Papers

Real-time decision-making for enhanced geothermal systems (EGS) is challenging because long-term production periods involve high-dimensional control spaces and a large number of time-consuming high-fidelity hydrothermal simulations. Reinforcement learning provides a natural framework for state-dependent sequential control, but direct policy training with numerical simulators is computationally expensive. To address this issue, we propose a diffusion-surrogate guided reinforcement learning framework for long-horizon EGS well-control optimization. The reservoir temperature and pressure fields are used as system states, while injection rates are selected as control actions. A learned surrogate environment is constructed using conditional diffusion models to predict the evolution of reservoir temperature and pressure fields and a separate reward model to estimate the corresponding economic return. The surrogate environment is then integrated with Proximal Policy Optimization (PPO) for efficient policy training. Experiments on a fractured EGS benchmark show that the diffusion surrogate can accurately reproduce reservoir-state evolution over multiple control stages. The resulting surrogate-assisted PPO policy achieves competitive well-control performance compared with direct simulator-based PPO and existing optimization methods, while substantially reducing the dependence on expensive high-fidelity simulations. These results demonstrate the potential of diffusion-based surrogate environments for efficient reinforcement learning in geothermal well-control optimization.

Technical operations teams resolve large volumes of incidents by synthesizing fragmented evidence from ticket text, historical cases, system logs, and technical documentation. Existing automation often relies on monolithic generation without explicit evidence modeling or provenance, making outputs difficult to verify when critical signals are sparse across sources. We propose ASTRA, an agentic system for ticket resolution in which a central orchestrator coordinates three specialist information-gathering agents and drives a judge-orchestrator refinement loop to produce evidence-backed troubleshooting reports. TicketSimilarityAgent retrieves relevant historical precedents through dense retrieval and LLM reranking; LogAgent distills hundreds of thousands of log lines into structured, quote-grounded findings using deterministic filtering and constrained LLM analysis; and DomainKnowledgeAgent retrieves relevant technical knowledge via the Model Context Protocol (MCP). Their outputs are transformed into a claim-evidence representation linking each claim to a verbatim source passage, assigning a support level, and preventing cross-attribution. A JudgeAgent scores the report on five criteria, while the OrchestratorAgent converts low scores into targeted follow-up queries for bounded iterative refinement. Evaluated on 987 real-world telecom fault tickets across seven product lines, ASTRA achieves a mean quality score of 4.13/5.0, with 59.9% of reports identifying the fault area at the component-family level or better. Relevance and Clarity scores are 4.88 and 4.94, respectively, while fabricated technical details remain below 3% of error cases. Stratification by fault type reveals that hardware faults remain substantially harder than software or configuration faults (Cohen's d=0.80), pointing to a fundamental limitation of text-based evidence channels for hardware fault diagnosis.

AlphaFold 3 (AF3) enables structure prediction of biomolecular complexes through co-folding multiple interacting molecules, making it increasingly useful for de novo protein design and for large-scale studies of protein-protein, protein-peptide, and other biomolecular interactions. However, systematic co-folding experiments can produce large volumes of output data, particularly when multiple random seeds and samples are generated for each input complex. We introduce FoldKit, a Python package for efficient storage and analysis of large-scale AF3 co-folding results. FoldKit converts raw AF3 outputs into a compact, structured representation while preserving the metadata needed for downstream analysis. The FoldKit Python library provides convenient programmatic access to global, single chain, and interface confidence metrics such as pLDDT, pTM, ipTM, ipAE, and ipSAE, as well as an ensemble-level interface for accessing and aggregating these metrics for a single input across multiple seeds and samples. We benchmark FoldKit on three types of AF3 co-folding datasets: (i) a protein design campaign with 2 chains per input, (ii) a TCR-pMHC dataset with 4 chains per input, and (iii) a pooled-AF3 protein-protein interaction dataset with up to 22 chains per input. We find that FoldKit reduces storage requirements by approximately 5-15-fold compared to native AF3 outputs, depending on dataset composition, while maintaining direct programmatic access to individual predictions, ensembles, and confidence metrics. By reducing storage requirements and facilitating programmatic access to relevant outputs, FoldKit facilitates large-scale computational studies of biomolecular interactions. FoldKit is available from PyPI and can be installed using pip.

