware behavior to assess production suitability. Our findings show that: (1) reproducing the LLaMA 3.1-8B fine-tune with a LoRA rank of 8 achieves 96.75% F1, only 0.20 points below the rank-32 baseline; (2) Qwen 3.5 4B with JSON-only prompting reaches 96.60% F1, within 0.35 points of the 8B baseline while using roughly half the parameters; (3) the 0.8B Qwen 3.5 model achieves 94.75% F1, matching models 2.5-4x larger and offering an attractive latency-accuracy trade-off; (4) chain-of-thought fine-tuning generally improves F1 by 0.3-1.8 points across most models, although Qwen 3.5 4B performs best with direct JSON-only prompting; and (5) Qwen 3.5 Think and Nothink training templates produce nearly identical results (F1 differences <0.004), indicating that explicit reasoning supervision is unnecessary for structured extraction tasks. We further deploy all 14 fine-tuned sub-8B models as Databricks Model Serving endpoints and observe that benchmark performance transfers reliably to production, with an average F1 change of only 0.8 points. Aya 3.35B, based on the Cohere2 architecture, is the sole exception, exhibiting a 3-5 point decline under serving conditions. Based on these results, we provide deployment recommendations across accuracy and latency requirements, ...

AI agents increasingly turn past experience into reusable artifacts such as code, workflows, and procedural memories. Reuse can improve efficiency, but it also creates a lifecycle reliability problem: artifacts that succeed once may fail under environment drift, underspecified tasks, or changing task distributions, especially in web automation. We introduce SKILL.nb, a framework for governing reusable agent workflows with evidence-calibrated lifecycle policies. SKILL.nb uses selective formalization: execution evidence decides which workflow steps should become executable code, which should remain natural-language guided, and when those choices should be revised. Workflows are stored as auditable, versioned notebooks that interleave natural-language guidance, multi-language executable cells, validation gates, fallback paths, and multimodal evidence such as outputs, screenshots, and error traces. At runtime, gate-conditioned execution lets each step run code when its gates validate, or fall back locally when drift invalidates the executable realization. On WebArena-Verified, SKILL.nb achieves 53.7% single-round success, improving over the strongest baseline by 3.9 percentage points. Across three re-executions, it retains 91.7% of initially successful tasks, 15.5 points above the next best method. Under bounded repair, it recovers 72.9% of subsequent failures while limiting post-repair regressions to 4.2%, compared with 15.0% to 17.0% for persistent baselines. It also leads on Mind2Web cross-website and cross-domain splits. In a GitLab migration test, SKILL.nb preserves performance when reusing frozen state learned on GitLab 15.7, with frozen-versus-fresh target-version gaps of -1.7 points on GitLab 16.11 and +0.6 points on GitLab 18.9. These results identify lifecycle governance and gate-conditioned execution as reliability axes beyond one-shot task success.

Diffusion language models (DLMs) offer substantial speed advantages through parallel decoding, but the lack of token dependencies limits generation quality compared to autoregressive (AR) models. Recent progress attempts to bridge the gap via importance sampling, with DLM being the proposal and AR being the target. However, due to the huge gap between their distributions, the sampling requires a large number of particles and is thus expensive to compute. In this paper, we introduce PoE-Bridge, a novel decoding framework that drastically improves generation speed and accuracy by introducing an intermediate distribution to bridge the gap. The distribution is constructed as a Product-of-Experts (PoE) of the DLM proposal and the AR target. With the intermediate distribution, we first use the DLM to draft multiple continuations in parallel, then apply rejection sampling to verify the drafted tokens and move the resulting candidates toward the PoE. We then use importance sampling to further correct the PoE-aligned candidates toward the AR target. We further propose several improved techniques, including mixed-temperature sampling for enhanced diversity and elastic rejection windows for reducing wasted verification. Empirically, PoE-Bridge achieves significantly improved accuracy with $5\times$ speedup over the standard DLM decoding approach, and recovers at least 95% of the target AR model's performance, efficiently advancing most of the quality gap on challenging mathematical reasoning and coding tasks. Our code is available at https://github.com/juntongshi48/poe-bridge.

