Papers

Model compression techniques such as pruning and quantization facilitate the efficient deployment and acceleration of Large Language Models (LLMs). However, recent studies show that weight sparsification methods, such as SparseGPT, can amplify existing biases in models, with outputs varying significantly depending on persona cues in the prompt. In this paper, we introduce Debias-SparseGPT, a post-training pruning method incorporating representational debiasing using a second-order term defined over demographically contrasting inputs. We perform empirical validation of our method over a wide range of generative LLMs. Across models and sparsity regimes (25%, 50%, and structured 2:4 sparsity), Debias-SparseGPT consistently reduces pruning-induced bias compared to SparseGPT while preserving model perplexity and zero-shot accuracy. Under the most restrictive 2:4 structured sparsity pattern, which most aggressively degrades model quality, augmenting the calibration set with long-context, content-rich examples further improves both downstream performance and fairness. Overall, Debias-SparseGPT advances the bias-performance trade-off while preserving the computational efficiency of sparse models.

Differential privacy (DP) has emerged in the computer science literature as a measure of the impact on an individual's privacy resulting from the publication of a statistical output such as a frequency table. This paper provides an introduction to DP for official statisticians and discuss its relevance, benefits, and challenges from a National Statistical Organisation (NSO) perspective. We motivate our study by examining how privacy is evolving in the era of big data and how this might prompt a shift from traditional statistical disclosure techniques used in official statistics–which are generally applied on a cell-by-cell or table-by-table basis–to formal privacy methods, like DP, which are applied from a perspective encompassing the totality of the outputs generated from a given dataset. We identify an important interplay between DP's holistic privacy risk measure and the difficulty for NSOs in implementing DP, showing that DP's major advantage is also DP's major challenge. This paper provides new work addressing two key DP research areas for NSOs: DP's application to survey data and its incorporation within the Five Safes framework.

Accurate scene understanding in confined, cluttered spaces such as shelves is essential for service robots, as many everyday tasks require them to locate and retrieve objects reliably. Yet, it remains challenging due to severe occlusions, restricted accessibility, and the need to avoid excessive scene changes. In this paper, we propose Multi-Skill Manipulation-Enhanced Mapping (MS-MEM), an evidential framework for uncertainty-aware mapping that integrates active viewpoint selection, object pushing, and grasping. MS-MEM combines scene-level metric-semantic evidential belief estimators with an uncertainty-aware grasp representation. This representation is learned using a novel full-evidential grasp estimator that models both grasp affordance and orientation uncertainty. In our framework, candidate perception and manipulation actions are evaluated within a unified action selection pipeline using a common information gain criterion. For manipulation actions, we further introduce a collateral disturbance constraint (CDC) that discourages excessive changes to confident regions of the scene belief. This enables MS-MEM to select actions that effectively reduce map uncertainty while limiting collateral scene changes. Experimental results show that, compared with single-skill and unconstrained baselines that ignore scene disturbance, MS-MEM achieves higher mapping accuracy while substantially reducing scene disturbance, highlighting the synergistic effects of active viewpoint selection, push, and grasp actions.

In many systems, the interaction between a deformable surface or interface and the surrounding bulk fluid is coupled with intrinsic spatiotemporal dynamics within the moving surface. Examples include tumor growth, biological tissue morphogenesis, cardiac mechanics, multi-phase surfactant chemistry, additive manufacturing, clothing wear-and-tear, and reactive combustion flows. Solving such problems requires both geometric computing algorithms to track and resolve the surface and numerical methods to solve the coupled governing equations in the surface and the surrounding bulk phase. Here, we present a fully meshfree numerical solver for such coupled bulk-surface problems with deformable interfaces. The presented solver tracks the surface implicitly, solving for the dynamic surface geometry based on stress balance coupled to surrounding fluid phases. We show convergence for a mass-conserving case on a growing sphere and solve bulk-surface problems with incompressible Navier-Stokes fluids coupled to in-surface nonlinear reaction-diffusion dynamics. Finally, we show a model of biological morphogenesis, solving simultaneously for the dynamic surface shape and the fields on the curved surface with two-way coupling.

