Inference-time scaling for LLM reasoning has focused on three axes: chain depth, sample breadth, and learned step-scorers (PRMs). We introduce a fourth axis, granularity of external verbal supervision, via Verbal Process Supervision (VPS), a training-free framework that uses structured natural-language critique from a stronger supervisor to guide an iterative generate-critique-refine loop up to a round budget R. Across GPQA Diamond, AIME 2025, and LiveCodeBench V6 (covering both closed and open models), VPS yields three key results. First, on GPQA Diamond, GPT-5.4 (High) | GPT-5.4 (Low) reaches 94.9% at R=4, surpassing the 94.1% state of the art without gradient updates. Second, on AIME 2025, VPS enables strong weak-actor rescue, boosting scores from 11.7-26.7% to 63.3-90.0% (up to +63.3 points). Third, at matched compute, VPS outperforms Reflexion by +8.5 to +12.1 points and Self-Consistency@5 by +5.0 pp (GPQA) and +8.3 pp (LiveCodeBench), isolating critique granularity as the key driver. Performance scales with the supervisor-actor capability gap (Pearson r=0.90) and degrades when errors are not linguistically expressible (e.g., code synthesis), motivating hybrid verbal-executable methods. These results establish critique granularity as a new axis of inference-time scaling.

As LLMs reduce English-centric bias, a surprising trend emerges: non-English responses sometimes outperform English on reasoning tasks. We hypothesize that language functions as a latent variable that structurally modulates the model's internal inference pathways, rather than merely serving as an output medium. To test this, we conducted a Polyglot Thinking Experiment, in which models were prompted to solve identical problems under language-constrained and language-unconstrained conditions. Results show that non-English responses often achieve higher accuracy, and the best performance frequently occur when language is unconstrained, suggesting that multilinguality broadens the model's latent reasoning space. Based on this insight, we propose polyGRPO (Polyglot Group Relative Policy Optimization), an RL framework that treats language variation as an implicit exploration signal. It generates polyglot preference data online under language-constrained and unconstrained conditions, optimizing the policy with respect to both answer accuracy and reasoning structure. Trained on only 18.1K multilingual math problems without chain-of-thought annotations, polyGRPO improves the base model (Qwen2.5-7B-Instruct) by 6.72% absolute accuracy on four English reasoning testset and 6.89% in their multilingual benchmark. Remarkably, it is the only method that surpasses the base LLM on English commonsense reasoning task (4.9%), despite being trained solely on math data-highlighting its strong cross-task generalization. Further analysis reveals that treating language as a latent variable expands the model's latent reasoning space, yielding consistent and generalizable improvements in reasoning performance.

Recent breakthroughs in 3D generative modeling have yielded remarkable progress in static shape synthesis, yet high-fidelity dynamic 4D generation remains elusive, hindered by temporal artifacts and prohibitive computational demand. We present Sculpt4D, a native 4D generative framework that seamlessly integrates efficient temporal modeling into a pretrained 3D Diffusion Transformer (Hunyuan3D 2.1), thereby mitigating the scarcity of 4D training data. At its core lies a Block Sparse Attention mechanism that preserves object identity by anchoring to the initial frame while capturing rich motion dynamics via a time-decaying sparse mask. This design faithfully models complex spatiotemporal dependencies with high fidelity, while sidestepping the quadratic overhead of full attention and reducing network total computation by 56%. Consequently, Sculpt4D establishes a new state-of-the-art in temporally coherent 4D synthesis and charts a path toward efficient and scalable 4D generation.

Modern industrial applications increasingly demand language models that act as agents, capable of multi-step reasoning and tool use in real-world settings. These tasks are typically performed under strict cost and latency constraints, making small agentic models highly desirable. In this paper, we introduce the AgenticQwen family of models, trained via multi-round reinforcement learning (RL) on synthetic data and a limited amount of open-source data. Our training framework combines reasoning RL and agentic RL with dual data flywheels that automatically generate increasingly challenging tasks. The reasoning flywheel increases task difficulty by learning from errors, while the agentic flywheel expands linear workflows into multi-branch behavior trees that better reflect the decision complexity of real-world applications. We validate AgenticQwen on public benchmarks and in an industrial agent system. The models achieve strong performance on multiple agentic benchmarks, and in our industrial agent system, close the gap with much larger models on search and data analysis tasks. Model checkpoints and part of the synthetic data: https://huggingface.co/collections/alibaba-pai/agenticqwen. Data synthesis and RL training code: https://github.com/haruhi-sudo/data_synth_and_rl. The data synthesis pipeline is also integrated into EasyDistill: https://github.com/modelscope/easydistill.