We present OSMGraphCLIP, a CLIP-style geospatial representation model that learns global location embeddings from freely available OpenStreetMap (OSM) data. OSMGraphCLIP represents geographic environments as heterogeneous graphs of typed OSM features, preserving the topological and semantic relationships among roads, buildings, land-use regions, and points of interest. A multi-scale graph encoder captures both fine-grained local structure and broader landscape composition, and supervises a spherical-harmonics location encoder through a contrastive alignment objective. We evaluate OSMGraphCLIP across a diverse suite of downstream geospatial regression and classification tasks spanning climate, ecology, socioeconomic indicators, public health, land cover, biodiversity, and wildfire forecasting, and show that structured OSM data alone supports strong global location representations across domains. OSMGraphCLIP matches or exceeds satellite-based baselines on the majority of benchmarks, with the most pronounced advantage on socioeconomic and public-health tasks, where OSM's explicit semantic annotation of the built environment encodes patterns of human activity that satellite pixels can only capture indirectly. On ecological and environmental tasks, the model remains closely competitive with imagery-based methods despite using no Earth observation data. Qualitative analysis confirms that the learned embeddings organize geographic space coherently, recovering biome boundaries, urban gradients, and tropical--temperate distinctions from map topology alone.

Large Language Model (LLM) safety has often been evaluated at the behavior level, which provides limited evidence of internal robustness, as these evaluations target outputs rather than representation-level vulnerability under intervention. We formalize this discrepancy as the audit gap: the difference between behavioral safety and robustness under intervention. To study this gap, we construct dissociated models that preserve safe outward behavior while remaining vulnerable in the latent space. We introduce an intervention-based evaluation framework to test model robustness through soft interventions in parameter and latent spaces, including harmful fine-tuning and layer-wise latent perturbations. To formalize the evaluation, we propose the Latent Vulnerability Score (LVS) to measure how easily harmful behavior can be elicited by bounded latent perturbations. Using this evaluation framework, we show that behavioral safety metrics are insufficient measures of representation-level robustness across multiple safely and unsafely aligned state-of-the-art models. Notably, dissociated models show substantially elevated LVSs despite comparable refusal behavior under harmful intervention, with intermediate representations being the most sensitive to intervention. Our results suggest that behavioral safety evaluation alone provides an incomplete picture of model robustness, motivating representation-aware audits of latent vulnerability and observable behavior.

Robotic simulators are a cornerstone of modern research in aerial robotics, serving both as a vehicle for the development of new control algorithms and as the data source for training reinforcement learning (RL) policies. Yet, existing quadcopter learning environments often face a trade-off between physical fidelity, multi-agent support, and the throughput required by modern deep RL pipelines. In this paper, we present MuJoCo-Drones-Gym, an open-source Gymnasium-compatible multi-drone environment built on top of the MuJoCo physics engine. MuJoCo-Drones-Gym supports an arbitrary number of Bitcraze Crazyflie 2.x nano-quadcopters and exposes a modular API for selecting (i)~the physics model (rigid-body MuJoCo, explicit Python dynamics, or any subset of ground effect, blade drag, and inter-drone downwash), (ii)~the action interface (per-motor RPMs, collective normalized thrust, velocity setpoints, or PID waypoint commands), and (iii)~the observation space (kinematic state vectors, RGB / depth / segmentation cameras, or neighbourhood adjacency information). A PettingZoo ParallelEnv wrapper enables drop-in multi-agent reinforcement learning, while a suite of seven task environments, hover, velocity tracking, multi-drone hover, waypoint navigation, formation flight, gate racing, and a generic multi-agent template, demonstrates the breadth of the interface. We describe the environment design, the underlying physics and quadcopter dynamics, and illustrate its use through control and learning examples that mirror those of the closely related gym-pybullet-drones project, while taking advantage of MuJoCo's improved contact handling, rendering, and parallelizability.

The scale of speech anti-spoofing datasets has grown exponentially over the past decade, driven by the assumption that larger data leads to better performance. However, it remains unclear whether indiscriminate scaling commensurately improves model generalization. This study challenges the "scale-first" paradigm by decoupling the impacts of training data scale versus diversity. Through experiments on representative datasets, we report two key findings: (1) Larger is not always better. Expanding data scale excessively under fixed generation methods yields negligible returns and may even degrade cross-domain generalization due to overfitting.(2) Diversity outweighs scale. A smaller composite training set featuring diverse attacks significantly outperforms larger-scale datasets with limited diversity in cross-dataset evaluations. We conclude that future dataset construction should prioritize the diversity of generation methods over scale to effectively enhance model generalization.

