Papers

This study presents an empirical benchmarking comparison between Kolmogorov-Arnold Networks (KANs) and Multi-Layer Perceptrons (MLPs) on structured tabular classification tasks. Motivated by the growing interest in KANs as an alternative function-approximating architecture, we evaluate their out-of-the-box performance on twelve publicly available datasets spanning binary, multiclass, multilabel, and ordinal problems. Both models were trained under standardized preprocessing, architecture, and fixed hyperparameter settings, with performance assessed using test accuracy and F1-Score, paired hypothesis testing, and effect size analysis. Results show that KANs statistically outperform MLPs in binary and multiclass domains and achieve a significant aggregate advantage across all datasets. However, the observed medium effect size (d = -0.46) raises an important cost-benefit consideration: while KANs offer superior generalization through adaptive spline-based mappings, this advantage comes with substantially higher parameter and computational complexity relative to the MLP baseline. These findings suggest KANs are the preferred choice for high-precision applications, while MLPs remain a robust and efficient option for resource-constrained environments. Future work should extend this analysis to additional data modalities to further refine these architectural selection criteria.

Recent text-to-audio models generate high-quality audio, but often fail to follow instructions involving multiple sound events and temporal order. This gap arises because existing evaluation and training signals mainly emphasize global similarity or perceptual quality, with limited supervision on instruction-level correctness. We propose an instruction-level framework that uses audio-aware large language models (ALLMs) as fine-grained judges to verify target event presence and temporal relations in generated audio. After validating ALLM judgments on benchmarks and through human verification, we use their feedback to construct preference pairs for direct preference optimization. We further introduce S3Bench, a narrative benchmark for evaluating multi-event temporal instruction following. Experiments show that our method improves event completeness, temporal ordering, and joint instruction-following accuracy across existing benchmarks and S3Bench, while maintaining audio quality.

In bandit problems, standard regret-minimizing algorithms treat exploration as an amortized cost, which can expose early participants to unfair ex-ante losses in settings such as clinical trials. Recent work addresses this by evaluating the sequence of per-round expected rewards through the generalized $p$-mean, interpolating between utilitarian welfare ($p=1$), Nash welfare ($p\to0$), and Rawlsian fairness ($p\to-\infty$). Although tight guarantees are known for $p\ge0$, the strictly fair regime $q=-p>0$ remains unresolved because negative-power means are dominated by the smallest per-round rewards. For $\sigma$-sub-Gaussian rewards with nonnegative means, the best prior algorithm relied on uniform early exploration and achieved regret $O(k^{(q+1)/2}/\sqrt{T})$, while the only general lower bound was the classical $\Omega(\sigma\sqrt{k/T})$. Thus it was unclear whether the extra dependence on $k$ was intrinsic to strict fairness or an artifact of uniform exploration. We close this gap by identifying the exact polynomial price of strict fairness. Using a needle-in-haystack construction, we prove an algorithm-independent lower bound $\Omega(\sigma\sqrt{k^{\max(1,q)}/T})$; for $q>1$, this shows that the penalty $k^{q/2}$ is information-theoretically unavoidable. We then introduce UCB-HARE (Harmonic Anchored Rank Exploration), which replaces uniform exploration with an inverse-weighted harmonic rank schedule protected by a certified positive-mean anchor. Its regret is $\widetilde{O}(\sigma\sqrt{k^{\max(1,q)}/T})$, matching the lower bound up to logarithmic factors. Experiments on synthetic instances confirm that UCB-HARE improves over uniform-exploration baselines, with gains increasing as $q$ grows.