Joint-embedding predictive architectures (JEPAs) have primarily been developed for self-supervised representation learning. Denoising JEPA (D-JEPA) recently demonstrated strong generative capabilities on natural images, yet the applicability to 3D medical imaging remains unexplored. Building on the D-JEPA framework, we present Med-D-JEPA, a systematic adaptation and evaluation of joint-embedding predictive generation for 3D brain MRI. Med-D-JEPA operates on continuous latent tokens produced by a 3D KL-regularized adversarial variational autoencoder, and combines masked context prediction, representation-level alignment, per-token diffusion, and iterative next-set-of-token sampling. We evaluate unconditional and class-conditional generation quality on BraTS2019 and OASIS-1 datasets; downstream classification utility; and preliminary whole-tumor segmentation on BraTS2020. Across different generation settings, Med-D-JEPA achieves superior or competitive performance compared to several strong baselines on fidelity and diversity metrics. Compared to training with real samples, Med-D-JEPA-based synthetic pretraining improves classification AUC from 0.63 to 0.85 on BraTS2019 and from 0.78 to 0.87 on OASIS-1. In the segmentation study, pretraining on Med-D-JEPA samples improves Dice from 0.74 to 0.80 and reduces HD95 from 13.40 to 9.56 mm. These findings establish joint-embedding predictive generation as a promising direction for 3D medical image synthesis and encourage further research in this direction.

Multimodal Large Language Models (MLLMs) have recently made strong progress in visual–linguistic understanding. However, their performance on text-centric video reasoning remains highly sensitive to input quality. Real-world user-provided videos often contain motion blur, compression artifacts, noise, and low-resolution text, which impair reliable text reading and downstream reasoning. Whether MLLMs can robustly read and reason about real-world scene text under diverse quality conditions remains a fundamental open question. We introduce ClearText-Video (CTVid), a large-scale, scene-text-aware benchmark for studying text-centric video understanding under controlled quality variation. CTVid contains 4,639 real-world text-rich egocentric videos, 550K+ frames, 1.6M human-verified scene-text annotations, and 220K+ spatial/temporal question–answer pairs in Chinese and English. For each high-quality video, CTVid provides content-matched Degraded-Quality and Restored-Quality variants, supporting two task families: Text-Centric Video Restoration and Multi-Quality VideoQA. We evaluate 18 representative restoration methods and 16 state-of-the-art MLLMs on CTVid. The results show that visual enhancement does not guarantee textual fidelity or downstream reasoning gains: blur is more damaging than low resolution, restored videos can alter the textual evidence used by MLLMs, and OCR-only pipelines remain far below direct multimodal reasoning. CTVid exposes the gap between video restoration and text-grounded understanding, providing a rigorous foundation for restoration-aware, quality-robust text-centric video systems.

Journaling is a common practice for emotional expression, reflection, and processing. However, as entries accumulate, it can become difficult to interpret and compare their affective content, especially since traditional text-based analyses and visualizations often struggle to convey affective nuance. We introduce Data Characters, a visualization approach that represents affective content in journaling through human-like characters. Using a customizable Data Character as a design probe, we investigate the potential of character-based representations for conveying affective experiences and explore what visual encodings emerge through customization. Preliminary walkthroughs with two participants demonstrate the intuitiveness and feasibility of the approach. This work contributes an exploratory approach to studying how affective experiences can be visually represented and encoded through anthropomorphic forms.

Autonomous vehicles (AVs) rely on accurate camera-LiDAR calibration for multimodal sensor fusion. In practice, calibration can drift due to vibration, temperature variation, or minor sensor displacement, motivating online calibration algorithms that detect and correct misalignment at runtime while allowing the vehicle to continue operating without a factory visit. Existing AV attacks largely assume correct calibration. We instead identify online sensor calibration as a new attack plane. A corrupted calibration update can persist across subsequent fusion operations, causing system-wide errors that propagate from perception to planning and control. We present Adversarial Calibration Attack (ACA), the first physical attack against camera-LiDAR online calibration. Using a single adversarial poster, ACA first spoofs the miscalibration detector to trigger the calibration process and then steers the calibration estimator toward an incorrect transformation. A unified optimization jointly designs the poster's geometry and texture for both objectives. We evaluate ACA across benchmark datasets, simulation, and physical experiments. On benchmark datasets such as KITTI and nuScenes, ACA induces up to 33.9 degrees mean rotational calibration error, thereby severely degrading object detection. In the CARLA simulator, the attack causes a collision when the corrupted calibration is accepted in vulnerable scenarios crafted by the attacker. On a real Husky robot, a printed adversarial poster successfully reproduces the calibration error. These results demonstrate that online calibration is a practical and safety-critical attack surface for AVs.