Electrocardiogram (ECG) classification models often suffer from severe label scarcity, making semi-supervised learning (SSL) an attractive strategy for reducing annotation costs. In clinical settings, however, unlabeled pools frequently contain out-of-distribution (OOD) anomalies or diagnostic groups absent from the labeled set. Standard SSL forces incorrect pseudo-labels onto these unseen classes, producing overconfident predictions. To address this, we propose SafeECGMatch, a calibration-aware safe SSL framework for single-label ECG classification under label distribution mismatch. Methodologically, SafeECGMatch employs a dual-branch architecture extracting time-frequency latent representations via ECG-specific augmentations. Crucially, it dynamically aligns confidence with empirical accuracy through adaptive label smoothing and temperature scaling, calibrating both the multiclass classifier and the OOD detector across temporal and spectral domains. This joint optimization allows trustworthy OOD rejection and reliable pseudo-labeling. Evaluated on the PTB-XL and PhysioNet/CinC Challenge benchmarks, SafeECGMatch achieves state-of-the-art accuracy and calibration, advancing reliable knowledge discovery in physiological time-series. Code is available at https://github.com/labhai/SafeECGMatch.

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a leading paradigm for enhancing visual reasoning in Multimodal Large Language Models (MLLMs). However, existing RLVR methods optimize primarily for the reasoning outcome, fundamentally overlooking the fine-grained cross-modal coordination required during the generation process. Through token-level analyses and controlled interventions, we reveal that during Chain-of-Thought (CoT) reasoning, MLLMs frequently fail to dynamically alternate between extracting visual evidence and synthesizing textual context-a coordination breakdown that is causally linked to reasoning failures. Motivated by these findings, we propose DyCo-RL, which integrates dynamic cross-modal coordination into RLVR optimization. Specifically, DyCo-RL uses the Fisher-Rao geodesic distance to measure within-modality attention shifts, assigning tokens to either visually-oriented or text-oriented functional roles. It then evaluates the alignment between a token's actual attention allocation and its assigned role, leveraging this score for alignment-guided advantage reweighting during policy optimization. Extensive experiments demonstrate that the algorithm-agnostic DyCo-RL, when applied to Qwen2.5-VL-3B/7B, consistently improves four representative RLVR algorithms across seven benchmarks spanning visual-centric and mathematical reasoning.

Symbolic benchmarks have emerged as a key approach to assess model robustness under minor modifications to STEM-related questions. However, existing symbolic benchmarks mostly remain limited to mathematical reasoning, lack visual grounding, and are predominantly in English. In this work, we introduce Sci-Rho (Science Rhobustness), a dynamic benchmark for visually-grounded STEM problems spanning five subjects and seven languages, comprising 4,242 problem templates (606 per language) crafted by domain experts, including Olympiad medalists. Each template is implemented as executable Python code that generates diverse but equivalent problem instances by varying numerical values, visual patterns, geometric shapes, color schemes, and function types, resulting in 42,420 instances in total, each paired with reasoning steps and ground-truth solutions. We evaluated 17 state-of-the-art VLMs and discovered a noticeable gap between worst-case accuracy (defined as the proportion of problem templates that a model answers correctly across every generated variation) and average accuracy. We also discovered that smaller models show noticeable performance degradation across languages, whereas proprietary and larger models remain robust. Step-level evaluation reflects this same trend, revealing a significant gap between average F1 and worst-case F1 scores. Finally, our inspection of attention heads of a VLM reveals substantial cross-lingual variation in the relative attention allocated to image tokens compared to text tokens. Our work highlights the importance of evaluation beyond static benchmarks as a metric to measure the quality of VLMs.