Feature discovery from complex unstructured data is fundamentally a reasoning problem: it requires identifying abstractions that are predictive of a target outcome while avoiding leakage, proxies, and post-outcome signals. With the introduction of ever-improving Large Language Models (LLMs), our method provides a structured method for addressing this challenge. LLMs are well suited for this task by being able to process large amounts of information, but unconstrained feature generation can lead to weak features. In this work, we study reasoning control in LLMs by inducing cognitive behaviors for improving feature discovery. We introduce CoFEE (Cognitive Feature Engineering Engine), a reasoning control framework that enforces cognitive behaviors in how the LLM reasons during feature discovery. From a machine learning perspective, these cognitive behaviors act as structured inductive biases over the space of candidate features generated by the model. These behaviors have been exploited with success in ML models, and include backward chaining from outcomes, subgoal decomposition, verification against observability and leakage criteria, and explicit backtracking of rejected reasoning paths. In a controlled comparison, we show that enforcing cognitive behaviors yields features with higher empirical predictability than those under unconstrained vanilla LLM prompts. CoFEE achieves an average Success Rate Score that is 15.2% higher than the vanilla approach, while generating 29% fewer features and reducing costs by 53.3%. Using held-out feature evaluation, we assess whether cognitively induced features generalize beyond the data used for discovery. Our results indicate that, in our evaluated setting, reasoning control is associated with improvements in quality and efficiency of LLM-based feature discovery.

Fitting an underlying body model to 3D clothed human assets has been extensively studied, yet most approaches focus on either single-modal inputs such as point clouds or multi-view images alone, often requiring a known metric scale. This constraint is frequently impractical, especially for AI-generated assets where scale distortion is common. We propose OmniFit, a method that can seamlessly handle diverse multi-modal inputs, including full scans, partial depth observations, and image captures, while remaining scale-agnostic for both real and synthetic assets. Our key innovation is a simple yet effective conditional transformer decoder that directly maps surface points to dense body landmarks, which are then used for SMPL-X parameter fitting. In addition, an optional plug-and-play image adapter incorporates visual cues to compensate for missing geometric information. We further introduce a dedicated scale predictor that rescales subjects to canonical body proportions. OmniFit substantially outperforms state-of-the-art methods by 57.1 to 80.9 percent across daily and loose clothing scenarios. To the best of our knowledge, it is the first body fitting method to surpass multi-view optimization baselines and the first to achieve millimeter-level accuracy on the CAPE and 4D-DRESS benchmarks.

Spatial transcriptomics (ST) enables spatially resolved gene profiling but remains expensive and low-throughput, limiting large-cohort studies and routine clinical use. Predicting spatial gene expression from routine hematoxylin and eosin (H&E) slides is a promising alternative, yet under realistic leave-one-slide-out evaluation, existing models often suffer from slide-level appearance shifts and regression-driven over-smoothing that suppress biologically meaningful variation. CHRep is a two-phase framework for robust histology-to-expression prediction. In the training phase, CHRep learns a structure-aware representation by jointly optimizing correlation-aware regression, symmetric image-expression alignment, and coordinate-induced spatial topology regularization. In the inference phase, cross-slide robustness is improved without backbone fine-tuning through a lightweight calibration module trained on the training slides, which combines a non-parametric estimate from a training gallery with a magnitude-regularized correction module. Unlike prior embedding-alignment or retrieval-based transfer methods that rely on a single prediction route, CHRep couples topology-preserving representation learning with post-hoc calibration, enabling stable neighborhood retrieval and controlled bias correction under slide-level shifts. Across the three cohorts, CHRep consistently improves gene-wise correlation under leave-one-slide-out evaluation, with the largest gains observed on Alex+10x. Relative to HAGE, the Pearson correlation coefficient on all considered genes [PCC(ACG)] increases by 4.0% on cSCC and 9.8% on HER2+. Relative to mclSTExp, PCC(ACG) further improves by 39.5% on Alex+10x, together with 9.7% and 9.0% reductions in mean squared error (MSE) and mean absolute error (MAE), respectively.