The ability to move stably over terrain with varying slopes and textures is essential for mobile agricultural robots operating in complex and dynamic environments such as greenhouses, where small terrain irregularities can lead to significant navigation errors. This article presents a novel terrain-adaptation strategy based on the carried payload, ensuring accurate and robust trajectory tracking. The proposed approach is based on: (i) the experimental characterization of the most common types of greenhouse soil, concrete, compacted sand, and gravel, and (ii) the direct measurement of terrain slope using the IMU, in order to estimate the force with which this angle affects the motor input. Based on this information, a cascade trajectory-tracking scheme has been designed, consisting of a model-based predictive controller (MPC) in the outer loop and a PI controller in the inner loop. The system incorporates an adaptive feedforward control through gain scheduling approach, capable of adjusting to disturbances caused by variations in slope and terrain type. Simulation results demonstrate that the differential-drive robot achieves a significant improvement both in error indices and in control signal efficiency, highlighting the effectiveness and robustness of the proposed approach.

Document Visual Question Answering (DocVQA) often leverages Retrieval-Augmented Generation (RAG), where late-interaction encoders are commonly used to identify document pages relevant to a user query, before answer generation by a Large Vision-Language Model (LVLM). Existing approaches typically retrieve a fixed top-$k$ number of pages regardless of query complexity, which increases LVLM latency and may degrade answer accuracy. We introduce ViSAR (Visual Semantic Activation Retrieval), a training-free adaptive-$k$ retrieval method for late-interaction visual document retrieval. ViSAR operates directly in the embedding space to construct a query-conditioned page-level similarity matrix that highlights query-relevant semantics and dynamically determines the number of pages to retrieve. Across multiple encoders and LVLMs, ViSAR retrieves compact, query-adapted page sets that reduce RAG latency by up to 58.7%, while maintaining or improving answer accuracy compared with fixed top-$k$ and adaptive retrieval heuristics. Furthermore, we show that the similarity matrix structure correlates with answer accuracy, suggesting future directions for retrieval quality-aware document understanding.

Magnetic induction communication (MIC) is a promising technology for through-the-earth (TTE) communication. Previous studies on the MIC range have often overlooked the impact of eddy losses caused by underground materials. For TTE MIC, significant eddy losses complicate the analysis of the effective MIC range, which is vital for optimizing performance but has never been addressed in the literature. Accounting for the conductivity and permittivity of the underground medium, this paper derives the effective MIC range in TTE MIC, along with a closed-from expression that predicts the optimal carrier frequency to maximize this range. Finite element simulations validate the analysis, demonstrating that the optimal carrier frequency can significantly enhance the MIC range. It is also revealed that optimizing the antenna radius is effective in extending the MIC range for TTE and vehicle MIC applications.

In this paper, we analyze how Large Language Models (LLMs) employ worldbuilding strategies, focusing on setting as one measurable dimension of storyworld construction. We compare 1,000 AI-generated stories per model in English and German with human-authored fiction from Project Gutenberg. Building on prior work, we operationalize setting through five types of narrative space: "action", "perceived," "visual," "descriptive" and "no space", identified using fine-tuned BERT classifiers for German and English. We generate narratives using GPT 4.1, LlaMA 3.3, Mistral 3.2, and Gemma 3 and compare their spatial distributions to a human-authored baseline. We find that human-authored texts predominantly employ "action space," grounding narratives in embodied character-environment interaction, whereas LLMs systematically overproduce "perceived space," emphasizing atmosphere and affect. This divergence remains stable across narrative time. Overall, our findings show that LLMs exhibit worldbuilding patterns that differ consistently from human-authored fiction in ways that are both model-specific and language-sensitive.

Synthetic dialogue generation can support research in privacy-restricted service settings, but generated conversations must preserve communicative intent, affective meaning, and natural dialogue flow. We introduce PragAlign, a feedback-guided framework for controlled synthetic dialogue generation conditioned on service context, target intent, and target emotion, with auxiliary trait-style controls. PragAlign uses a generate–evaluate–revise loop in which an LLM-based evaluator scores intent alignment, emotion alignment, coherence, fluency, and aggregate quality, then provides criterion-specific feedback for up to three refinement rounds. On 800 matched dialogue specifications, PragAlign achieves 99.50% evaluator-defined acceptance, compared with 72.25% for one-shot generation and 95.88% for repeated generation without structured feedback. This indicates that repeated attempts account for much of the gain over one-shot generation, while structured feedback primarily improves last-mile multi-constraint satisfaction rather than broad average quality. Refinement gains are concentrated in emotion alignment, which is also the dominant failure mode in ablations. A separate human evaluation of 1,200 generated dialogues shows that intent expression and dialogue flow are highly recognizable to annotators, while emotion appropriateness is less stable and more subjective. These results support PragAlign as a quality-control framework for improving evaluator-defined communicative constraint satisfaction, while showing that affective realization and independent human-perceived quality remain open challenges.