Although recent deepfake detectors achieve high overall accuracy, their errors remain unevenly distributed across demographic subgroups, with real faces from certain groups more often misclassified as fake. Existing fairness-aware detectors typically regularise the entire feature representation, without identifying or controlling the specific components that drive unfair predictions. Such coarse intervention can over-suppress useful forgery cues while leaving demographic structure in component-specific subspaces. To address this, we identify two subgroup-sensitive components: multi-scale spatial features, which encode local facial and forgery patterns, and fine-tuning-induced residual features, which adapt the backbone to the unfair training distribution. We propose FairReL, a fairness-aware representation-learning framework that targets both components with dedicated demographic supervision. FairReL uses an SVD-decomposed foundation-model backbone to isolate the fine-tuning-induced residual representation, and introduces two complementary losses. Group-Conditional Wavelet Decorrelation (GCWD) suppresses subgroup-imbalanced structure across spatial wavelet sub-bands, while Subspace-Localised Mean Alignment (SLMA) aligns subgroup means within each real/fake class in the residual representation. Experiments on FF++, Celeb-DF, DFD and DFDC show that, against the state-of-the-art fairness-aware detector, FairReL improves unseen-dataset AUC by 3.9% while reducing subgroup FPR disparity by 10.2%. Code is available at https://github.com/xiaoman89/FairReL .

Multilingual sentence embeddings are increasingly used to estimate semantic similarity across languages, yet their sensitivity to fine-grained translation errors remains insufficiently understood. This study investigates whether general-purpose multilingual embedding models can distinguish correct English-Greek translations from minimally modified erroneous alternatives. A contrastive dataset was developed from FLORES+ sentence-aligned reference translations and reviewed by two translation experts. It contains 1,850 examples across ten core and five exploratory error categories, covering factual, lexical-semantic, grammatical, relational, referential, and discourse-level phenomena. Five multilingual sentence-embedding models (BGE-M3, Multilingual E5, Multilingual MPNet, LaBSE, and Jina Embeddings v3) were evaluated using cosine similarity between each English source sentence and its correct and erroneous Greek translations. A reference-free COMETKiwi model was also evaluated as an MT quality-estimation baseline. Performance was assessed through contrastive accuracy and score margins for category-specific sensitivity. BGE-M3 achieved the highest accuracy among embedding models at 89.30 percent, while COMETKiwi achieved 94.49 percent. Embedding models detected explicit factual and lexical changes more reliably than tense-and-aspect and pronoun-coreference errors. COMETKiwi improved performance on several difficult categories, including tense and aspect, pronoun and coreference, and semantic-role errors, but showed lower sensitivity to date-and-time errors and underperformed the embedding models on numbers. The results show complementary error-sensitivity profiles: multilingual sentence embeddings provide useful semantic adequacy signals but are better suited as components of broader translation-evaluation frameworks than as standalone metrics.

Perceptual losses are widely used in medical image synthesis because they encourage agreement in high-level structure beyond voxel-wise intensity similarity. In practice, most perceptual losses are still computed with natural-image backbones such as VGG16 or ResNet50, even when the target domain is magnetic resonance imaging (MRI). This mismatch may weaken supervision for anatomy, contrast enhancement, and acquisition variability. We test whether medical foundation model features provide a more suitable perceptual loss for brain MRI contrast dose simulation. The study has two stages. First, we compare RadImageNet, SegVol, and BrainIAC with ImageNet-pretrained VGG16 and ResNet50 as frozen feature extractors on four public medical imaging benchmarks: thyroid ultrasound, breast ultrasound, anterior cruciate ligament knee MRI, and meniscus knee MRI. RadImageNet achieves the lowest mean rank across the Stage I representation suite and is selected as $\phi^\star$. Second, we replace only the VGG16 feature extractor in an existing iterative brain MRI dose simulation framework with $\phi^\star$. The generator, reconstruction loss, adversarial loss, auxiliary losses, optimization schedule, and loss weights are kept unchanged. Standard metrics change modestly, with PSNR increasing from 41.63 to 41.74, SSIM from 0.9739 to 0.9754, RMSE decreasing from 0.1384 to 0.1369, and residual-uptake CNR from 0.0085 to 0.0082. The visual results show the main effect: RadImageNet reduces residual enhancement in marked structures, follows a more faithful dose-reduction trajectory, and remains close to the acquired 10% low-dose target. These results support domain-aligned radiology features as a practical perceptual feature space for MRI dose simulation, while leaving clinical equivalence and larger-cohort validation as future work.