This work focuses on per-video unsupervised action segmentation, which is of interest to applications where storing large datasets is either not possible, or nor permitted. We propose to segment videos by learning in deep kernel space, to approximate the underlying frame distribution, as closely as possible. To define this closeness metric between the original video distribution and its approximation, we rely on maximum mean discrepancy (MMD) which is a geometry-preserving metric in distribution space, and thus gives more reliable estimates. Moreover, unlike the commonly used optimal transport metric, MMD is both easier to optimize, and faster. We choose to use neural tangent kernels (NTKs) to define the kernel space where MMD operates, because of their improved descriptive power as opposed to fixed kernels. And, also, because NTKs sidestep the trivial solution, when jointly learning the inputs (video approximation) and the kernel function. Finally, we show competitive results when compared to state-of-the-art per-video methods, on six standard benchmarks. Additionally, our method has higher F1 scores than prior agglomerative work, when the number of segments is unknown.

Current model training approaches incorporate user information directly into shared weights, making individual data removal computationally infeasible without retraining. This paper presents a three-layer architecture that decouples personal data from shared weights by combining a static base model, composable domain-expert LoRA adapters that shape behavior without imparting user data, and per-user proxy artefacts whose deletion constitutes deterministic unlearning. Evaluation on Phi-3.5-mini and Llama-3.1-8B confirms per-user differentiation in which personal data influences outputs while remaining isolated, verified by a return to baseline after proxy removal (KL divergence of approximately 0.21 nats, 82-89% verification pass rate) and near-zero cross-user contamination. Because user-specific information never enters shared weights, the architecture mitigates model inversion, membership inference, and training-data extraction against shared model components by construction. The approach converts machine unlearning from an intractable weight-editing problem into a deterministic deletion operation that preserves personalization alongside privacy-enhancing guarantees and is compatible with differentially private stochastic gradient descent (DP-SGD) for privacy-preserving shared model improvement.

Autonomous robots deployed in mass casualty incidents (MCI) face the challenge of making critical decisions based on incomplete and noisy perceptual data. We present an autonomous robotic system for casualty assessment that fuses outputs from multiple vision-based algorithms, estimating signs of severe hemorrhage, visible trauma, or physical alertness, into a coherent triage assessment. At the core of our system is a Bayesian network, constructed from expert-defined rules, which enables probabilistic reasoning about a casualty's condition even with missing or conflicting sensory inputs. The system, evaluated during the DARPA Triage Challenge (DTC) in realistic MCI scenarios involving 11 and 9 casualties, demonstrated a nearly three-fold improvement in physiological assessment accuracy (from 15\% to 42\% and 19\% to 46\%) compared to a vision-only baseline. More importantly, overall triage accuracy increased from 14\% to 53\%, while the diagnostic coverage of the system expanded from 31\% to 95\% of cases. These results demonstrate that integrating expert-guided probabilistic reasoning with advanced vision-based sensing can significantly enhance the reliability and decision-making capabilities of autonomous systems in critical real-world applications.

Supply chain resilience and efficiency are vital in industries characterized by volatile demand and uncertain supply, such as textiles and personal protective equipment (PPE). Traditional forecasting and optimization approaches often operate in isolation, limiting their real-world effectiveness. This paper proposes a Hybrid AI Framework for Demand-Supply Forecasting and Optimization (HAF-DS), which integrates a Long Short-Term Memory (LSTM)-based demand forecasting module with a mixed integer linear programming (MILP) optimization layer. The LSTM captures temporal and contextual demand dependencies, while the optimization layer prescribes cost-efficient replenishment and allocation decisions. The framework jointly minimizes forecasting error and operational cost through embedding-based feature representation and recurrent neural architectures. Experiments on textile sales and supply chain datasets show significant performance gains over statistical and deep learning baselines. On the combined dataset, HAF-DS reduced Mean Absolute Error (MAE) from 15.04 to 12.83 (14.7%), Root Mean Squared Error (RMSE) from 19.53 to 17.11 (12.4%), and Mean Absolute Percentage Error (MAPE) from 9.5% to 8.1%. Inventory cost decreased by 5.4%, stockouts by 27.5%, and service level rose from 95.5% to 97.8%. These results confirm that coupling predictive forecasting with prescriptive optimization enhances both accuracy and efficiency, providing a scalable and adaptable solution for modern textile and PPE supply chains.