Wireless Capsule Endoscopy (WCE) captures and streams video while passing through a patient's Gastrointestinal (GI) tract and is used to examine its irregularities. Although advantageous over conventional endoscopy, WCE suffers from limitations related to capsule size and wireless transmission, resulting in images with coarser resolution. This work presents UnCapsTSR, an unsupervised transformer-based Generative Adversarial Network (GAN) framework for improving the spatial resolution of Low-Resolution (LR) WCE images. The proposed method accomplishes SR without explicit degradation estimation of real-world LR data and eliminates the need for true LR-HR pairs. UnCapsTSR employs a Bilateral Total Variation (BTV) loss to ensure spatial continuity in SR images. A newly curated dataset from the Kvasir Capsule dataset is also presented for training WCE SR models. Generalizability is validated on KID and GIANA datasets that are not used during training. A new non-reference metric, Endoscopy Quality Metric (EndoQM), is introduced for quantitative evaluation of domain-specific WCE data. Experiments demonstrate consistent improvement over state-of-the-art unsupervised SR approaches using NIQE, BRISQUE, PIQE, and EndoQM. Statistical evaluation shows 40 to 80 percent improvement in EndoQM from LR to SR across the evaluated datasets.

Ontology rankers remain useful for rare-disease diagnosis because each candidate can be traced to matched patient phenotypes. Large language models (LLMs) can generate differential diagnoses from the same patient description, but their predictions lack an equally clear evidence trail. Rather than asking which system should replace the other, we ask whether an LLM can improve the ranker without giving up its evidence. Our behavior-based fusion model examines the two ranked lists, their agreement, and the ontology support behind each candidate, and learns how much to rely on each system for the individual case. Before comparison, we remove a documented test-set leakage pathway caused by benchmark cases and ontology annotations being derived from the same publications. Across eight open LLMs, fusion improves Phenomizer Recall@1 by 7.86 percentage points on Phenopacket Store and 20.18 points on RAMEDIS. When paired with DeepSeek-V4-Flash through an API, a fusion model trained only on the other LLMs improves Recall@1 from 0.1657 to 0.2176, a 5.19-point gain, without retraining. For 90.8% of correct fused diagnoses, the disease retains candidate-level ontology evidence that can be inspected. These results show that LLMs can strengthen an established diagnostic tool without discarding the structured evidence that makes it useful.

Target templates define what a visual tracker searches for, yet the templates available at inference trade off localization certainty with appearance freshness: the initial ground-truth template is exact but becomes stale, whereas recent templates better reflect the current appearance but are cropped from uncertain predictions. We quantify this bottleneck with a non-deployable oracle that supplies an exact current-frame target crop, improving AUC on LaSOT by 15.2 percentage points. This gap reveals a training-only opportunity: frame-level ground truths provide exact current- and future-frame target crops, although such crops are unavailable at deployment. We introduce Privileged Appearance Transfer for Tracking (PATT), a teacher-student training framework that transfers these privileged appearances to a deployable tracker through multi-level representation prediction. The privileged teacher observes exact target crops from past, current, and future frames, whereas the student receives only past-frame templates and learns to predict the teacher's search representations. To avoid transferring unreliable teacher signals, PATT weights this transfer by the teacher's relative localization advantage over the student and its absolute localization accuracy. After training, the teacher, latent predictor, reliability weights, and privileged crops are removed, leaving standard student-only inference. Across seven benchmarks at two model scales, PATT achieves consistent gains under both long- and short-term tracking protocols.

Standard Delay Format (SDF)-annotated switching simulation distinguishes delay-dependent activity among functionally equivalent rewrites, but evaluating every candidate repeats timing, compilation, and replay. A zero-delay proxy can remove timed evaluations, but generating that proxy candidate by candidate can cost more than the timed work it saves. We present Batch Before You Time (BBYT), which compiles all candidates of one rewrite decision into one scoped zero-delay image and either commits a well-separated proxy winner or invokes the unchanged timed chain. Across 12 counterbalanced holdout sequences, BBYT reduces complete candidate-selection time by 18.05% on average; the design-level reductions are 5.84% and 30.25%, with both confidence intervals above zero. On a counterbalanced 8,192-transition C6288 workload, BBYT is 9.24% faster than the same gate executed with candidate-wise proxy launches. In the five-workload corpus, BBYT removes 32.52% of timed candidate evaluations and matches exhaustive timed selection on all 250 evaluated decisions. For on-demand timing-aware rewrite selection, proxy execution and fidelity continuation should use the same decision scope.