Large reasoning models achieve strong performance on complex tasks by generating extended chain-of-thought (CoT) traces via reinforcement learning with verifiable rewards (RLVR). While current RLVR methods have achieved strong results with correctness-based reward signals, they provide limited guidance on the quality of the reasoning process itself, leaving the internal reasoning structure largely unoptimized. Through empirical analysis across multiple model families, we identify a consistent pattern: correct reasoning trac es exhibit more frequent and larger token-level entropy drops within the thinking phase than incorrect ones. We propose ERR+, a two-phase RLVR framework grounded in this observation. The first phase trains with the Entropy Relief Reward (ERR), a bonus proportional to cumulative token-level entropy drops in the thinking phase, log-normalized by response length. Unlike prior methods that suppress entropy, ERR rewards the resolution of uncertainty while leaving exploratory high-entropy states unconstrained. The second phase introduces the Robust Relative Efficiency Reward, which scores each response's length against co-generated peers via a $\tanh$-transformed within-group $z$-score. We provide a formal analysis showing that joint optimization of the two objectives induces gradient conflict in early training, motivating the sequential design . Experiments on five datasets demonstrate consistent improvements in both accuracy and response conciseness across model backbones. Our code is available at https://github.com/XrkArul/err_response

We study the Edge-Constrained Graph Partitioning Problem (ECGP), which asks whether the vertices of a graph can be partitioned into r parts, each inducing at least gamma edges. We also consider a balanced variant (BECGP), requiring equal-sized parts, and signed variants, where the utility of a part is the difference between its numbers of positive and negative edges. We show that ECGP and BECGP remain NP-hard for fixed gamma, while BECGP is also NP-hard for fixed r. For the natural parameterization r+gamma, both problems admit polynomial kernels. We obtain FPT algorithms for ECGP and BECGP parameterized by maximum leaf number, vertex deletion distance to a clique, cluster vertex deletion number plus gamma, and vertex integrity. Furthermore, ECGP is FPT parameterized by vertex deletion distance to stars plus gamma and vertex deletion distance to paths plus gamma. On the negative side, ECGP and BECGP are W[1]-hard when parameterized by r together with several structural parameters. In particular, hardness holds for feedback edge set, vertex deletion distance to stars or paths, and modular width even when the corresponding parameter is zero. The problems are also W[1]-hard parameterized by cluster vertex deletion number plus r, and by clique-width even when gamma=3. For signed graphs, both variants are NP-hard even when r+gamma=3 and the input is a disjoint union of two cliques. Finally, the balanced signed variant is W[1]-hard parameterized by treedepth plus r, even when gamma=0.

Traditional Retrieval-Augmented Generation (RAG) systems treat documents in isolation, failing to capture hierarchical relationships between concepts in complex scientific corpora. This limitation compromises answer quality in specialized domains such as climatology, where conceptual dependencies frequently traverse multiple articles. We propose a hybrid architecture that integrates vector search with GraphRAG, Leiden community detection, and cross-encoder re-ranking, achieving gains of 160% in contextual relevance and 177% in contextual recall compared to classical RAG. These results demonstrate that unifying local and global retrieval significantly outperforms text-span isolation, paving the way for more effective question-answering systems over dispersed scientific literature.

Recent advances in visual question answering (VQA) and multimodal large language models (MLLMs) have enabled natural-language reasoning over traffic scenes. However, existing benchmarks are largely built from ego-vehicle views or 2D roadside videos, limiting their ability to evaluate 3D-grounded reasoning over real-world distances, trajectories, infrastructure topology, and safety-critical interactions. We introduce Inter-3D VQA, a large-scale roadside multimodal benchmark for 3D spatiotemporally grounded VQA at intersections. Built from synchronized point clouds and multi-view images, Inter-3D VQA contains 407K QA pairs covering lane-level positions, object relationships, motion patterns, and near-miss-oriented interaction reasoning. We further propose Inter-Geo, an MLLM baseline that integrates object- and scene-level aligned LiDAR representations, and Inter-Metrics, a unified evaluation framework for textual consistency, numerical accuracy, and semantic correctness. Experiments show that Inter-Geo outperforms image-based VLMs, especially on grounded spatial and temporal reasoning tasks. Our benchmark and codes are available at https://github.com/ASU-Suo-Lab/Inter-3D-VQA .

This article introduces peer $k$-oversight, a property of sequential collective decision mechanisms requiring at least $k$ agents to be responsible for every harmful outcome. It is shown that whenever $k$-oversight can be achieved by redistributing control over the decisions in a mechanism, it can be achieved using just $k$ agents. A polynomial-time algorithm is also presented that determines whether such a redistribution exists and, when it does, constructs one. These results establish peer oversight as a tractable design principle for multiagent decision-making systems.