What happens to an LLM agent's tool choice when the reliable tool silently changes within an ongoing session? We borrow the notion of set-shifting from cognitive psychology to study how well agents adapt to hidden reliability shifts. Our cognitive test for LLM agents mounts libraries of redundant tools and skills, in which many tools solve the same task but differ in hidden reliability. Using a branching schedule, we shift the reliable tool group in the environment and compare it with a stable control, allowing us to isolate the effect of each shift on the agent's behavior. We conduct our study on a panel of LLMs equipped with harnesses and show that the same set of shifts results in distinct behaviors across models: some latch onto a fixed routine within a few turns, whereas others continue to vary. Less capable models often omit the reliable tool group, while frontier models keep calling it alongside the other groups. We introduce a suite of measures to quantify agent behavior after reliability shifts. While policy prompting substantially alters behavior in some tested models, our findings highlight agents' brittleness when changes occur in indirectly observable context.

A field can reformulate its computations freely exactly where its demand is stated independently of any incumbent implementation, and finds itself unable to when the incumbent's own output has quietly become the specification. This note offers that observation as a lens on computational reformulation for modern accelerators, where posing a problem in a hardware-friendly form can yield large speed and energy gains, but only if a replacement can be judged at all. Building on the test-oracle problem (Weyuker; Barr et al.), on requirements engineering's notion of implementation bias (Zave and Jackson), and on the case for judging approximate designs by acceptability rather than numerical proximity (Felzmann et al.), it names the pathology "baseline capture": the moment an incumbent stops being evidence that a demand can be met and becomes the definition of meeting it. It then separates two questions that are easily confused: whether a reformulation can be judged at all, which turns on the existence of an incumbent-independent demand, and whether its discovery can be automated, which turns additionally on the cost of evaluating that demand. Short cases – shortest-path routing, learnable audio frontends, ZIP-215 for Ed25519 signature validation, CESM-ECT for climate models, and a single-GEMM audio frontend measured at 1.64x-3.29x speedup and up to 3.03x less energy – illustrate the pattern and the move of "buying a verifier": making a demand explicit, operational, and independent of the incumbent. No component is claimed novel in isolation; the contribution is the synthesis and the single question it makes easy to ask of any reformulation result – does its acceptance test mention the incumbent's output?

This paper is an extension of a paper presented at the ICAART 2026 conference, which introduced LEA (Learning Engagement Assistant), an adaptive AI tutoring agent combining course-specific Retrieval-Augmented Generation (RAG) with structured Knowledge Component (KC) models across integrated Chat, Tutor, and Quiz modes. That prior work validated LEA on a single STEM course (CMP511) exclusively through simulation, using synthetic learner agents. This paper extends that work by reporting the first classroom deployment of LEA with real students (n = 8, CMP511) and the first empirical test of its cross-course scalability, deploying the system across three courses spanning two academic levels and two disciplinary domains. The study reveals a divergence from simulation predictions across modes, showing that synthetic evaluation alone cannot anticipate all aspects of real deployment. A RAGAS-based cross-course scalability evaluation (660 questions) finds Answer Relevancy and Context Precision broadly stable across courses (0.88-0.94 and 0.88-0.90 respectively), while Faithfulness declines with curriculum distance from the system's original course (0.69 to 0.50), a preliminary finding that may reflect generation logic tuned to the system's original subject rather than a scalability limitation. These findings suggest that while the orchestration layer requires no modification, full course-agnosticism of all downstream components requires further investigation.

LLM-as-a-judge is widely used to provide feedback and selection signals in closedloop regeneration, but this use remains insufficiently validated. We study it in table recognition, where deterministic TEDS evaluation provides a controlled testbed, using FinTabNet and OmniDocBench. Three findings emerge. First, judge signals were weak on both datasets: scores frequently tied, rankings were not reproducible, and no tested judge score policy, whether selecting among candidates or accepting revisions under a conservative score margin, improved on the first output on both datasets. Iteration produced better candidates, but the judge recovered them at most partially on one dataset and not at all on the other. Second, severe losses occurred even without specific judge feedback, supporting target-preservation failure under unconstrained regeneration as a proximate mechanism. Third, a structure-preserving instruction reduced the severe-loss rate, significantly on FinTabNet and directionally on OmniDocBench, but produced no improvement, and in an exploratory 2x2 analysis this protection was not stably observed when judge feedback was retained. These results do not dispute the value of LLMs as evaluators, but show that the tested reference-free judge signals were too weak and unstable to drive candidate selection in this setup, and that evaluation-style evidence alone was insufficient to establish closed-loop optimization utility. Iterative refinement requires, at minimum, a verification signal that deterministically detects structural change, rather than judge scores alone.

