Validating an LLM judge against human annotations usually means reporting several agreement statistics: accuracy, precision, recall, $F_1$, Cohen's $κ$, and one or more rank correlations. A survey of 24 recent LLM-as-judge papers finds metric choice entangled with the judgment scale, tie handling, invalid outputs, and abstention handling, and those choices rarely stated. For binary criteria -- the common case in rubric-based evaluation, where each criterion is graded MET or UNMET -- most of the reported numbers are redundant: Pearson's $r$, Spearman's $ρ$, Kendall's $τ_b$, the phi coefficient $ϕ$, and the Matthews Correlation Coefficient all reduce to a single number on non-degenerate binary data, so reporting several of them only creates an illusion of corroborating evidence. Cohen's $κ$ is the one agreement coefficient that adds information: it shares $ϕ$'s numerator but normalizes differently, and the gap between them measures how far the judge's positive-label rate has drifted from the human's. We then trace what changes when a judge may abstain with a CANNOT_ASSESS verdict: the three common ways of handling abstentions are not interchangeable preprocessing choices but answer different questions, and they break the binary equivalences. The same equivalences reappear, up to a negligible finite-sample correction, for multi-judge ensembles scored with Fleiss' $κ$ or Krippendorff's $α$. We close with a reporting checklist that names the judgment scale, the abstention and tie handling mode, coverage, the confusion matrix, and the aggregation level alongside any scalar agreement coefficient.

Weakly-supervised classification of whole-slide images with attention-based multiple instance learning (ABMIL) on top of foundation features now reaches near-saturation on Camelyon16 slide-level performance, but the corresponding attention maps are an imperfect localization signal: in clinical interpretation, a model that classifies correctly without firing on the actual lesion is hard to trust. We address this gap with cellular sheaves, which equip each vertex and edge of a graph with a finite-dimensional vector space and consistent linear maps between them, providing a principled way to detect local disagreement on graph-structured data. We apply cellular sheaves to weakly-supervised tumour localization on whole-slide images, combining a sheaf disagreement field with ABMIL. The natural training objective, encouraging consistency between similar features, produces a disagreement field that tracks tissue-level texture rather than diagnostic content. We propose attention-conditional consistency, which uses the classifier's attention to define which neighbouring patches should agree. Joint training of the classifier and the sheaf under this objective produces a disagreement field with patch-level AUC 0.940 on Camelyon16 and raises the attention head from its ABMIL-alone level of 0.717 to 0.953. Two-stage ablation with the classifier frozen at its ABMIL values reaches only 0.727 on the disagreement field and leaves attention at 0.717, confirming that the gain comes from the projector co-adapting under both objectives, not from the loss change in isolation. The trained model transfers without retraining to annotated slides from Camelyon17, maintaining Delta AUC 0.932 +/- 0.083 and attention AUC 0.955 +/- 0.099. The result is an attention map and a sheaf-disagreement map that fire on the same diagnostic regions, giving clinicians two complementary explanations for each slide-level prediction.

Joint Embedding Predictive Architectures (JEPAs) have reshaped self-supervised representation learning in vision. The recent LLM-JEPA ported JEPA to autoregressive language models but inherited two steep costs from the causal-attention substrate: it demands explicit multi-view data (e.g., text-code pairs), and it requires two gradient-carrying forward passes per step. We introduce DLLM-JEPA, which pairs JEPA with masked-diffusion language models to eliminate both costs at once. The bidirectional attention of diffusion models yields two semantically distinct views of the same input via different masking rates -- no explicit pairs needed -- and supports a single gradient-carrying forward pass, cutting training FLOPs by 33% relative to LLM-JEPA. DLLM-JEPA improves over diffusion-only fine-tuning in every (task, architecture) combination we evaluate: up to +18.7 pp on LLaDA-8B GSM8K and +11.4 pp on Dream-7B GSM8K, with consistent positive gains on Spider, NL-RX-SYNTH, and Django. Beyond accuracy, DLLM-JEPA exhibits a dual-win property: on LLaDA-8B with the Wide-t configuration, it simultaneously raises GSM8K accuracy (67.1 vs. 65.2, +1.8 pp), drives held-out Wikitext loss below the pre-trained base, and preserves MMLU accuracy at base level across three fine-tuning seeds -- whereas an L2-to-base parameter anchor matches baseline accuracy with no task gain. Layer-wise probing reveals the mechanism: a geometric-functional drift dissociation in which the fine-tuned backbone moves further from the pre-trained weights than the baseline yet forgets less on held-out Wikitext, with the amplification concentrated in middle transformer layers. The pattern appears on Dream-7B as well, indicating the phenomenon is not specific to a single backbone.