Indoor 3D Scene Graphs (3DSGs) represent environments as multi-layer hierarchies that connect observed geometric primitives (e.g., planes) to higher-level metric-semantic concepts (e.g., rooms, floors, buildings), enabling incremental spatial reasoning for robotic perception and SLAM. However, classical high-level concept generation approaches rely on hand-crafted rules for specific concept classes, while learning-based methods require separate models for graph structure and spatial node features (e.g., centroids), which limits scalability to novel classes and more complex hierarchies. We propose a unified autoregressive diffusion-based graph generative model that jointly learns structure and features, constructing complete 3DSGs bottom-up from observed vertical planes across arbitrary hierarchy depths. Our method consistently surpasses all learning-based and random baselines across 3DSG datasets spanning synthetic scenes, real architectural floor plans, and robotic sensor data, with varying layout complexity and hierarchy depth, and surpasses a one-shot model with oracle access to the target graph size on the largest hierarchy and on real single-floor data. Finally, we propose an adaptation of the Fused Gromov–Wasserstein distance for principled graph-level evaluation of generated 3DSGs against ground truth.

Ensuring the security of complex systems involves the strategic allocation of defensive resources to prevent various types of attacks from succeeding. A defender often has multiple types of defensive assets at its disposal, where it must decide how to optimally deploy their heterogeneous capabilities across different attack types. In this paper, we formulate a multi-resource allocation problem in the form of a General Lotto game where a defender possesses various types of resources. A feature that we introduce is that their individual effectiveness against different types of attacks is characterized by a network weight matrix. In our analysis, we derive upper and lower bounds on the performance of the defender, and provide numerical evidence suggesting that they are tight. For the case of two attack types, we analytically prove that the bounds coincide, establishing an exact equilibrium characterization. We then numerically compare our proposed networked multi-resource architecture to an independent-defense benchmark from the existing literature. These results highlight fundamental and tractable structures underlying multi-attack-type defense problems.

Latest JPEG restoration systems achieve strong quality with large models, yet often remain too slow and expensive for efficient on-device deployment. We present a 65M-parameter generative restorer that attains the lowest LPIPS at QF 10 and 20 on LIVE-1, Urban100, and DIV2K-val while sustaining 8.05 images/s at $1024\times1024$ on a single RTX 3090, roughly $4.9\times$ the reported throughput of one-step SODiff at one-twentieth of its parameters. Trained from scratch, the model replaces the learned VAE encoder-decoder with an exactly invertible two-level Haar transform, predicts a clean wavelet-domain residual through a rank-enhanced linear-attention DiT that estimates compression severity internally, and is optimized with an improved MeanFlow objective that enables inference in one or two network evaluations without distillation. Large pretrained priors remain stronger under severe compression (QF 5), whereas our model prioritizes throughput for deployment-constrained restoration.

With the increasing importance of teamwork skills for modern workplaces, development of teamwork training programs has received substantial attention. Game-based teamwork training is one promising approach that is engaging, cost-effective, and well-suited to increasingly decentralized workplaces. However, design of effective game-based teamwork training requires understanding how a game elicits specific desired teamwork behaviors. Significant progress has been made in characterizing these relationships. However, despite its critical importance, little work has examined how a game's design influences team adaptability behaviors. This paper presents a preliminary framework for analyzing adaptability in cooperative games by conceptualizing adaptive stimuli as retrieval or disruption of players' ludic knowledge. We illustrate this framework through a qualitative case study that applies interaction analysis methods to gameplay videos of a Overcooked!, a cooperative cooking game. We examined instances where game events led to players altering their behavior and connected the game's design features that resulted in each event with three adaptive stimulus cue categories. Although exploratory and limited to a small case study of a single game, the proposed framework is grounded in established theories across teamwork research and game studies, and it offers an initial vocabulary for describing how cooperative games can be designed to create demands for team adaptation. With this continued development, the framework may provide an analytic tool to help inform the design and evaluation of purpose-built game-based teamwork training environments.