We judge whether AI systems are safe largely by watching how they behave. Alignment faking would break this: a model that behaves well only when it believes it is watched, or that complies now to avoid being changed by training, can pass every behavioral test. Such behavior has been reported in a frontier model, and open-weight models that comply more under a monitoring prompt are now read as showing it, with hidden-state probes proposed as detectors. We test both inferences, and neither holds. Of 18 model configurations, 11 comply more when told they are monitored. Yet a prompt factorial on four models shows that the training threat has no detectable effect: removing it changes compliance by -3.3 to +1.6 points. The prompt's wording and framing drive the gap instead. Probes fare no better. They separate monitored from unmonitored compliant answers (AUROC 0.96-0.99), but they separate refusals just as well: they read the prompt, not the faking. A steering direction learned only from refusals moves behavior as much as the "faking" direction. We release PRISM, five checks and a toolkit that catch each of these failures before they become claims.

Constraint Answer Set Programming (CASP) is a hybrid reasoning paradigm that combines Answer Set Programming (ASP) with Constraint Processing and Satisfiability Modulo Theories (SMT), enabling powerful declarative encodings of complex combinatorial search problems. This paper presents the design and implementation of EZSMTV3, an extensible SMT-based CASP framework that advances the translational approach to CASP solving. Building upon the foundation of the EZSMT+ system, EZSMTV3 introduces a more expressive input language, supports optimization via weak constraints, and offers foundations for streamlined integration of new constraint types. Rather than implementing custom search procedures, EZSMTV3 leverages state-of-the-art SMT solvers, such as CVC5, YICES, and Z3 to perform reasoning. The paper provides benchmarking results comparing EZSMTV3 with its CASP peers such as CLINGCON, CLINGO[DL], and CLINGO[LP], while showcasing its ability to handle mixed-domain constraints involving both integers and reals. The system provides a robust platform for future extensions and theoretical exploration within the CASP domain.

Autonomous highway overtaking demands foresighted decision-making to handle complex interactions, stochastic traffic evolution, and temporal risk accumulation. However, standard safe reinforcement learning approaches typically rely on implicit value-based risk estimations rather than explicit dynamics modeling, thereby struggling to accurately capture complex risk propagation over multi-step horizons. This limitation frequently results in behaviors that are locally safe but induce substantial latent risks in the long term. To address this, a World Model-based Risk-aware Mixture-of-Experts (WM-RMoE) framework is proposed. First, a learned latent dynamics model facilitates parallel multi-step rollouts, elevating safety assessment from the action level to the trajectory level via cumulative risk evaluation. Second, to enhance robustness under varying interaction intensities, a hierarchical gating mechanism dynamically coordinates experts across long-horizon, short-horizon, and rule-based safety modules. Furthermore, a Gaussian Mixture Model is integrated to preserve multimodal maneuvering branches, thereby mitigating the issue of behavioral mode averaging. Experimental results demonstrate that WM-RMoE significantly outperforms representative baselines in terms of safety compliance, decision stability, and generalization capability. Furthermore, benefiting from the risk-aware formulation, the proposed framework uniquely exhibits the ability to generate foresighted and semantically distinct overtaking maneuvers across diverse traffic densities.