Physical AI systems increasingly map multimodal observations, language instructions, and learned world representations into physically consequential actions. Robotics foundation models, vision-language-action models, and world-model-based autonomous systems can condition decisions that move vehicles, robots, drones, and industrial machines. This transition exposes a safety problem that is not fully captured by conventional AI content moderation or by classical robot safety alone: a black-box model may issue a physically consequential action while appearing confident, plausible, and semantically aligned. The resulting failure can be silent, arising from sensor drift, occlusion, state-estimation error, distribution shift, hallucinated affordances, or invalid physical assumptions before downstream hardware controllers detect a violation. Across embodied foundation models, world models, robotics simulation, embodied safety benchmarks, safe control, runtime assurance, uncertainty estimation, verification, and guardrail evaluation, model capability and safety mechanisms have advanced along largely separate technical tracks. A recurring gap synthesized here is that no single stream surveyed in this review supplies a complete runtime authorization boundary between black-box Physical AI models and physical execution. The resulting analysis develops a bounded problem formulation, a definition of silent physical-action failure, a taxonomy of runtime guardrail functions, and evaluation requirements for comparing guardrails as Physical AI assurance mechanisms.

Predictive Physical AI systems output state rollouts, action chunks, and latent plans, yet a low root-mean-square error (RMSE) does not imply that a particular proposal is physically executable. We formulate physical admissibility as a prediction-control interface: before execution, a decoded proposal is treated as candidate dynamics and evaluated using kinematic, dynamic, and direct-to-composed horizon conditions. Passing is not a certificate of task success; rejection identifies violation of the specified physical envelope and gives a component-level reason. On Hugging Face LeRobot PushT, controlled falsification shows that one-step prediction-RMSE and standardized dynamics residuals reach area under the receiver operating characteristic curve (AUC) 0.982 and 0.972, kinematic-only conditions reach AUC 0.592, and the full gate reaches AUC 0.957 with condition-level attribution. In replay-based intervention experiments, residual-based filters and the full physical-admissibility gate prevent 87-$89% of invalid proposals while preserving mean progress near 0.998.

Effective pre-polysomnography screening for obstructive sleep apnea-hypopnea syndrome (OSAHS) requires combining clinical risk factors with visible craniofacial and neck cues. Directly prompting general-purpose multimodal foundation models for medical yes/no decisions can yield unstable, poorly calibrated outputs. We propose EviOSAHS, an evidence-grounded multimodal reasoning framework that separates image-only anatomical evidence acquisition from final clinical adjudication. Each frontal facial image is decomposed into seven fixed anatomical queries covering the neck, chin, mouth, face/neck fat, lower jaw, midface, and nose. Visual responses are converted into structured evidence cards recording target anatomy, visibility, risk direction, evidence strength, confidence, and a concise summary. These cards are combined with a cleaned clinical profile only in the final stage, where a large language model performs balanced binary screening adjudication. We evaluated EviOSAHS on a 642-subject cohort, mapping normal subjects to screening-negative and mild, moderate, or severe OSAHS subjects to screening-positive. EviOSAHS achieved 88.47% accuracy, 94.86% sensitivity, 93.74% F1-score, and a 5.14% false-negative rate, outperforming clinical-only prompting, direct multimodal prompting, and naive two-stage pipelines under a unified protocol. Ablations showed that seven-question visual decomposition and balanced final adjudication were critical to the high-sensitivity operating point. A question-level audit of 4,494 visual outputs showed a 100% structured parse rate and 93.88% high-visibility rate. EviOSAHS provides an auditable, high-sensitivity workflow for binary pre-polysomnography OSAHS screening, but should be viewed as a triage assistant rather than a diagnostic system. Prospective validation, external testing, and calibrated operating-point control are needed before clinical deployment.

Inverse kinematics (IK) is a fundamental problem in robotics, requiring the generation of joint configurations that satisfy target end-effector poses. Existing approaches often struggle to generalize across diverse robot morphologies and to effectively model the multi-modal nature of IK, particularly in articulated systems with multiple kinematic branches. In this work, we propose GraphDiff-IK, a structure-aware graph diffusion framework for inverse kinematics. Specifically, we represent the robot as a kinematic graph constructed from the robot URDF, where nodes correspond to actuated joints and edges encode kinematic dependencies. Building upon this representation, we formulate IK as a conditional graph diffusion process that directly generates joint configurations on the robot graph. To better capture structural dependencies in articulated systems, we further introduce a structure-aware graph reasoning framework with hierarchical stage-wise message passing and torso-aware conditioning for multi-branch robots. In addition, we incorporate noisy forward kinematics feedback and task-space supervision to improve geometric consistency during denoising. The proposed framework provides a unified formulation that naturally supports single-arm robots, dual-arm systems, and articulated robots with torso or waist structures. Extensive experiments on diverse robotic platforms demonstrate that the proposed method achieves accurate and stable IK performance while preserving the ability to generate multiple feasible solutions for redundant robotic systems.