Intrusion Detection has been revolutionized due to the integration of Machine Learning. Improved detection rates, reduced false alarms, and optimized algorithms contribute to the perception of improved systems with optimal accuracy and near-perfect performance, the illusion of model efficacy. However, the value of this effectiveness diminishes when confronted with unseen attacks. In this paper, we go beyond solely algorithmic enhancements and metric adjustments in ML-based Network Intrusion Detection Systems. We design an experiment to test the generalization capabilities of certain classifiers on unseen attacks. Our approach examines the dimensionality parameter's impact through two experimental methodologies, which are applied in two distinct settings. The experimental findings reveal how effectively the models could identify even a fraction of unseen attacks and underscore structural weaknesses in ML-based IDS research and evaluation techniques. Finally, seven evaluation criteria are outlined to address these challenges.

We explore predicting eCommerce user preferences for product aspects such as brand, size, and color - a task we define as Aspect Affinity. Solving this task improves customer understanding and enables fine-grained personalization in recommendation, search, and marketing. We frame Aspect Affinity as a temporal prediction task: forecasting a users future aspect choices from their time-ordered interaction history, capturing long-term preferences that evolve beyond the current session. To this end, we propose DeepAffinity, which leverages Small Language Models (SLMs) with structured prompts and specialized prediction heads fine-tuned for this task. We show DeepAffinity outperforms standard generative fine-tuning methods, while general-purpose open-source LLMs perform poorly without task-specific tuning, highlighting their limits in modeling nuanced behavior. Finally, DeepAffinity enhances recommendation quality on a large-scale multinational eCommerce platform.

Security operations centers (SOCs) face large numbers of false alerts, making detection of cyberattacks difficult under typical resource constraints. Risk-based alerting (RBA) has been proposed as a means to reduce false alerts and has reportedly succeeded in doing so in various enterprise deployments. However, RBA has not been comprehensively evaluated until now, leaving implementation mostly guesswork based on anecdotal evidence. In this paper, we present the first systematic evaluation of RBA. To this end, we reformulate it as a continuous alert prioritization problem rather than a binary decision problem (i.e., whether an alerting threshold is exceeded), allowing us to evaluate performance across all possible thresholds and thus model SOCs of varying sizes and alert volumes. We distill five fundamental risk hypotheses, formalize them as independently parametrizable modules, and implement them in our novel experimentation suite CATS. We thoroughly assess the hypotheses across eight diverse alert datasets, six of which we created or extended to make such an evaluation possible. Our results show that certain combinations of hypotheses achieve a remarkable alert prioritization performance (AUROC $\mu=0.92$, $\sigma=0.09$ across the eight datasets), outperforming a straightforward prioritization by alert severity level (AUROC $\mu=0.72$, $\sigma=0.21$). We conclude that RBA can substantially reduce the number of false alerts that analysts have to review and thus has the potential to mitigate cybersecurity alert fatigue. In addition, it serves as a strong baseline for more complex, resource-intensive alert triage approaches (e.g., based on large language models).

End-to-end autonomous driving in urban environments requires robust decision-making under partial observability and complex multi-agent interactions. Severe occlusions and dense traffic at intersections limit the perception capability of single-agent systems, motivating recent efforts on Vehicle-to-Infrastructure (V2I) cooperation for perception and planning. However, existing evaluation protocols face a fundamental trade-off: open-loop evaluation fails to capture error accumulation and recovery from deviations, while closed-loop evaluation is costly, difficult to scale, and often relies on simulated environments that may suffer from domain gaps. To bridge this gap, we propose VIPS, a benchmark for cooperative autonomous driving in V2I settings based on pseudo-simulation. VIPS extends pseudo-simulation by integrating vehicle and infrastructure observations. This enables scalable yet realistic evaluation of robustness and error propagation without full simulation. We further present CoS-V2X, a cooperative planning framework based on sparse representations. CoS-V2X models vehicle-infrastructure interactions using compact features for efficient communication and robust decision-making under heterogeneous observations. Code and dataset are available at https://vips2026.github.io.