The problem of probabilistic verification of a neural network investigates the probability of satisfying the safe constraints in the output space when the input is given by a probability distribution. It is significant to answer this problem when the input is affected by disturbances often modeled by probabilistic variables. In the paper, we propose a novel neural network probabilistic verification framework which computes a guaranteed range for the safe probability by efficiently finding safe and unsafe probabilistic hulls. Our approach consists of three main innovations: (1) a state space subdivision strategy using regression trees to produce probabilistic hulls, (2) a boundary-aware sampling method which identifies the safety boundary in the input space using samples that are later used for building regression trees, and (3) iterative refinement with probabilistic prioritization for computing a guaranteed range for the safe probability. The accuracy and efficiency of our approach are evaluated on various benchmarks including ACAS Xu and a rocket lander controller. The result shows an obvious advantage over the state of the art.

Sentence embedding techniques aim to encode key concepts of a sentence's meaning in a vector space. However, the majority of evaluation approaches for sentence embedding quality rely on the use of additional classifiers or downstream tasks. These additional components make it unclear whether good results stem from the embedding itself or from the classifier's behaviour. In this paper, we propose a novel method for evaluating the effectiveness of sentence embedding methods in capturing sentence-level concepts. Our approach is classifier-independent, allowing for an objective assessment of the model's performance. The approach adopted in this study involves the systematic introduction of syntactic noise and semantic negations into sentences, with the subsequent quantification of their relative effects on the resulting embeddings. The visualisation of these effects is facilitated by Concept Separation Curves, which show the model's capacity to differentiate between conceptual and surface-level variations. By leveraging data from multiple domains, employing both Dutch and English languages, and examining sentence lengths, this study offers a compelling demonstration that Concept Separation Curves provide an interpretable, reproducible, and cross-model approach for evaluating the conceptual stability of sentence embeddings.

Under the EU AI Act, translating AI governance requirements into software development practice remains challenging. While AI governance frameworks exist at industry and organizational levels, empirical evidence of team-level implementation is scarce. We address this "Last Mile" Challenge through insider action research embedded within an AI startup. We present a legal-text-to-action pipeline that translates EU AI Act requirements into actionable strategies through internal expert collaboration by extracting requirements from legal text, engaging practitioners in assessment and ideation, and prioritizing implementation through collective evaluation. Our analysis reveals three patterns in how practitioners perceive regulatory requirements: convergence (compliance aligns with development priorities), existing practice (current work already satisfies requirements), and disconnection (requirements perceived as administrative overhead). Based on these patterns, we discuss when governance might be treated genuinely or performatively. Practitioners prioritize requirements that serve end-users or their own development needs, but view verification-oriented requirements as box-ticking exercises. This distinction suggests a translation challenge: regulatory requirements risk superficial treatment unless practitioners understand how compliance serves system quality and user protection. Expert collaboration offers a practical mechanism for transforming governance from external imposition to shared ownership and making previously invisible governance work visible and collective.

Estimating the prevalence of a category in a population using imperfect measurement devices (diagnostic tests, classifiers, or large language models) is fundamental to science, public health, and online trust and safety. Standard approaches correct for known device error rates but assume these rates remain stable across populations. We show this assumption fails under covariate shift and that multicalibration, which enforces calibration conditional on the input features rather than just on average, is sufficient for unbiased prevalence estimation under such shift. Standard calibration and quantification methods fail to provide this guarantee. Our work connects recent theoretical work on fairness to a longstanding measurement problem spanning nearly all academic disciplines. A simulation confirms that standard methods exhibit bias growing with shift magnitude, while a multicalibrated estimator maintains near-zero bias. While we focus the discussion mostly on LLMs, our theoretical results apply to any classification model. Two empirical applications -- estimating employment prevalence across U.S. states using the American Community Survey, and classifying political texts across four countries using an LLM -- demonstrate that multicalibration substantially reduces bias in practice, while highlighting that calibration data should cover the key feature dimensions along which target populations may differ.