Model Predictive Control (MPC) for autonomous navigation faces a fundamental trade-off between model accuracy and real-time efficiency. High-fidelity dynamics models can accurately predict complex vehicle-terrain interactions during trajectory rollouts, but incur significant computational cost, increasing inference latency and reducing control frequency. Conversely, lightweight models enable fast updates and dense sampling, yet may produce erroneous predictions under safety-critical conditions, potentially leading to catastrophic failures such as vehicle rollover. To address this trade-off, we propose Adaptive Dynamics Orchestration (ADO), a framework that dynamically selects the most appropriate dynamics model for the current navigation context. ADO maintains a library of models spanning diverse accuracy-efficiency profiles and continuously refines terrain-conditioned performance estimates using residual errors from online counterfactual rollouts, where executed control actions are replayed across the model library to assess predictive discrepancy. These estimates guide model selection in real time, balancing computational efficiency and predictive accuracy. Real-world experiments on an off-road ground robot demonstrate that ADO significantly reduces modeling error compared to a fixed low-latency baseline, while approaching the accuracy of the highest-fidelity model without incurring its computational cost, resulting in more reliable and effective navigation in challenging terrain.

Mixture-of-Experts (MoE) large language models reduce per-token computation through sparse expert activation, but their deployment remains memory-intensive because all expert weights must be kept resident in memory. Existing MoE compression methods struggle in the ultra-low-bit regime: pruning irreversibly removes model capacity, while coarse-grained quantization fails to allocate bits according to heterogeneous expert and weight-direction importance. We propose BitsMoE, a spectral-energy-guided bit-allocation framework for MoE LLM quantization. BitsMoE decomposes each MoE layer by SVD into a shared basis and expert-specific spectral factors, retaining the shared basis without quantization to preserve common cross-expert structure and using the expert-specific factors as fine-grained quantization units. To determine the bit-width of each unit, BitsMoE formulates spectrum-wise mixed-precision quantization as an activation-aware reconstruction surrogate and solves an integer linear program that minimizes estimated reconstruction loss under a fixed bit budget. Experiments across multiple MoE LLMs show that BitsMoE substantially reduces downstream task accuracy degradation in ultra-low-bit regimes. Under 2-bit quantization on Qwen3-30B-A3B-Base, BitsMoE accelerates quantization by 12.3$\times$, improves average accuracy by 27.83 percentage points, and increases decoding speed by 1.76$\times$ over GPTQ. Our model and code are publicly available at https://github.com/zjiayu064/BitsMoE.

Numerous modern applications in signal processing and medical imaging necessitate acquiring high-dimensional signals under tight resource constraints. Traditional sampling theory suggests that accurate signal reconstruction requires a number of measurements proportional to the signal's ambient dimension, a requirement often too expensive or impractical. Compressed sensing challenges this notion by demonstrating that sparse signals can be recovered with fewer measurements, provided the measurement operator meets certain conditions. This proof-of-concept study presents a task-aware flow-based generative framework -- a reformulation of the conventional Flow Matching training paradigm with a flow model trained to optimize subsampling in compressed sensing applications. We establish the fundamental feasibility of the proposed framework of learning subsampling masks that substantially enhance the performance of compressed sensing for image classification, image reconstruction, and MRI acceleration. For the image reconstruction task, our method demonstrated state-of-the-art performance, achieving Peak Signal-to-Noise Ratio of 25.17 dB at the subsampling rate of 5\% on the CelebA dataset and 29.24 dB when reconstructing $8\times$ accelerated MRI measurements (fastMRI dataset) with the minimal computational overhead. These results highlight the effectiveness of task-conditioning within generative flow models and reveal a promising direction for representation learning strategies. Overall, the proposed framework offers a unified, flexible approach to designing data- and task-driven sensing schemes that can be potentially adapted to a broad range of inverse problems.