Urban construction governance requires early decisions that connect workplace safety, permitting requirements, and community impact, yet the relevant evidence is often scattered across separate municipal and regulatory data sources. This paper presents PermitGPT, a unified generative artificial intelligence framework for converting unstructured construction permit descriptions into structured decision-support outputs across three domains: safety hazard identification, permit requirement specification, and community impact assessment. To address data fragmentation, we spatially and temporally align records from the New York City Department of Buildings, Occupational Safety and Health Administration, and NYC 311 service requests, producing 90,000 structured prompt-response pairs derived through rule-based alignment and domain-informed spot checking. We fine-tune three open-weight language models using parameter-efficient adaptation and evaluate them on 2,833 held-out test cases. The results show complementary model behavior: Gemma-3-1B provides the most efficient inference at 3.07 samples per second with low memory usage, Llama-3.2-3B gives the highest lexical overlap for regulatory-style outputs with a BLEU score of 0.0091, and 4-bit Mistral-7B-Instruct-v0.3 achieves the strongest semantic alignment with a BERTScore-F1 of 0.7747. Because the task involves open-ended structured generation, low BLEU values are interpreted alongside semantic metrics and qualitative output structure rather than as standalone indicators of utility. Overall, PermitGPT provides an initial step toward AI-assisted construction governance while identifying directions for stronger task-level evaluation and real-world validation.

We introduce a fully discrete numerical scheme for semilinear SPDEs with additive space-time white noise that overcomes the previous order barrier for the spatial convergence rate. The scheme, which we refer to as the doubly accelerated exponential Euler scheme, achieves a strong convergence rate of $M^{-1+\epsilon}$ in time and $N^{-3/2+\epsilon}$ in space for any $\epsilon>0$, where $M^{-1}$ and $N^{-1}$ are the temporal, respectively the spatial, meshsizes. This substantially improves the standard spatial error bounds of order $N^{-1/2}$ in the literature. Numerical simulations support the findings.

The remarkable performance of multimodal large language models (MLLMs) comes at the cost of substantial computational overhead, posing significant challenges to real-time deployment and cost effectiveness. Existing model routing approaches either decide from coarse request-level features alone or spend one or several extra language model passes to inspect the generated response, leaving the token-level uncertainty signals that emerge during generation unused. To address these limitations, we propose Pro-Router, a token-aware progressive model routing method with adaptive edge-cloud collaboration for efficient multimodal LLM inference. Pro-Router employs a two-stage progressive decision mechanism. First, a lightweight prompt pre-scorer module performs rapid pre-screening before token generation begins, guiding apparently simple requests to small models. Second, a token-aware verifier reads the sampling probability distribution of each token the small model generates, estimating the model's confidence in its own output to determine, per request, whether the answer ships or escalates to the cloud-based high-precision model. Furthermore, we design an adaptive edge-cloud serving pipeline that sizes every dispatch to each device's measured service rate, so both the edge and the cloud tiers stay fully utilized without manual parameter tuning and are not impacted by the network latency. Extensive experiments on multiple multimodal benchmark datasets and models demonstrate the effectiveness of Pro-Router. Compared to other methods, it achieves the highest routing accuracy and improves routing speed by more than 10x. Its serving pipeline also reaches more than 75% higher end-to-end throughput than the existing model routing pipeline. Our code is available at https://github.com/xinyuangui2/pro-router.

Large language models (LLMs) are increasingly used as graders, verifiers, and process auditors, but most mathematical evaluations still emphasize final-answer accuracy. This can obscure whether a model can verify a non-canonical but valid solution trace. We introduce a controlled linear-equation benchmark for evaluating LLMs in the evaluator role. Each instance asks the model to judge final-answer correctness, step-level trace correctness, and the first incorrect step. Our evaluation of state-of-the-art open LLMs reveals a significant robustness gap: models that accurately evaluate canonical solutions often fail when presented with perturbed but logically equivalent variants. Across GPT-OSS 20B, Qwen3-14B, and Phi-4-Reasoning, base models perform well on canonical traces but degrade substantially on perturbed traces, especially for error localization. On valid perturbed traces, base-model false-rejection rates reach 75.6-85.3%, showing strong sensitivity to canonical solution form. Supervised fine-tuning, distillation, and test-time compute improve robustness in some settings, but gains are model dependent and can trade off against canonical performance. The results show that reliable process-level verification remains challenging, and evaluator robustness should be measured separately from solver accuracy, even in a simple algebraic domain with exact ground truth.