Out-of-Distribution (OOD) detection requires sensitivity to subtle shifts without overreacting to natural In-Distribution (ID) diversity. However, from the viewpoint of detection granularity, global representation inevitably suppress local OOD cues, while patch-based methods are unstable due to entangled spurious-correlation and noise. And neither them is effective in detecting compositional OODs composed of valid ID components. Inspired by recognition-by-components theory, we present a training-free Component-Based OOD Detection (CoOD) framework that addresses the existing limitations by decomposing inputs into functional components. To instantiate CoOD, we derive Component Shift Score (CSS) to detect local appearance shifts, and Compositional Consistency Score (CCS) to identify cross-component compositional inconsistencies. Empirically, CoOD achieves consistent improvements on both coarse- and fine-grained OOD detection.

Wheel-legged robots combine the efficiency of wheeled locomotion with the versatility of legged systems, enabling rapid traversal over both continuous and discrete terrains. However, conventional designs typically employ fixed wheels as feet and limited degrees of freedom (DoFs) at the hips, resulting in reduced stability and mobility during legged locomotion compared to humanoids with flat feet. In addition, most existing platforms lack a full upper body with arms, which limits their ability to perform dexterous manipulation tasks. In this letter, we present X2-N, a high-DoF transformable robot with dual-mode locomotion and manipulation. X2-N can operate in both humanoid and wheel-legged forms and transform seamlessly between them through joint reconfiguration. We further propose a reinforcement learning (RL)-based whole-body control framework tailored to this morphology, enabling unified control across hybrid locomotion, transformation, and manipulation. We validate X2-N in a range of challenging locomotion and manipulation tasks, including dynamic skating-like motion, stair climbing and package delivery. Results demonstrate high locomotion efficiency, strong terrain adaptability, and stable loco-manipulation performance of X2-N, highlighting its potential for real-world deployment.

Identifying critical nodes in complex networks is a fundamental task in graph mining. Yet, methods addressing an all-or-nothing coverage mechanics in a bipartite dependency network, a graph with two types of nodes where edges represent dependency relationships across the two groups only, remain largely unexplored. We formalize the CriticalSet problem: given an arbitrary bipartite graph modeling dependencies of items on contributors, identify the set of k contributors whose removal isolates the largest number of items. We prove that this problem is NP-hard and requires maximizing a supermodular set function, for which standard forward greedy algorithms provide no approximation guarantees. Consequently, we model CriticalSet as a coalitional game, deriving a closed-form centrality, ShapleyCov, based on the Shapley value. This measure can be interpreted as the expected number of items isolated by a contributor's departure. Leveraging these insights, we propose MinCov, a linear-time iterative peeling algorithm that explicitly accounts for connection redundancy, prioritizing contributors who uniquely support many items. Extensive experiments on synthetic and large-scale real datasets, including a Wikipedia graph with over 250 million edges, reveal that MinCov and ShapleyCov significantly outperform traditional baselines. Notably, MinCov achieves near-optimal performance, within 0.02 AUC of a Stochastic Hill Climbing metaheuristic, while remaining several orders of magnitude faster.

Lung adenocarcinoma (LUAD) grading depends on accurately identifying growth patterns, which are indicators of prognosis and can influence treatment decisions. Common deep learning approaches to determine the predominant pattern rely on patch-level classification or segmentation, requiring extensive annotations. This study proposes an attention-based multiple instance learning (ABMIL) framework to predict the predominant LUAD growth pattern at the whole slide level to reduce annotation burden. Our approach integrates pretrained pathology foundation models as patch encoders, used either frozen or fine-tuned on annotated patches, to extract discriminative features that are aggregated through attention mechanisms. Experiments show that fine-tuned encoders improve performance, with Prov-GigaPath achieving the highest agreement (\k{appa} = 0.699) under ABMIL. Compared to simple patch-aggregation baselines, ABMIL yields more robust predictions by leveraging slide-level supervision and spatial attention. Future work will extend this framework to estimate the full distribution of growth patterns and validate performance on external cohorts.