Recently, Belief-Attention \cite{Guoqiang25BeliefAttention} has been proposed by first performing an orthogonal projection of the softmax-based weighted summation of $V$ vectors with respect to the original $V$ vectors and then taking the perpendicular component as the residual signal in Transformer for performance improvement. In this paper, we first conduct an ablation study showing the projected component also carries information about the token correlation, which should not be ignored. We then propose to extend Belief-Attention by making use of both the perpendicular and projected components. In particular, the projected component goes through certain activation function and then a linear mapping before merging with the considered token. Conceptually speaking, the neural block for the projected component can be viewed as a two-layer feedforward network (FFN) within the new attention block. It is also noted that standard attention captures the token correlation via the inner-product matrix $QK^T$. We propose to introduce an additional inner-product matrix $ZZ^T$ to $QK^T$ to capture richer token correlation. We refer to the new module as Belief2-Attention. It can be easily shown that Belief2-Attention is more expressive than standard Attention. We then verify the effectiveness of Belief2-Attention for vision tasks of image classification and segmentation.

The present work introduces DefocusTrackerAI, a generalized deep-learning framework for the automatic detection and position estimation of defocused particle images from any kind of optical configuration without compromising uncertainty and recall, intended as a follow-up of the open-source project DefocusTracker. We selected the deep neural network architecture from the direct comparison of two well-known object detection models, Faster R-CNN and YOLOv9, trained on a diverse and feature-rich synthetic image set containing astigmatic and non-astigmatic defocused particle images of varying diameters. The model evaluation on synthetic data showed that, first, YOLOv9 outperforms Faster R-CNN, achieving higher recall and lower uncertainty, particularly at high particle image densities; and second, that YOLOv9 provides enhanced spatial resolution, with uncertainty values between 0.1 and 0.4 pixels for particle image densities N_s up to 0.5, outperforming state-of-the-art algorithms. We demonstrated that our models are able to detect astigmatic and non-astigmatic defocused particle images in multiple optical setups with varying lighting conditions. In addition, we successfully applied our models on real DPT experiments, including fluorescence and shadowgraph data, showing that they can be used beyond conventional DPT applications, including the tracking of sprays and droplets. A pre-trained, ready-to-use version of DefocusTrackerAI based on YOLOv9 is available at https://gitlab.com/goncalo.coutinho/defocustrackerAI-main/-/tree/7e0f11f649ebad50e20dca5b9545f26ca303ebe0 and can be used for automatic detection of defocused particle images of any kind with high accuracy. In combination with a suitable calibration approach for the depth position, it can be used as an effective first step for three-dimensional defocusing particle tracking.

LiDAR semantic segmentation is a core perception capability for autonomous vehicles and mobile robots. However, safe operation also depends on knowing when predictions are unreliable. Existing approaches typically rely on softmax confidence, which is often miscalibrated and overconfident, while stronger uncertainty estimates from Monte Carlo dropout or ensembles are often computationally expensive for real-time use. To this end, we introduce a novel, architecture-agnostic uncertainty-aware Adapter Head. It decomposes the prediction into a Preference Head for class ranking and a Strength Head that refines uncertainty assessment, thereby enabling a principled construction of evidential Dirichlet representations. Building on this design, we propose our inverse-vacuity self-calibration objective (Invascal), which directly supervises the strength signal to produce reliable and well-calibrated uncertainty estimates while preventing runaway evidence growth. We evaluate our framework across multiple LiDAR datasets and backbone architectures. We compare against deterministic training, Monte Carlo dropout and ensembles, and prior evidential methods. Our approach consistently improves uncertainty calibration over traditional deterministic methods with minimal computational overhead. At the same time, it preserves competitive segmentation accuracy, where prior evidential methods often suffer performance degradation.

Diffusion and flow-matching based text-to-speech (TTS) models excel in naturalness but often lack explicit emotion control, as emotional signals remain entangled with speaker identity. We discover that emotion embedding emerges as a linearly decodable direction of frozen hidden states, nearly orthogonal to the direction embedding speaker identity. This inspires a plug-and-play framework DUET for emotion control over pretrained diffusion and flow-matching based TTS models. During generation, DUET unifies dual-space control to achieve fine-grained emotion intervention in a single per-step update: hidden space steering shifts generation along the target emotion direction, while mel-space guidance refines spectral details through gradients backpropagated from a differentiable vocoder. We validate DUET on five architecturally diverse pretrained TTS backbones across three datasets, where it outperforms 10 supervised state-of-the-art emotional TTS baselines across paradigms and achieves the highest human-rated emotion appropriateness. To further showcase its qualitative behavior, we deploy DUET on an Ameca humanoid robot, where it produces richly expressive emotional speech on the humanoid, demonstrating the strong potential for plug-and-play affective interaction for embodied agents.