Neural Architecture Search (NAS) has emerged as a powerful paradigm for automatically designing deep neural networks; however, its practical adoption is often limited by substantial computational cost. To alleviate expensive full-training evaluations, surrogate-based methods have been introduced to estimate network performance efficiently. Nevertheless, existing approaches-particularly model-based surrogates-require training many candidate architectures and involve additional optimization overhead. In this work, we propose a Model-Free Surrogate PSO Network (MFSPNet) for evolving convolutional neural network architectures. The proposed method integrates a lightweight model-free surrogate predictor within a particle swarm optimization (PSO) framework, eliminating the need for pre-trained surrogate models. Specifically, MFSPNet introduces two key contributions: (1) a validation-loss-driven exponential moving average estimator (VLE-EMA) that captures early generalization behavior for reliable architecture ranking; and (2) a block-based dense connection strategy that enables effective stacking of evolved blocks while mitigating vanishing-gradient issues. This design also facilitates transferability of learned blocks across datasets. Extensive experiments demonstrate that MFSPNet achieves competitive performance with reduced computational cost. Under a consistent training protocol with ten independent runs, the proposed method attains error rates of 3.91%, 17.68%, and 1.91% on CIFAR-10, CIFAR-100, and SVHN, respectively, along with top-1/top-5 error rates of 28.29%/12.82% on ImageNet, while requiring less than three GPU days for architecture search. Due to computational constraints, the ImageNet result is based on a single run and should be interpreted as indicative of scalability. Overall, MFSPNet provides an efficient and reliable framework for cost-aware neural architecture search.

We present CivBench, an open-source benchmark for evaluating language model agents in long-horizon, tool-mediated environments through the Model Context Protocol (MCP). A single episode can span over 300 turns and produces thousands of tool calls over a large action space, requiring sustained planning, state monitoring, and execution under partial observability. The environment exposes 76 MCP tools and a deterministic narration layer that retrieves and formats game state as structured text. We use CivBench to characterise agent behaviour across four model families in 23 admissible runs. The sample is a pilot, not a model ranking: aggregate outcomes do not reliably discriminate models at this scale. Instead, we introduce two interface-level metrics that the environment makes measurable: Proactive Monitoring Rate (PMR), capturing whether agents actively query latent strategic state, and RAG@10, capturing whether commitments stated in structured planning reflections are executed within ten subsequent turns. Across runs we observe two consistent patterns under a shared playbook protocol. Agents under-monitor strategically relevant state that is available but requires explicit querying: despite playbook guidance to query victory progress every 20 turns, agents do so only every 30-75 turns, and in 8 of 20 detectable defeats they failed to query within the 20-turn warning window before game end. Agents also frequently fail to execute near-term commitments stated in their own planning reflections (RAG@10 between 48.2% and 65.8% across models). Both patterns arise despite tool access and explicit guidance, and we interpret them as deviations under instruction rather than absences of capability. We release the environment, scenarios, logs, metrics, and analysis pipeline at https://github.com/lmwilki/civ6-mcp.

Given the safety-critical nature of many embedded systems, their safety assurance is essential. Because such systems are typically stochastic, probabilistic model checking is a particularly important technique. However, there is a well-known scalability issue due to state-space explosion, especially when verifying complex properties. To mitigate this issue, we propose specification-guided path shortcutting for probabilistic systems, focusing on Markov chains (MCs) and $\omega$-regular properties. The key idea is that, when the verified property is fixed, certain sequences of transitions in an MC can be replaced with a single transition without changing the satisfaction probability, and thus, we can reduce the state space of the MC. We implement the proposed path shortcutting and evaluate its contribution to the performance of probabilistic model checking, using Storm as the baseline model checker. The results suggest that our approach often outperforms the baseline, particularly on benchmark instances with complex specifications.

Digital rights increasingly exist in law but remain difficult to exercise through the information systems that mediate them. Using disciplined conceptual synthesis and problematization, this critical-conceptual IS paper explains the gap through the interaction of legal heterogeneity, conflicting organizational and commercial incentives, fragmented architectures, and asymmetrical control over rights-relevant acts. It then theorizes a human-compatible rights layer: a governed sociotechnical capability for standardized, machine-readable, bidirectional, and jurisdictionally plural communication of requests, consent, refusal, withdrawal, objection, records, and support. Comparing California-style opt-out signals, the EU's mixed lawful-basis regime, and P3P, DNT, GPC, and ADPC, the paper derives seven normative requirements, develops rights by architecture as a bounded emancipatory policy argument, and treats a proposed GDPR provision on automated and machine-readable privacy management as a policy case for moving from banner-based compliance toward rights-supporting digital infrastructure.