Cracks are a critical indicator of building health, and early stage identification is fundamental to prevent harmful damages. Advances in deep learning (DL), particularly convolutional neural networks (CNNs), have enabled scalable solutions for automated crack detection. However, CNN performance strongly depends on the availability of large and diverse datasets, which is particularly challenging for complex surfaces such as masonry. Collecting sufficient real data is time-consuming, while publicly available datasets may not be adequate. To address this limitation, we explored generating synthetic crack data, which complements real data and improves training effectiveness. The real dataset consists of masonry crack images collected from buildings in Bologna and surrounding areas. In contrast, the synthetic dataset was generated using a crack overlay tool that adds cracks to background images in a controlled orientation and placement. The real dataset was used to train several DL architectures, to identify the best-performing model (InceptionV4) employed for experiments with generated data. Six training scenarios were tested in InceptionV4 by varying the ratio of real and synthetic data, with evaluation performed on a test set composed of real images using the F1-score and mean Intersection over Union (mIoU) metrics. Results show that training on synthetic data plus a modest addition of 20% real data achieves results comparable to training on real data only. Moreover, the 20/80 scenario (synthetic/real) achieved an 76% F1-score and 80% mean IoU, outperforming the real-only case. As can be seen, the method demonstrates the potential of synthetic data to reduce collection efforts while enhancing crack detection accuracy.

Object navigation requires a robot to search for an unobserved target in an unknown environment by deciding where to explore next under partial observability. Effective search resembles human-like exploration: selectively probing visually promising frontiers while relying on spatial memory to avoid redundant revisits. We propose IntentNav, a spatial-visual imitation framework that learns human-like ObjectNav policies from human demonstrations. To infer high-level search intent from low-level human actions, we introduce Frontier-based Human-Intent Labeling, which looks ahead in human demonstrations and labels the frontier that best explains the demonstrator's future search direction. We construct a spatial-visual candidate space, where BEV memory tracks explored regions, unexplored frontiers, and trajectory history, while egocentric visual memory provides semantic cues for each candidate. A VLM policy is trained to select among these grounded candidates, using Intent-Aligned Objective to encourage consistent and human-like exploration. IntentNav achieves state-of-the-art performance on the MP3D, HM3D-v1 and HM3D-v2 ObjectNav benchmarks. The proposed candidate-level navigation interface transfers zero-shot to wheeled, quadruped, and humanoid robots without further VLM fine-tuning. \href{https://anonymous.4open.science/w/IntentNav/}{Project page}.

We study high-probability regret bounds for online convex optimization (OCO) with strongly convex losses and establish three results that resolve open questions at the intersection of noise adaptivity, feedback structure, and constraint satisfaction. For the full-information setting with sub-Gaussian stochastic gradients, we prove a noise-adaptive high-probability regret bound in which the martingale deviation term scales with the noise level $σ$ rather than the gradient bound $G$, yielding a multiplicative improvement of $G/σ$ over the classical Azuma-Hoeffding baseline. Our analysis introduces an exponential supermartingale argument that bypasses the bounded-difference requirement of Freedman's inequality, enabling direct treatment of unbounded sub-Gaussian noise without truncation artifacts. For bandit feedback, we prove a minimax lower bound: the high-probability regret scales linearly in $\log(1/δ)$, in contrast to the $\sqrt{\log(1/δ)}$ confidence cost under full information. This constitutes a formal separation in the confidence cost of strongly convex OCO across feedback models. Regarding constrained OCO with stochastic constraints satisfying a Slater condition, we provide simultaneous high-probability guarantees for both cumulative regret and long-run constraint violation, achieving $\mathcal{O}(\sqrt{T\log(m/δ)})$ regret and $\mathcal{O}(\sqrt{T}/(ζδ) + m\sqrt{T\log(m/δ)})$ violation. Synthetic experiments corroborate all theoretical predictions.