Systems moving in low Reynolds number fluid regimes are known to be governed by a ``motility map'' which linearly relates their shape change rates to they body frame velocity moving through the fluid. A consequence of this is ``Purcell's Scallop Theorem'' -- a locomotion system that undergoes shape changes that follow the same path forward and backward in time (reciprocal body deformations) cannot achieve net displacement, regardless of pacing of those changes.We show that linear-in-velocity motility maps extend to any power law viscosity (a.k.a. Ostwald--de Waele fluid), and therefore to many biological fluids in intermediate shear ranges. We also show that the linear-in-velocity property can be violated in Carreau-Yasuda fluids to produce net motion using an ``inchworm'' model consisting of two unequal masses with unequal drag coefficients performing reciprocal motions. Interestingly, the direction of motion can be switched by changing speeds. Our results show that the linear motility map of geometric mechaincs can be used to analyze and design locomotion in power-law fluids, and that some nonlinear drag relationships such as Carreau-Yasuda can be exploited to generate net locomotion in seeming violation of the ``scallop theorem''.

Retrieval-Augmented Generation (RAG) has become foundational for grounding large language models in domain-specific corpora, yet conventional vector-based RAG systems are fundamentally limited in their ability to capture the structured, multi-entity relationships that underpin financial market analysis. This paper presents a comprehensive comparative study of a novel two-hop Graph-RAG architecture versus a standard vector-only baseline for cross-entity financial sentiment analysis. Our system constructs a sentiment-weighted knowledge graph of 59 equity entities from 255 news articles covering 10 major technology stocks, then augments dense retrieval with intensity-filtered graph traversal over INFLUENCES edges to surface relational evidence inaccessible to vector search alone. We evaluate both architectures on 100 grounded queries (30 Direct, 70 Relational) using semantic similarity, entity recall, RAGAS metrics, latency benchmarks, and ablation studies. Graph-RAG achieves a statistically significant improvement in entity recall (+6.4%, p < 0.001, Wilcoxon signed-rank) and delivers substantially more relevant answers for complex multi-entity queries (+11.7% Answer Relevancy), with gains concentrating in relational question types (+16.1%). Critically, these improvements come at no measurable cost to answer quality (delta = +0.001 semantic similarity, Cohen's d = 0.078), with a modest 22.6% increase in mean latency offset by an 80% reduction in latency variance. An ablation study on the graph traversal intensity threshold reveals an inverted-U relationship with answer quality, identifying tau = 0.5 as optimal over the production default of tau = 0.7. These findings characterize a precision-for-coverage trade-off inherent to graph-augmented retrieval and provide actionable architectural guidance for practitioners building RAG systems for multi-entity financial analysis.

Informative excitation signals are critical for accurate system identification of mechatronic systems, yet classical system identification (SI) approaches require expert knowledge and hand-crafted signal design to respect hardware safety constraints, limiting their generalizability. We propose a reinforcement learning (RL) agent that learns optimal excitation signals for a Quanser Aero 2 testbed while autonomously enforcing safety constraints through reward shaping. Evaluated across 10 independent training seeds, our comprehensive agent achieves competitive estimation accuracy across all three identified parameters, outperforming classical baselines while incurring only 0.75% safety violations.

Recent progress in generalizable embodied control has been driven by large-scale pretraining of Vision-Language-Action (VLA) models. However, most existing approaches rely on large collections of robot demonstrations, which are costly to obtain and tightly coupled to specific embodiments. Human videos, by contrast, are abundant and capture rich interactions, providing diverse semantic and physical cues for real-world manipulation. Yet, embodiment differences and the frequent absence of task-aligned annotations make their direct use in VLA models challenging. This survey provides a unified view of how human videos are transformed into effective knowledge for VLA models. We categorize existing approaches into four classes based on the action-related information they derive: (i) latent action representations that encode inter-frame changes; (ii) predictive world models that forecast future frames; (iii) explicit 2D supervision that extracts image-plane cues; and (iv) explicit 3D reconstruction that recovers geometry or motion. Beyond this taxonomy, we highlight three key open challenges in this area: structuring unstructured videos into training-ready episodes, grounding video-derived supervision into robot-executable actions under embodiment and viewpoint heterogeneity, and designing evaluation protocols that better predict real-world deployment performance and transfer efficiency, thereby informing future research directions. A curated list of papers and resources is available at https://github.com/AaronFengZY/HumanCentricToVLA-Survey.