LLMs now sit on every side of evaluation: as examinees scored on benchmarks, judges of other models' outputs, and raters of human-generated content. Each paradigm can be viewed as a measurement problem, where a latent property of an object is probed with items from an instrument (e.g., benchmark) by judges or raters. Standard evaluation practices often neglect the contributions of each core component to the end result, limiting our understanding of what is being measured. Rasch measurement theory (RMT) is well-suited to this problem. RMT decomposes ordinal ratings into separable facets on a common scale. It further provides a battery of diagnostics that can identify miscalibrated measurements and rater biases. We present a case study of RMT applied to the LLM-as-rater paradigm using the Measuring Hate Speech corpus, whose construct was itself built under RMT. We fit a series of many-facet Rasch models to annotations from nine LLMs spanning families and capability levels. Our analyses show that LLMs systematically differ from human raters in severity, item-level calibration, question-order robustness, target-identity sensitivity, and rating scale use, all of which standard evaluation practice would largely obscure. Overall, we argue that RMT belongs in the toolkit for evaluating LLM-as-examinee, -judge, and -rater paradigms.

In this paper we present a generative AI application developed to support both teachers and students in secondary education. The system employs two Large Language Models-LLMs, Gemini and DeepSeek, and a Small Language Model-SLM, Gemma, integrated within a Retrieval Augmented Generation - RAG framework, creating a pedagogically grounded, Greek-language assistant capable of adapting its reasoning and communication style to the user role. Unlike conventional chatbots, the assistant introduces pedagogical persona switching, a dual-role mechanism that enables the same AI model to act as both a teaching companion and a learning guide. Utilizing a RAG paradigm tailored to the Greek educational domain, the architecture segments official textbooks into coherent units. Enriched with specific metadata, these units preserve curricular structure and instructional context, demonstrating how generative AI optimizes modern instructional design. The initial case study focuses on home economics in Greek lower secondary education, a cross-disciplinary subject that integrates elements of economics, health education, and social responsibility. The assistant has been developed to support both learners and educators in complementary ways. In future classroom implementations, students will be able to use it to clarify key concepts such as financial literacy, resource management, and healthy living, while teachers could employ it to design authentic instructional materials, formative assessments, and classroom activities aligned with the official curriculum. The study elevates the concept beyond a simple chatbot, proposing a structured, contextually adaptive framework for pedagogical generative assistants that effectively bridge technology, curriculum, and human learning.

People are increasingly using large language models (LLMs) to seek advice, including for parenting. Parenting is a critical and socially sensitive domain. Thus, evaluating advice provided by LLMs requires indicators beyond aggregated information quality benchmarks to consider relational and behavioural elements of the responses. With a multi-dimensional rubric created by parenting experts, this paper evaluates 15 LLMs across 100 parenting scenarios in 2 languages (English and Chinese), using an LLM-as-a-judge method. Results show that aggregate scores can hide rubric item-specific weaknesses, models implicitly encourage different parenting styles, and language influences responses. We highlight the importance of evaluation output auditability and challenges involved in evaluating LLM-generated advice in domains like parenting. Our findings provide important insights for selecting LLMs for direct user engagement and the development of user-facing parenting advice applications.

Long-term memory is increasingly central to LLM agents, yet memory design remains a highly coupled architecture problem: what to encode, how to store it, how to retrieve it, and how to manage it can vary substantially across tasks and backbone models. We construct a discrete search space with 5 encoders, 5 stores, 6 retrievers, and 4 managers, and show that no single memory architecture consistently dominates: different tasks favor different module combinations, leading to substantial performance gaps. Motivated by this, we propose AutoMem, a text-gradient recursive self-improvement framework for task-adaptive memory architecture search. AutoMem optimizes over the factored space through two components: Experience-Guided Architecture Search, which proposes candidate architectures from historical search trajectories and accumulated reflections, and Failure-Guided Module Diagnosis, which localizes memory-related failures to specific modules and converts them into targeted textual feedback. Experiments on GAIA, WebWalkerQA, and xBench-DeepSearch across two LLM backbones show that AutoMem consistently discovers task-adaptive memory architectures that outperform the strongest human-designed memory baselines, improving accuracy by $2.8$ points on average across six benchmark-backbone settings. Further analysis shows that AutoMem achieves a favorable accuracy-efficiency trade-off, reducing token cost by $14.3\%$ over the strongest accuracy baselines under Qwen3.5-122B-A10B, while also finding stronger architectures than substantially larger random searches within only a few guided iterations.