Vertical federated learning (VFL) is a distributed learning paradigm that leverages vertically partitioned features across isolated parties without sharing raw samples; however, it remains vulnerable to active sample reconstruction attacks. Existing defenses fail to achieve a satisfactory trade-off between model utility and privacy protection, due to either suppressing task-relevant information alongside privacy-sensitive features or relying on end-to-end supervised training to converge the defense module, which exposes the model to early-epoch vulnerability. To address this challenge, we adopt a structural causal model (SCM) insight and construct CausShield. From a task-learning standpoint, causal features within a raw sample are those that are directly relevant and contributory to the learning objective, whereas non-causal features are task-irrelevant but often encode sample-specific private information, thereby facilitating reconstruction. Importantly, we lay a theoretical foundation to prove this insight. CausShield thus decomposes the shared representations between the client and the coordinating server in VFL into task-relevant and task-irrelevant components to ensure full-cycle privacy protection. Nonetheless, the decomposition is inherently challenging due to the dual objectives of preserving model utility while mitigating privacy leakage. We address this via a carefully formulated optimization problem, which is solved through unsupervised representation learning. We further theoretically prove that CausShield preserves the convergence behavior of standard VFL. Extensive experiments compare CausShield against seven SOTAs, including InvL (USENIX Security'25), and evaluate robustness against advanced reconstruction attacks such as URVFL (NDSS'25). Results demonstrate that CausShield consistently outperforms in privacy protection, model utility, and computational efficiency.

As large language model (LLM) agents are integrated into autonomous cloud operations, distributed systems face a semantic reliability problem: proposer agents can generate production mutations, such as modifying IAM policies, opening firewall security groups, or executing data exports, that are syntactically valid and statically authorized but operationally unsafe. Classical distributed consensus protocols replicate deterministic state transitions but do not evaluate the safety of the proposed intent. To address this gap, we introduce Semantic Quorum Assurance (SQA), a control-plane primitive for governing non-deterministic agentic infrastructure. SQA represents proposals as declarative execution contracts bound to cryptographic evidence chains and routes them to a diverse panel of read-only, sandboxed validator agents. SQA aggregates their judgments under a risk-adaptive quorum predicate that enforces model and archetype diversity, adjusts weights based on calibrated assurance scores, and respects archetype-specific vetoes. Admitted proposals execute only through a sovereign execution gate. We instantiate SQA in a cloud-native control plane and formalize a correlated cognitive failure model for non-deterministic validators. On 500 infrastructure-inspired mutation scenarios, with safety results reported on held-out safe/unsafe trials excluding ambiguous scenarios, SQA reduces unsafe approval from 18.5% for single-agent validation to 0.3% while adding median validation latency of 1.45--4.12 seconds across the studied risk buckets.

Existing text-to-SQL benchmarks are largely centered on SQLite, making it difficult to evaluate whether models can generalize across heterogeneous SQL dialects. However, real-world database systems differ substantially in syntax, functions, type systems, and execution semantics, so the same natural language intent often requires dialect-specific SQL realizations. We introduce UniQL, a human-verified benchmark for cross-dialect text-to-SQL evaluation. UniQL aligns 1,534 natural language questions with executable SQL annotations across 16 SQL dialects, yielding 24,544 dialect-specific queries. All dialects share the same intents, aligned schemas and database contents, enabling controlled evaluation of dialect generalization. UniQL is constructed through a hybrid pipeline combining database migration, SQL translation, execution-guided verification, iterative rule summarization, and human validation. Experiments on both open-source and closed-source LLMs show that current models remain far from dialect-universal, with substantial performance variation across database systems and limited transfer from SQLite success to other dialects. These findings highlight the need for aligned cross-dialect benchmarks and more dialect-aware text-to-SQL methods. Code and data are available at https://github.com/JerryGao818/UniQL

Current image editing software often hinges on fixed filters or expert tuning, leaving a gap between amateur users' intent and outcomes. Creations by generative models may contain artifacts, implausible details, or stylistic drift away from photorealism and offer little insight into why an edit was made. We propose IEA, a conversational Image Editing Agent that learns to operate parameterized tools in an explicit, interpretable action space. IEA is trained via a three-stage multitask pipeline: (1) SFT on distilled expert edits, (2) GRPO with rewards for likeness improvement, tool usefulness, and intent summarization, and (3) large-scale synthetic fine-tuning to jointly master image editing, refinement, and user intent summarization. By manipulating 16 editing tools step by step, IEA produces transparent edit traces that can be inspected and debugged. In quantitative experiments, it attains a lower pixel distance on the edit task and a higher ROUGE-L on the summary task than strong baselines. In user studies, it ranks best among tool-calling methods for instruction following while surpassing generative methods in overall perceptual quality. Our results validate interpretable, tool-centric VLMs as a reliable path to human instruction-guided image retouching.