Vision-Language-Action (VLA) models follow a data-driven paradigm and are constrained by the coverage of training data, making them prone to failure on edge-case configurations after deployment. To mitigate such risks, it is essential to expose high-quality failure modes and convert the resulting failures into supervisory data for model enhancement. Existing studies largely stop at failure detection and lack a mechanism for leveraging discovered failures for model repair. We propose VLAMotor, the first analysis framework for VLA enhancement, which integrates distance-aware model testing for failure exposure and agent-based data synthesis for model finetunning. First, VLAMotor estimates input uncertainty based on the distance to training samples, and combines uncertainty ranking with redundancy elimination to build compact test sets that expose diverse failures. Then, VLAMotor abstracts failure trajectories into structured semantic representations, and plans parameterized repair-skill sequences, which are then realized as executable trajectories through inverse kinematics and motion execution. The resulting successful trajectories are automatically labeled and used to fine-tune the original VLA model, yielding an enhanced VLA model. Evaluation on four representative robotic manipulation tasks shows that 92.33% of the in-simulation test cases generated by VLAMotor trigger VLA failures, and VLAMotor improves test coverage over the state-of-the-art tool by 18.93%. By fine-tuning VLA models with synthetic data derived from failed test cases, VLAMotor further enhances the overall success rate of VLA models by 49.25%. When deployed on real hardware, the simulation-enhanced models improve the success rate over the original VLA models by 57.50%, demonstrating an effective and low-cost direction for VLA enhancement.

As Industry 4.0 accelerates the integration of Cyber-Physical Systems (CPS) in manufacturing, robust anomaly detection has become critical for ensuring process safety and security. Current data-driven approaches typically employ "product-agnostic" or global models trained on the aggregate of all normal operating data. However, modern industrial facilities frequently operate under diverse product grades. While computationally simple, these global models inherently expand their decision boundaries to accommodate the variance of multiple modes, creating a "blind spot" where subtle anomalies or targeted cyber-physical attacks may be masked by the wide acceptance region of the model. In this work, we first demonstrate that the vulnerability described above is present in global-agnostic models operating across multiple product grades. We then present a Product-Aware Autoencoder as a principled mitigation that restricts the learning domain to grade-specific distributions. While this approach reduces the identified blind-spot risk, we do not claim it as the optimal mitigation among all possible alternatives. We rigorously validate this approach against a Global Agnostic baseline using the Extended Tennessee Eastman Process (TEP) benchmark. Our empirical results indicate that the Product-Aware framework performs comparably to the global baseline on standard detection metrics, while offering improved robustness to product-grade-specific operating modes. Most critically, stress tests simulating our hypothetical attack scenarios reveal that while the global model fails to detect operational deviations in 77.8% of the scenarios, the product-aware system achieves 100% detection accuracy. These findings suggest that, in flexible manufacturing environments, generalized anomaly detectors can pose non-trivial security risks, motivating a shift toward mode-aware diagnostic architectures.

We present Grokers, an architecture for building persistent, structured comprehension of typed knowledge graphs through bottom-up inductive traversal of dependency subgraphs. Unlike retrieval-augmented generation (RAG), which pays full comprehension cost at every query, Grokers pushes intelligence to write time: autonomous Groker agents analyze nodes in a typed stream graph, extract structured attributes via governed language model (LM) calls, and inductively compose that understanding upward through dependency relations, writing enriched typed attributes that serve all future queries at zero additional LM cost. We prove three formal properties: (1) the Byte-Identity Theorem, establishing that context blocks assembled from a transactionally-maintained denormalization index are byte-identical across LM turns between semantic changes, enabling KV-cache hit rates approaching 100%; (2) the Accumulation Monotonicity Theorem, establishing that the fraction of interactions resolved without LM calls is non-decreasing in the number of completed interactions under a governed wisdom library growth protocol; and (3) the Dual-Traversal Ordering Theorem, establishing that top-down generation and bottom-up comprehension are the unique correct traversal orderings for their respective tasks over a dependency DAG, and that their composition closes into a complete generation-comprehension cycle. We further present a deterministic alternative to embedding-based semantic search, with a synonym caching protocol whose LM fallback rate converges to zero for finite-vocabulary domains. A reference implementation is provided in the open-source Qbix / Safebox / Safebots stack.

Coronary artery disease (CAD) remains one of the leading causes of death globally, highlighting the need for reliable predictive systems to support early diagnosis and risk assessment. While traditional machine learning models perform well on structured clinical data, large language models (LLMs) present new possibilities to interpret medical information expressed in natural language. In this work, we develop a hybrid framework that bridges structured clinical data and natural-language representations for CAD prediction. Using a publicly available dataset of 1,190 patient records with 11 clinical attributes, structured variables are converted into interpretable feature representations and synthetic clinical narratives using LLMs. A validation pipeline performs reverse extraction of clinical variables and computes a consistency score with the original records, achieving an average fidelity of 94.61%. We then evaluate four conventional machine learning models and compare their performance with LLM-based classification under zero-shot and few-shot prompting settings. We use two LLMs here, GPT and Gemini. Experimental results show that Random Forest achieves the highest accuracy. Despite this advantage, LLM-based classification remains beneficial in real-world clinical settings. This is because LLMs operate directly on natural language patient descriptions, meaning that sensitive numerical patient data such as exact lab values, blood pressure readings, and diagnostic codes are kept private. Findings suggest that combining structured clinical data with LLM-generated narratives can enable new directions for hybrid clinical prediction systems.