Machine learning (ML) education faces two persistent and connected obstacles: many educational tools present ML as an opaque black box, which leaves learners with a superficial understanding, and this same opacity prevents users from forming the calibrated trust that appropriate reliance on AI systems requires. We present ICE-T, a didactic framework that integrates three mutually reinforcing facets: intermodal transfer grounded in Bruner's enactive, iconic, and symbolic modes of representation, computational thinking operationalized through the Use-Modify-Create progression, and explanatory thinking supported by a process model. Connecting the framework to the empirical literature on algorithm aversion, AI literacy, and mental model formation, and to systematic reviews of the K-12 ML activity landscape, we argue that the three facets supply the cognitive mechanisms that the trust calibration literature identifies as drivers of appropriate reliance: representational richness, graduated process control, and the capacity to contextualize errors. On this basis, we propose that trust calibration be treated as an explicit educational objective, with ICE-T as a principled and scalable means of achieving it.

Knowing where and how fast objects are moving within a scene is important across various domains. Usually, cameras are used to capture the data necessary for this task, but adding cameras often raises privacy concerns, and the quality of captured frames is heavily influenced by lighting conditions. In this work, we explore using WiFi channel state information (CSI) instead of camera frames for optical flow estimation. We propose WiFlow, a CSI based flow estimator, a preprocessor evaluation for CSI, and three model architectures that offer different trade-offs between accuracy and complexity. Further, we create the first dataset for training and evaluating CSI-based optical flow estimators, and our experiments provide insights into key design elements for this task. Code and data are available at https://visinf.github.io/wiflow.

In this paper, we propose a scalable Kronecker-based approximation that captures cross-layer interactions without storing the entire Fisher matrix, enabling practical Hessian analysis for billion-parameter networks where full computation is infeasible. Our approach reveals consistent vulnerability patterns: value projection layers exhibit the highest sensitivity and strongest cross-layer correlations across multiple model families, while other components exhibit architecture-specific behaviors. Through extensive experiments on quantization, sparsification, inter-layer corruption, and post-corruption fine-tuning, we demonstrate that our approximation strongly correlates with both performance degradation and recovery. Our framework provides a practical, theoretically grounded tool for identifying fragile components in large models, opening new avenues for guided compression and optimization strategies, such as mixed-precision allocation, layer-wise sparsity, and adaptive low-rank decomposition across layers and even individual weight groups.

Semantic triggers in federated learning (FL) can be less conspicuous than synthetic patches, but sample-dependent placement may weaken backdoor implantation across aggregation rounds. This challenge is compounded in decentralized FL (DFL), where topology-dependent peer aggregation repeatedly mixes local models. CACTUS converts label-consistent semantic pairs into target-directed representation shifts. Mask-guided, modality-specific operators isolate trigger effects, couple them across samples, and apply the shifts counterfactually to clean non-target embeddings before peer aggregation. Experiments cover speech, text, tabular, and image tasks under nine aggregation rules. With 30% malicious nodes, CACTUS reaches a nine-rule mean attack success rate (ASR) of 51.2% on Speech Commands and the highest nine-rule mean ASR among evaluated attacks on three of four modalities. Sensitivity analyses show that ASR varies with network topology and increases with the malicious-node ratio. These results indicate that CACTUS can propagate backdoors through repeated DFL aggregation.

Digital elevation models (DEMs) can provide accurate height information, making it invaluable for analyzing the lunar surface. As the European Space Agency (ESA) prepares for future lunar missions that aim to land on the Moon, a precise method for height estimation will be essential for hazardous terrain that could endanger the landing approach. Traditional approaches to generate DEMs from imagery, such as shape from shading (SfS) and stereophotogrammetry (SPG) have been proven highly valuable for this task. However, due to advancements in machine learning, especially computer vision, the focus has shifted towards monocular depth estimation via deep learning. The lunar surface is covered by rocks and craters, and classic hazard detection methods rely solely on 2D image data. Our goal is to address this issue by developing a relative lunar surface height estimator that can provide additional information for hazard localization. In this letter, we present a methodology that builds on the well-known zero-shot relative depth estimation model Depth Anything V2 (DAV2). Other works have been using it as a state-of-the-art comparison for their proposed lunar DEM estimation method, but without adaptations to the target domain. Thus, it may underperform. Therefore, we propose a fine-tuning strategy with publicly available SPG-derived DEM data of the lunar surface. Our results demonstrate a significant improvement in performance compared to the zero-shot model, effectively transforming DAV2 into a reliable relative depth estimator of the lunar surface.