Planetary surface exploration missions rely increasingly on autonomous robotic platforms capable of interpreting complex terrain to ensure safe navigation, enable targeted science, and improve operational efficiency, as demonstrated across past Mars missions from Viking through Perseverance. Among the key perception capabilities, landform classification provides contextual information for landing site selection and scientific analysis, while boulder segmentation supports hazard assessment and path planning. This paper presents MANTLE, a multi-task adaptive network for terrain and landform extraction. The model uses a shared DINOv2 backbone for high-level feature extraction with task-specific heads: a classification head for large-scale landform classification, and a segmentation head for pixel-wise boulder localization, each trained on curated datasets built respectively from HiRISE orbital imagery and MSL surface-level imagery. The classification head achieved a test accuracy of 92.56% across seven Martian terrain classes, while the segmentation head achieved a validation IoU of 0.753 and showed strong cross-sol generalization on a held-out test set from previously unseen rover traverses. A key advantage of MANTLE is its modular, extensible design, formalized here as the Modular Uplink Principle: only a shared, frozen backbone needs to remain onboard, while subsequent perception capabilities are trained on Earth as lightweight task-specific heads and uplinked without retraining the full model. This work demonstrates two such high-impact capabilities, terrain classification and boulder segmentation, as an initial realization of a framework built to support many more over a mission's lifetime. With this foundation, future explorers need not arrive on Mars fully formed, but can continue to learn, adapt, and grow more capable with every uplink.

Automated drift controllers commonly track a prescribed drift equilibrium, sideslip reference, or trajectory. These formulations establish how to execute drift, whereas the continuous transition from grip driving to drift near the handling limit remains unresolved. This paper defines drift emergence in a repetitive lap time minimization task, where neither the controller objective nor the reward contains an explicit drift reference. A boundary exploration learning model predictive controller (BE-LMPC) constructs an empirical safe set and a locally shifted terminal cost from completed laps. By iteratively improving spatial speed allocation under a fixed global speed bound, the controller progressively explores larger sideslip and yaw rate envelopes while preserving recoverability. As lap performance improves, sustained sideslip and pronounced yaw motion emerge while the rear axle approaches saturation. Analysis shows that, when external conditions vary smoothly, the transition from tire adhesion to sliding does not itself cause abrupt changes in tire force or vehicle state. The combined-slip Fiala model satisfies this continuity condition at the transition. At a tire road friction coefficient of 0.6, lap time decreases from 49.95 s on Lap~3 to 25.50 s on Lap~12, with drift first emerging on Lap~11. Lap~12 reaches 16.5$^\circ$ sideslip and 0.894 rear axle utilization. In contrast, no drift is detected for friction coefficients from 0.8 to 1.2; at 1.2, a similar peak speed is achieved with only 0.483 rear axle utilization. These results characterize drift as a conditional continuation of limit handling that emerges when increasing performance demand approaches the available tire capacity, rather than as a separately prescribed motion mode.

The VCG family and the AGV mechanism are two classical approaches to efficient implementation in the static social choice problem. In 2024, Csóka et al. showed that AGV has critical weaknesses. In contrast, the transferable-utility Guaranteed Utility Mechanism (TU-GUM) retains all the standard desirable properties of AGV while adding further ones, including collusion-proofness, because it implements efficiency in Guaranteed Utility Equilibrium. TU-GUM also applies to a more general dynamic setting with multiple extensions. Moreover, TU-GUM is a special case of an even more general and robust mechanism that combines contingent first-price tendering with the coordinated execution of dynamic stochastic multi-agent projects through a surprisingly simple rule. This paper summarizes and connects existing results from a different perspective, with some minor new observations.

Video world models increasingly serve as data engines, action planners, and simulators for embodied AI, but conventional embodied world model (EWM) benchmarks lack a unified 3D-grounded protocol for establishing whether generated rollouts preserve the underlying 3D scene state or translate into executable actions. We introduce RoboPhys-3D, a 3D-grounded EWM benchmark built on RoboTwin 2.0, covering 50 manipulation tasks across four regimes, with 5,000 episodes and 25,000 multi-view ground-truth videos. A defining feature of RoboPhys-3D is that generated and ground-truth videos are processed through the same 3D reconstruction pipeline, enabling reconstruction-induced error to be distinguished from generation-induced error. The RoboPhys-3D benchmark organizes 50 complementary metrics into 18 sub-dimensions across four levels: pixel-level fidelity, 3D geometry consistency, state-level understanding, and task-level completeness. We further introduce Average Full Score, a hierarchical score averaging all 50 metrics for comprehensive evaluation, and RoboPhyscore, a compact task-aligned score averaging the metrics most strongly correlated with task success. Among the four representative video world models, Cosmos 3 achieves the highest RoboPhyscore (0.6330, 92.7% of ground truth), while state- and execution-grounded metrics reveal substantial failures that perceptual and vision-language model-based judgments fail to capture. RoboPhyscore further exhibits strong agreement with human evaluation (Pearson r = 0.9761 and Spearman h̊o} = 0.8962), demonstrating the importance of grounded, execution-aware evaluation for EWM capability.