Low-cost air quality sensors (LCS) provide a practical alternative to expensive regulatory-grade instruments, making dense urban monitoring networks possible. Yet their adoption is limited by calibration challenges, including sensor drift, environmental cross-sensitivity, and variability in performance from device to device. This work presents a deep learning framework for calibrating LCS measurements of PM$_{2.5}$, PM$_{10}$, and NO$_2$ using a Long Short-Term Memory (LSTM) network, trained on co-located reference data from the OxAria network in Oxford, UK. Unlike the Random Forest (RF) baseline, which treats each observation independently, the proposed approach captures temporal dependencies and delayed environmental effects through sequence-based learning, achieving higher $R^2$ values across training, validation, and test sets for all three pollutants. A feature set is constructed combining time-lagged parameters, harmonic encodings, and interaction terms to improve generalization on unseen temporal windows. Validation of unseen calibrated values against the Equivalence Spreadsheet Tool 3.1 demonstrates regulatory compliance with expanded uncertainties of 22.11% for NO$_2$, 12.42% for PM$_{10}$, and 9.1% for PM$_{2.5}$.

Span-level skill extraction from job advertisements underpins candidate-job matching and labor-market analytics, yet generative large language models (LLMs) often yield malformed spans, boundary drift, and hallucinations, especially with long-tail terms and cross-domain shift. We present SRICL, an LLM-centric framework that combines semantic retrieval (SR), in-context learning (ICL), and supervised fine-tuning (SFT) with a deterministic verifier. SR pulls in-domain annotated sentences and definitions from ESCO to form format-constrained prompts that stabilize boundaries and handle coordination. SFT aligns output behavior, while the verifier enforces pairing, non-overlap, and BIO legality with minimal retries. On six public span-labeled corpora of job-ad sentences across sectors and languages, SRICL achieves substantial STRICT-F1 improvements over GPT-3.5 prompting baselines and sharply reduces invalid tags and hallucinated spans, enabling dependable sentence-level deployment in low-resource, multi-domain settings.

Large Vision-Language Models (VLMs) are increasingly used to evaluate outputs of other models, for image-to-text (I2T) tasks such as visual question answering, and text-to-image (T2I) generation tasks. Despite this growing reliance, the reliability of these Evaluator VLMs remains under explored. In this work, we systematically evaluate the reliability of Evaluator VLMs across both I2T and T2I tasks. We introduce targeted perturbations that degrade output quality along key error dimensions, including object hallucinations, spatial reasoning, factual grounding, and visual fidelity. These perturbations test whether Evaluator VLMs can reliably account for these quality degrading errors in their evaluations. Using a comprehensive benchmark of over 4000 perturbed instances spanning 40 perturbation dimensions, we evaluate 4 prominent VLMs using single-answer scoring, pairwise comparison, and reference-guided paradigms. Our findings reveal that current VLM evaluators exhibit substantial blind spots: they often fail to detect perturbed outputs - in some cases exceeding 50%, struggle particularly with fine-grained compositional and spatial errors, and are often insensitive to hallucinated content that contradicts the input image. Pairwise comparison proves more reliable, though failure rates persist. These results highlight the unreliable nature of current Evaluator VLMs and urge caution in their deployment for benchmarking and development decisions. Code and data have been made publicly available.

In the automatic reassembly of fragments acquired using laser scanners to reconstruct objects, a crucial step is the matching of fractured surfaces. In this paper, we propose a novel local descriptor that uses the Gaussian Mixture Model (GMM) to fit the distribution of points, allowing for the description and matching of fractured surfaces of fragments. Our method involves dividing a local surface patch into concave and convex regions for estimating the k value of GMM. Then the final Gaussian Mixture Descriptor (GMD) of the fractured surface is formed by merging the regional GMDs. To measure the similarities between GMDs for determining adjacent fragments, we employ the L2 distance and align the fragments using Random Sample Consensus (RANSAC) and Iterative Closest Point (ICP). The extensive experiments on real-scanned public datasets and Terracotta datasets demonstrate the effectiveness of our approach; furthermore, the comparisons with several existing methods also validate the advantage of the proposed method.