We study the synthesis of polynomial invariants for probabilistic transition systems (PTS) based on martingale theory. We present tractable methods to verify that such polynomials are indeed invariants, in the sense that their expected value upon termination is the same as their value at the start of the computation. We do this by applying the Optional Stopping Theorem (OST) in the form of a specific precondition. This precondition requires the existence of an integrable dominating function for the martingale expression, which implies uniform integrability; we refer to this condition as dui. For linear PTS we simplify the dui property to proving finiteness of the expected value of an expression depending on the update matrix, the degree of the martingale expression, and the stopping time. Specifically, if all random samples have finite moments and we can verify a moment bound on the runtime of a linear loop, then we can automatically synthesise polynomial loop invariants that satisfy the OST. Notably, dui allows for the sampled distributions to have unbounded support, which is a novel contribution to the field.

Adversarially robust models often overfit to a specific attack budget, necessitating multiple specialized models for diverse and dynamic adversarial environments, a strategy that becomes fundamentally intractable as the threat space grows. This raises an open challenge: can we achieve strong robustness across a continuum of threat levels within a single model? We propose the Threat Conditional Network (TCN), grounded in a representation factorization framework that decomposes representation learning into a threat-invariant shared backbone and a lightweight threat-conditional adaptor. TCN conditions a single model on the perturbation level via Fourier-based embeddings and channel-wise affine modulation, and is trained against a distribution over perturbation budgets, enabling flexible and seamless adaptation across an infinite continuum of threat levels during inference. Extensive experiments on CIFAR-10, CIFAR-100, and Tiny-ImageNet show that TCN matches or surpasses a full ensemble of budget-specialized models with a single set of parameters, generalizes to unseen perturbation budgets, and transfers robustly under mismatched threat conditions, with only 4.6% parameter overhead. These contributions chart a promising path toward adaptive and generalizable robustness in dynamic and diverse threat environments.

Large language models can look capable of logical reasoning, but correct or incorrect answers alone tell us little about what the model represents internally. We study logical verification in five open-weight transformer models using matched valid–invalid premise–claim pairs that vary across inference families, semantic domains, templates, and difficulty levels. Despite near-chance behavioral performance, logical validity is often almost perfectly decodable from hidden states and remains strongly decodable under held-out templates, domains, and inference families. Validity also remains highly decodable on behaviorally incorrect examples in the conditions where correctness-conditioned evaluation is well defined. At the same time, exhaustive leave-one-out tests reveal clear limits to this generalization, and interventions along probe-derived validity directions have only weak, nonspecific effects compared with random controls. Our results suggest that representing validity, expressing it in behavior, and using it causally are distinct. Validity related information can be strongly decodable from a model's hidden states without being reliably expressed in its output.

Collaborative Brain-Computer Interfaces (cBCIs) offer a promising mechanism to augment team decision-making, but existing approaches rely exclusively on evidence available only after a decision has been made and reported, such as reaction time or stated confidence. This limits their use to explaining or discounting a decision after the fact, rather than informing a team's response before it is finalised. We tested whether spatial-covariance EEG features could instead provide a genuinely pre-emptive signal of an operator's decision correctness, available within the response window itself, and whether such a signal depends on cognitive workload. Using a continuous virtual reality target-detection task, participants (N = 23) completed a within-subject workload manipulation (High vs. Low). At the team level, weighting votes by this pre-emptive neural signal, available before a response is committed, produced substantial accuracy gains on contested (evenly-split) trials under High Workload (57% to 88% as team size increased from 2 to 16), but was actively detrimental under Low Workload. Critically, this advantage held even against post-hoc behavioural signals: confidence was the strongest single team-level signal overall, but by definition cannot inform a decision still in progress, whereas the neural signal can. These findings indicate that EEG-based decision-reliability signals are not a general-purpose team augmentation tool, but a workload-conditional one, with clear implications for when and how cBCI systems should be deployed in operational teams.