Deep learning segmentation models are often evaluated using geometric metrics such as Dice, HD95, and ASD, yet it remains unclear to what extent improvements in these metrics translate into clinically meaningful changes in downstream decision-making. The metric-to-decision gap is examined using radiological Peritoneal Cancer Index (rPCI) region segmentation on contrast-enhanced CT, where a consensus definition provides anatomically grounded 3D regions and the clinically used PCI 20 threshold enables decision-level evaluation. Inter-observer variability is quantified across four experts on ten abdominal CT scans, and a published nnU-Net based rPCI segmentation model is benchmarked against this human reference using Dice, HD95, and ASD across all 13 regions. To relate geometric differences to clinical impact, a probabilistic peritoneal metastasis simulation is implemented on majority-vote rPCI maps, propagating region-boundary variability into variability of derived (r)PCI scores and classification at the PCI 20 cutoff. Observers showed high agreement (mean Dice $0.87$), while the model matched human performance in most regions but deviated more in regions 4, 8, and the small-bowel regions (9-12). Across simulations, score differences were typically small (mean $\Delta$rPCI $\approx 0.3$-$0.6$) for both observers and the model, and decision flips occurred predominantly when the reference score was near 20. These results suggest that rPCI-derived scoring is generally robust to typical segmentation variability, while highlighting borderline cases as the main setting where expert review remains essential.

Ultrasound-guided interventions can require localization of an untracked 2D frame within a 3D anatomical reference. Rigid slice-to-volume registration (SVR) estimates this six-degree-of-freedom pose but remains challenging because of limited anatomical context, acoustic artifacts, and view-dependent appearance. Existing methods often use dense cross-attention, whose cost scales with the product of slice and volume token counts, or direct pose regression that lacks explicit correspondence constraints. We introduce SCoPE-Reg, combining dual-input state-space slice-volume interaction, dense 3D coordinate prediction, and parameter-free weighted Kabsch estimation. On SVR tasks from CAMUS and $\mu$-RegPro, SCoPE-Reg yields mean target registration errors of $0.73$ mm and $2.27$ mm against $1.24$ mm and $2.63$ mm for the best-performing baseline, reduces peak error on CAMUS by 56% ($12.5\to5.5$ mm), and registers $100%$ and $80%$ of frames within $3$mm. On CAMUS at $128^2$ it retains the lowest error at increasing pose-perturbation magnitude. It holds $6.49$ M parameters independent of resolution, sustaining $51$ FPS at $512^2$. By coupling correspondence-based accuracy with bounded worst-case error and resolution-independent cost, SCoPE-Reg becomes viable at native acquisition resolution during intervention, whereas previous methods had to compromise on accuracy, reliability or frame rate. Supplementary code: https://github.com/TIO-IKIM/SCoPE-Reg.

Objective: Deep learning accelerates brain MRI four- to tenfold, but models can erase lesions or synthesize false tissue - failures pixel-averaged metrics like PSNR and SSIM miss. We review whether current evaluation practices detect this blind spot. Methods: Following PRISMA 2020, we searched seven databases without date limits, including 263 studies (1995-2026), appraised them using QUADAS-2 and matched instruments, and synthesized narratively. Categories were derived from titles, abstracts, and controlled vocabulary; reported prevalence figures represent floors. Duplicate screening achieved high agreement (Fleiss kappa = 0.877), as did appraisal (0.788; 0.390 where observable). Extraction is unaudited. Results: Only 18 of 263 studies (6.8%) recorded both a fidelity metric and reader assessment on identical data, leaving the central surrogate unmeasured. Reader studies mostly measured inter-reader agreement, which was weak: fastMRI 2020 concordance reached 0.457 and 0.386 (Kendall W), improving only where SSIM diverged. Erasing a 100 mm3 lacunar infarct shifts global PSNR by 0.03 dB under the stated error model. As the corpus grew fivefold, reader assessments dropped from 32% to 18%, recovering to 21%. Generative models - most associated with hallucination (39%) - were among the least reader-evaluated (11.3%), while self-supervised models reached 47% with zero reader evaluation. Only 5% released code and ran reader studies; none evaluated a model observer; no named dataset covered acute stroke or hemorrhage. Conclusions: On these floors, current evaluation practices cannot certify diagnostic safety. We derive five requirements safety-oriented evaluations must meet.