We propose Q-Guided Value-Gradient Matching (Q-VGM), an off-policy reinforcement learning (RL) method that tackles a long-standing challenge in fine-tuning flow-matching vision-language-action (VLA) policies: efficiently improving an expressive flow-matching action expert with respect to a learned Q-function. Effective improvement must exploit the first-order (gradient) information of the critic, but this is difficult for flow policies, because directly back-propagating the value through their multi-step denoising process is numerically unstable at VLA scale, while the tractable action likelihoods required by policy-gradient methods are unavailable under iterative denoising. Existing value-based methods either backpropagate through the full denoising chain, use the critic only at test time without updating the policy, or distill critic-improved actions as terminal labels without supervising the velocity field. Q-VGM sidesteps these issues by leveraging VGG-Flow, a value-gradient view of flow alignment in generative modeling that transforms value gradient into a denoising-time value-gradient field rather than an unstable end-to-end objective. This requires no action likelihoods and no backpropagation through the denoising chain, and operates on a fixed replay buffer. The critic is an action-sensitive Cal-QL ensemble over compact RLT features with per-layer action injection. Q-VGM enables a practical few-shot initialization then learn-from-experience paradigm: starting from a few-shot-SFT pi0.5 VLA, the method leverages self-generated rollout data to substantially improve task performance without additional expert supervision. On LIBERO, Q-VGM raises the average success rate from 75.0% to 92.5%; on RoboTwin 2.0, from 76.4% to 87.2%; and on two real-robot tabletop tasks, from 40.0% to 67.5%, outperforming all same-backbone, same-critic baselines across all three settings.

Accurate 3D instance segmentation in point cloud data is critical for machine vision applications. Recent advancements leverage multiple pre-trained foundation models to generate 3D proposals, followed by the application of proposal aggregation methods, which significantly enhance performance. However, they often produce sub-optimal results due to inherent variations in confidence levels across different segmentation models, resulting in a bias toward the model with higher confidence. This bias is inherently model-dependent and is influenced by factors such as data preprocessing techniques and training strategies. To address this bias, we propose a novel, training-free 3D instance segmentation approach via Geometric Visual Correspondence (GVC-Seg), which exploits the correspondence between 3D geometric cues and 2D visual cues to mitigate the confidence bias. Additionally, a 3D proposal generation module and a mask-aware CLIP feature extraction module are introduced during the instance mask generation and instance semantic reasoning, respectively. In this way, GVC-Seg enhances proposal quality assessment, ensuring unbiased ensemble learning across different models. Extensive experiments demonstrate that our method achieves state-of-the-art performance on several challenging benchmarks, while also exhibiting strong potential in open-vocabulary semantic segmentation settings.

Catastrophic forgetting, the abrupt loss of previously acquired knowledge upon learning new information, remains the central challenge in Continual Learning. This project investigates whether the order in which a model learns information affects how well it retains knowledge. Specifically, we ask: does learning general categories first (like "animals" vs "vehicles") before learning specific classes (like "dog" vs "cat") reduce forgetting compared to learning all classes at once? We test three approaches on CIFAR-100: (1) Coarse-to-Fine: train on 2 super-classes, then expand to 10 specific sub-classes, (2) Fine-to-Coarse: train on 10 sub-classes, then group into 2 super-classes, and (3) Flat: train on all 10 classes from the start. We use Elastic Weight Consolidation (EWC) to prevent forgetting during transitions. Our hypothesis is that learning general patterns first creates a stable foundation that helps the model retain knowledge when learning more detailed distinctions. We evaluate using standard metrics (accuracy, precision, recall, F1) plus continual learning metrics like backward transfer and forgetting rates. This work could inform how we design learning sequences for real-world systems that need to learn incrementally.