ASPIC-style structured argumentation frameworks provide a formal basis for reasoning in artificial intelligence by combining internal argument structure with abstract argumentation semantics. A key challenge in these frameworks is ensuring compliance with five critical rationality postulates: closure, direct consistency, indirect consistency, non-interference, and crash-resistance. Recent approaches, including ASPIC$^{\ominus}$ and Deductive ASPIC$-$, have made significant progress but fall short of meeting all postulates simultaneously under a credulous semantics (e.g. preferred) in the presence of undercuts. This paper introduces Deductive ASPIC$^{\ominus}$, a novel framework that integrates gen-rebuttals from ASPIC$^{\ominus}$ with the Joint Support Bipolar Argumentation Frameworks (JSBAFs) of Deductive ASPIC$-$, incorporating preferences. We show that Deductive ASPIC$^{\ominus}$ satisfies all five rationality postulates under a version of preferred semantics. This work opens new avenues for further research on robust and logically sound structured argumentation systems.

While Large Language Models (LLMs) demonstrate remarkable reasoning, complex optimization tasks remain challenging, requiring domain knowledge and robust implementation. However, existing benchmarks focus narrowly on Mathematical Programming and Combinatorial Optimization, hindering comprehensive evaluation. To address this, we introduce OptiVerse, a comprehensive benchmark of 1,000 curated problems spanning neglected domains, including Stochastic Optimization, Dynamic Optimization, Game Optimization, and Optimal Control, across three difficulty levels: Easy, Medium, and Hard. The experiments with 22 LLMs of different sizes reveal sharp performance degradation on hard problems, where even advanced models like GPT-5.2 and Gemini-3 struggle to exceed 27% accuracy. Through error analysis, we identify that modeling & logic errors remain the primary bottleneck. Consequently, we propose a Dual-View Auditor Agent that improves the accuracy of the LLM modeling process without introducing significant time overhead. OptiVerse will serve as a foundational platform for advancing LLMs in solving complex optimization challenges.

Protein-ligand bioactivity data published in the literature are essential for drug discovery, yet manual curation struggles to keep pace with rapidly growing literature. Automated bioactivity extraction remains challenging because it requires not only interpreting biochemical semantics distributed across text, tables, and figures, but also reconstructing chemically exact ligand structures (e.g., Markush structures). To address this bottleneck, we introduce BioMiner, a multi-modal extraction framework that explicitly separates bioactivity semantic interpretation from ligand structure construction. Within BioMiner, bioactivity semantics are inferred through direct reasoning, while chemical structures are resolved via a chemical-structure-grounded visual semantic reasoning paradigm, in which multi-modal large language models operate on chemically grounded visual representations to infer inter-structure relationships, and exact molecular construction is delegated to domain chemistry tools. For rigorous evaluation and method development, we further establish BioVista, a comprehensive benchmark comprising 16,457 bioactivity entries curated from 500 publications. BioMiner validates its extraction ability and provides a quantitative baseline, achieving an F1 score of 0.32 for bioactivity triplets. BioMiner's practical utility is demonstrated via three applications: (1) extracting 82,262 data from 11,683 papers to build a pre-training database that improves downstream models performance by 3.9%; (2) enabling a human-in-the-loop workflow that doubles the number of high-quality NLRP3 bioactivity data, helping 38.6% improvement over 28 QSAR models and identification of 16 hit candidates with novel scaffolds; and (3) accelerating protein-ligand complex bioactivity annotation, achieving a 5.59-fold speed increase and 5.75% accuracy improvement over manual workflows in PoseBusters dataset.