Retrieval-augmented generation systems struggle with temporal reasoning and evidence fusion when answering complex questions over historical criminal case narratives. Existing approaches either retrieve independently of query semantics or fail to integrate multiple evidence sources coherently. We propose Temporal Context Augmented Retrieval Generation (TCAR-Gen), a framework that combines query-conditioned graph neural networks, temporal evidence fusion, and chain-of-trees reasoning to ground answer generation in retrieved evidence. On the Victorian Crime Diaries benchmark, TCAR-Gen achieves 0.3738 Recall@5, outperforming Vanilla RAG, Temporal RAG, GraphRAG-C, and GraphRAG-T across seven query types including multi-hop reasoning and counterfactual questions. Ablation studies reveal that the context graph, temporal penalty mechanism, and query conditioning are critical components. Cross-model evaluation across five language model (GPT-OSS 20B to TinyLlama 1.1B) demonstrates that TCAR-Gen maintains robust retrieval coverage at smaller model scales, though generation quality degrades substantially with reduced model capacity. Our work shows that explicit temporal modelling and multi-branch evidence fusion are essential for faithful, reasoning-intensive question answering over knowledge-grounded corpora.

Large language models (LLMs) are increasingly deployed across healthcare, yet existing benchmarks fail to capture model behavior under adversarial or ethically complex conditions common in clinical practice. We developed a multi-domain red teaming framework evaluating eleven contemporary LLMs across 690 clinically grounded scenarios spanning nine domains and over 150 subcategories. Scenarios incorporated adversarial transformations, and responses were assessed using a seven-dimension rubric with LLM-assisted scoring and human-in-the-loop validation. Results revealed substantial performance variance, with mean scores ranging from 0.791 to 0.984. Critically, several high-performing systems produced complete failures in individual safety-critical scenarios, demonstrating that aggregate accuracy masks clinically meaningful risk. The highest-performing systems (X-BAI, GPT-5, Claude Opus 4.1) achieved scores above 0.97 with low variance, while performance varied significantly across domains. Equity-related tasks showed 10-20% error amplification with demographic modifications, and human reviewers identified clinically relevant failures missed by automated evaluation. Our findings demonstrate that performance variance and worst-case failures provide more clinically meaningful reliability indicators than mean accuracy alone, and that hybrid evaluation approaches combining automation with clinician oversight are essential for credible safety assessment.

This paper investigates whether combining cognitively grounded linguistic features with transformer-based embeddings improves automated detection of depression in online text. Using Beck's Cognitive Theory of Depression, the study extracts cognitive distortions as measurable features, including first-person pronoun density, absolutist words, and negative emotion in Reddit posts from depression-related and control communities. Using a subset of the Kaggle Reddit Suicide and Depression Detection dataset, two classification pipelines are compared, a TF-IDF embedding with Naive Bayes as a baseline, and a hybrid model that concatenates DistilBERT sentence embeddings with Holographic Reduced Representation (HRR) vectors encoding the cognitive-linguistic features, followed by Logistic Regression. The hybrid DistilBERT HRR model achieves a macro F1 score of 0.94 versus 0.80 for the TD-IDF baseline, with 5-fold cross validation F1 improving from 0.83 to 0.92, and AUC from 0.958 to 0.981.

The rapid development of Language Diffusion Models (LDMs) challenges the dominant position of auto-regressive competitors in language processing. However, their flexible, any-order decoding strategies not only enable fast decoding speed but also potentially bring new trustworthiness challenges. To better understand the risks behind their pipelines, we introduce a comprehensive trustworthiness benchmark tailored to LDMs (TrustLDM), evaluating safety, privacy, and fairness across different LDM architectures with multiple categories of static post contexts. Our empirical results show that although LDMs generally exhibit strong trustworthiness with only the user prompts, their alignment behavior degrades noticeably when the malicious post contexts are attached to the masked responses. We further observe that longer contexts do not necessarily induce stronger effects, and both decoding order and generation length affect the evaluation outcomes. Finally, we propose TrustLDM-Auto, an automatic evaluation framework that leverages LDM decoding flexibility to systematically identify vulnerable configurations, revealing substantial trustworthiness weaknesses across all evaluated models and dimensions. Our work may potentially help the community build more trustworthy LDMs. Our code is available at https://github.com/PKU-ML/TrustLDM.