Underwater instance segmentation integrates pixel-level mask prediction and instance-level discrimination for marine resource exploration, ecological monitoring, and underwater robotic perception. Recent prompt-based and auxiliary-modality methods improve mask quality, but their reliance on large foundation models, prompt generation, or extra modality estimation complicates efficient deployment. This work introduces Lightweight Underwater Salient Instance Segmentation Detection Transformer (LUSIS-DETR), a compact detection-transformer framework built around the Aqua Boundary-Saliency Attention Module (AquaBSAM). AquaBSAM embeds underwater boundary, contrast, attenuation, chroma, dark-channel, and center-prior cues into DINOv2-initialized multi-scale features through bounded residual modulation, while auxiliary mask supervision and small-object copy-paste are training-only. Extensive evaluation on four recent underwater instance segmentation datasets, UIIS, UIIS10K, USIS10K, and USIS16K, shows competitively leading performance against previous state-of-the-art works across category-aware and salient-instance protocols. TensorRT half-precision (FP16) benchmarking on an NVIDIA T4 graphics processing unit (GPU) achieves 4.31-6.34 milliseconds (ms) latency, supporting real-time inference under an accessible reproduction setting.

The progress of large language models (LLMs) has fueled claims that model-generated summaries rival or even surpass human-written references, raising questions about whether summarization remains an open research problem. We re-examine this narrative through a multi-track evaluation covering five diverse datasets and five state-of-the-art LLMs, combining controlled human assessment, bias-mitigated LLM-as-Judge protocols, factuality verification against external knowledge, and corpus-level linguistic analysis. Our findings reveal a more nuanced landscape in which human reference summaries continue to demonstrate advantages in informativeness and faithfulness, whereas LLM outputs are preferred mainly for surface-level coherence and fluency. Factuality verification indicates that human references remain more reliable, particularly for claims involving reasoning or synthesis, and linguistic analysis uncovers a pattern of stylistic homogeneity across different models. These observations suggest that current LLMs have raised the floor of summarization quality, but the ceiling of their performance remains below human capabilities.

Pretraining is fundamental to the development of Large Language Models (LLMs), yet the opacity of pretraining data complicates model analysis and raises ethical, legal, and fairness concerns. Detecting whether specific datasets were used during pretraining is, therefore, critical. Existing state-of-the-art methods typically rely on access to model probability distributions, making them unsuitable for closed-source LLMs that provide only input-output interfaces. To address this limitation, we introduce Masked Corpus-level Pretraining Data Detection (MC-PDD), a novel method inspired by the masked language modeling paradigm. MC-PDD masks highly specific tokens in each text and prompts the LLM to predict the missing content. It then assesses whether the difference in prediction hit rates between a candidate corpus and a reference non-member corpus is statistically significant. Based on this comparison, MC-PDD determines whether the candidate texts were likely included in the model's pretraining data. Experimental results demonstrate clear and consistent differences in prediction hit rates between pretrained and unseen data across three datasets, for both open-source and closed-source LLMs. Despite operating under a stricter black-box setting, MC-PDD achieves performance comparable to existing detection methods. Our approach enables practical applications such as model auditing and data copyright verification using only standard API access. Upon acceptance, we will publicly release the code and datasets.

Understanding customer shopping trajectories is essential for enabling personalized shopping experiences. However, shopping records (i.e., customer's search, clicks, purchases, etc.) often span long time horizons over multiple years, resulting in extremely long trajectories that pose significant challenges for existing large language models (LLMs). Despite the importance of this problem, existing benchmarks are limited to short customer trajectories, while real-world trajectories from large e-commerce platforms are rarely accessible due to data privacy constraints. To address this gap, we introduce ShopTrajQA, a long-context evaluation benchmark constructed from real-world product information and simulated shopping trajectories. The dataset includes variants of up to 32k and 64k tokens, enabling systematic evaluation of model robustness under varying context lengths. Through comprehensive benchmarking of frontier LLMs, we identify critical performance gaps in reasoning over long shopping trajectory data. To address these challenges, we propose a Customer Agent Framework for ultra-long context management. Leveraging a Reinforcement Learning with Verifiable Rewards (RLVR) agentic training paradigm, our approach stores trajectories as external local files and trains the agent to autonomously retrieve and parse them through code-interpreter interactions (e.g., SQL queries), effectively bypassing the fixed in-context window constraints of LLMs. Experimental results demonstrate that our framework achieves strong performance for ShopTrajQA and shows generalization to other complex reasoning tasks.