Numerical reasoning over expert-domain tables often exhibits high in-domain accuracy but limited robustness to domain shift. Models trained with supervised fine-tuning (SFT) on specific datasets tend to rely on header-operation shortcuts rather than structural reasoning. We introduce TaNOS, a continual pre-training framework comprising three components: (i) header anonymization to reduce lexical memorization, (ii) operation sketches that provide minimal structural cues, and (iii) self-supervised pretraining that constructs correctness-guaranteed program-question pairs from given tables in a program-first manner. By decoupling domain semantics and numerical operation structure, TaNOS improves the transferability of numerical reasoning. Applied to an 8B instruction-tuned model, TaNOS achieves 80.13% execution accuracy on FinQA with only 10% train data, outperforming SFT baseline (73.97%) with full train data and proprietary models such as GPT-5, Gemini-2.5-Pro. Furthermore, in the domain-shift experiments, TaNOS displays nearly-negligible cross-domain gap (<2pp) when standard SFT shows over 10pp gap. These results suggest that structural guidance with operation sketches, header-agnostic representations, and correctness-guaranteed self-supervision can improve the robustness of numerical reasoning across diverse expert-domain tables.

Multi-modal trajectory generation is essential for safe autonomous driving, yet existing diffusion-based planners suffer from high inference latency due to iterative neural function evaluations. This paper presents MISTY (Mixer-based Inference for Single-step Trajectory-drifting Yield), a high-throughput generative motion planner that achieves state-of-the-art closed-loop performance with pure single-step inference. MISTY integrates a vectorized Sub-Graph encoder to capture environment context, a Variational Autoencoder to structure expert trajectories into a compact 32-dimensional latent manifold, and an ultra-lightweight MLP-Mixer decoder to eliminate quadratic attention complexity. Importantly, we introduce a latent-space drifting loss that shifts the complex distribution evolution entirely to the training phase. By formulating explicit attractive and repulsive forces, this mechanism empowers the model to synthesize novel, proactive maneuvers, such as active overtaking, that are virtually absent from the raw expert demonstrations. Extensive evaluations on the nuPlan benchmark demonstrate that MISTY achieves state-of-the-art results on the challenging Test14-hard split, with comprehensive scores of 80.32 and 82.21 in non-reactive and reactive settings, respectively. Operating at over 99 FPS with an end-to-end latency of 10.1 ms, MISTY offers an order-of-magnitude speedup over iterative diffusion planners while while achieving significantly robust generation.

Crowdsourced pairwise evaluation has emerged as a scalable approach for assessing foundation models. However, applying it to Text to Speech(TTS) introduces high variance due to linguistic diversity and multidimensional nature of speech perception. We present a controlled multidimensional pairwise evaluation framework for multilingual TTS that combines linguistic control with perceptually grounded annotation. Using 5K+ native and code-mixed sentences across 10 Indic languages, we evaluate 7 state-of-the-art TTS systems and collect over 120K pairwise comparisons from over 1900 native raters. In addition to overall preference, raters provide judgments across 6 perceptual dimensions: intelligibility, expressiveness, voice quality, liveliness, noise, and hallucinations. Using Bradley-Terry modeling, we construct a multilingual leaderboard, interpret human preference using SHAP analysis and analyze leaderboard reliability alongside model strengths and trade-offs across perceptual dimensions.

Large language models (LLMs) are increasingly deployed as customer-facing agents, yet evaluating their reliability remains challenging due to stochastic, multi-turn interactions. Current evaluation protocols rely on linear Monte Carlo rollouts of complete agent-user conversations to estimate success. However, this approach is computationally inefficient, repeatedly regenerating identical early prefixes, and often fails to uncover deep failure modes that arise from rare user behaviors. We introduce DIVERT (Diversity-Induced Evaluation via Branching of Trajectories), an efficient, snapshot-based, coverage-guided user simulation framework for systematic exploration of agent-user interactions. DIVERT captures the full agent-environment state at critical decision points and resumes execution from these snapshots, enabling reuse of shared conversation prefixes and reducing redundant computation. From each junction, the framework branches using targeted, diversity-inducing user responses, allowing directed exploration of alternative interaction paths. By focusing evaluation on semantically diverse and underexplored trajectories, DIVERT improves both efficiency and coverage. Empirical results show that it discovers more failures per token compared to standard linear rollout protocols, while expanding the set of tasks on which failures are identified.