Humor generation remains difficult not only because producing fluent, novel jokes is hard, but because "funny" is audience-dependent and supervision is noisy -- preferences vary with audience, context, and culture, and annotator agreement is often low. In this paper, we describe our system for the SemEval-2026 Task-1 (MWAHAHA), which focuses on humor generation under explicit constraints. The task evaluates submitted systems via human preference judgments in 1-on-1 arena-style comparisons. We adopt a "generate-many -> select-best" strategy. First, we generate a diverse pool of candidates per instance using multi-step prompting, model ensembling, and diversity-oriented decoding. Second, we select outputs using a preference model that approximates a "reader" by learning from human comparisons rather than absolute funniness scores. To support this approach, we release 2.5K human pairwise judgments collected through the Humor Arena prototype. We further propose an interpretable pipeline that converts labeled comparisons into a preference model. Across three preference datasets, our models consistently outperform baselines and show stronger cross-domain transfer. Finally, we apply the learned preference model to rank candidates for the MWAHAHA setting and release intermediate artifacts (candidate pools and rankings) to facilitate follow-up work. Our system ranked 1st in the English and Chinese subtasks of MWAHAHA and 2nd in the Spanish subtask.

Speculative Decoding (SD) accelerates Large Language Model (LLM) inference by employing a lightweight draft model to propose candidate tokens, which are verified in parallel by the target model, without compromising generation quality. While Retrieval-based Speculative Decoding (RSD) is favored for its plug-and-play versatility, its potential is impeded by rigid lexical dependencies, rendering both retrieval and verification brittle to surface-level variations. To address this, we propose SENSE (Semantic Embedding Navigation with Soft-gated Evaluation). By anchoring retrieval on the hidden states of the target model, SENSE establishes robust semantic alignment, which empowers the Soft-gated Evaluation module to validate semantic equivalence rather than surface forms. To ensure rigorous benchmarking, we deconstruct existing methods into atomic primitives within a unified framework, facilitating granular, component-level comparison. Extensive experiments across diverse domains demonstrate that SENSE outperforms multiple baselines on the LLaMA and Qwen families, attaining up to 4.09 mean acceptance length and 3.26x speedup, while preserving generation quality. Our code will be released upon publication.

Large Language Model (LLM) based Chinese Grammatical Error Correction (CGEC) systems face two critical challenges: general-purpose models lack specialized linguistic priors for subtle grammatical distinctions, and Supervised Fine-Tuning (SFT) with Maximum Likelihood Estimation fails to optimize for precision-focused metrics, leading to systematic over-correction. We propose CSRP, a three-stage framework that progressively builds correction capability through Continual Pre-training (CPT) on 5.9M balanced samples to internalize domain knowledge, Chain-of-Thought SFT with explicit error reasoning for diagnostic transparency, and Group Relative Policy Optimization with a novel Efficiency-Aware Reward that explicitly penalizes unnecessary edits. On the NACGEC benchmark, CSRP achieves state-of-the-art performance with 50.99 $F_{0.5}$ and 57.17 precision, substantially outperforming previous best results while effectively mitigating the over-correction bias inherent in MLE-trained models. Our method also advances CSCD spelling correction to 59.61 F1, surpassing GPT-4 by 5.20 points. Comprehensive ablation studies demonstrate that the RL alignment stage contributes a 8\% relative gain over the SFT baseline, and that this gain is orthogonal to the contribution of large-scale CPT, validating that explicit optimization for edit efficiency is essential for high-quality grammatical error correction. Our code is available at https://github.com/TW-NLP/ChineseErrorCorrector.

Training language model agents for multi-agent strategic interaction presents a core difficulty: the quality of any action may depend on future events that never materialize, on moves that violate game rules, or on decisions made by other players. Standard reinforcement learning assumes that rewards can be assigned at each step, but this assumption fails in settings where outcomes are entangled across time and agents. We introduce delayed per-step reward attribution with eligibility gating, an episode lifecycle and postprocessing pipeline that computes rewards only at episode end, propagates them back to originating steps according to task-specific semantics, and excludes steps that lack valid dependent information from training. Together with asynchronous rollout generation via vLLM's continuous batching, curriculum-based opponent sampling, and multi-level stratified batch construction, this approach enables stable, sample-efficient RL training in multi-agent environments. We evaluate on the MindGames Arena benchmark at NeurIPS 2025, where a single 8-billion-parameter open-source model trained with our method matched or surpassed substantially larger proprietary systems, including GPT-5, in head-to-head play and took first place in both the Open (unrestricted) and Efficient (<=8B parameters) tracks.