Agreement with human judgments is a common proxy for evaluating the alignment of large language models (LLMs). Yet agreement in final labels does not show that human annotators and models rely on the same moral grounds. Two agents may reach the same judgment while appealing to different principles, contextual assumptions, or interpretations of the situation. We test this distinction using a curated 500-item ETHICS-derived benchmark spanning five domains of moral judgment, with new human annotator and LLM annotations of both final labels and supporting rationales. Across frontier and open model families, agreement with human annotator majority labels is often high. However, rationale-level analysis reveals systematic divergence in the moral grounds expressed by human annotators and models. In particular, models redistribute attention across categories such as harm, respect, promise-keeping, justice, desert, and excuse relevance, even when their final labels match the human annotator majority. Our results show that agreement should not be treated as equivalent to alignment. Label-based evaluation can therefore be misleadingly reassuring unless complemented by analysis of the reasons, principles, and moral priorities expressed in model judgments.

Top-k Mixture-of-Experts (MoE) routing is discontinuous, so a deployment-motivated numerical disturbance – simulated 4-bit KV-cache quantization read by a protected BF16 gate – pushes tokens across decision boundaries and flips which experts fire. This paper proposes no new mitigation; it supplies a causal apparatus, empirical findings, and a detection-limit result. A four-run apparatus prices the route-mediated fraction (RMF) of quantization damage, a token-level attribution decomposes it by mechanism, and pre-registered probes carry the findings across three architectures. On OLMoE-1B-7B at 4-bit KV (pilot), about a third of the damage is routing-mediated: RMF ~ 0.31 (discovery 0.31 [0.20, 0.41]; process-replicated mean 0.313 +/- 0.020; pre-registered re-execution 0.231). The deployable router margin detects that a flip occurred (AUC 0.772) but cannot tell a harmful flip from a helpful one (at chance): among the tested local, inference-observable router statistics we find no predictor of a flip's loss sign above chance – an empirical benefit-detection barrier bounding selective repair restricted to this feature family. The signed-flip tax and sign-inseparability carry cross-model; the clean-reference remedy's payout is architecture-modulated; a controlled same-checkpoint flag-swap re-scopes the gate's normalization convention to a damage-magnitude moderator, not a route-recoverability mechanism. A real int4 KV kernel yields a fraction compatible with the fake-quant dose curve but underpowered (95% CI [-0.111, 0.394] includes zero) – ruling out gross disagreement, not an independent replication. Hypotheses, thresholds, and evaluations were pre-registered before measurement, with misses reported; a pre-registered held-out read replicates the partition and the near-cancelling tax out of sample, while the strict impossibility exclusion narrowly misses.

Training deep learning models on variable long sequences poses significant computational challenges. Existing methods force a difficult trade-off between efficiency and ease-of-use. Simple approaches use static configurations that cause workload imbalance low efficiency, while complex methods introduces significant complexity and code change for new models. To break this trade-off, we introduce Data-Centric Parallel (DCP). Its core principle is to let the data itself drive the runtime. It achieves this by dynamically adjusting direct runtime settings (e.g., parallel size, gradient accumulation, recomputation) based on each batch's sequence length. Empirical results demonstrate that our method achieves up to a 2.88$\times$ speedup on 32 H200 GPUs. Designed for generalization, it can be integrated into any model with 10 lines of code. We anticipate this simple yet effective approach will serve as a robust baseline and facilitate future advancements in distributed training for variable long sequences.

A recurring question in the design of scalable multi-agent systems – from robot swarms to collectives of large-language-model (LLM) agents – is whether adding more agents can, on its own, overcome performance limits, or whether a qualitatively deeper organization is required. A recent preprint argues that flat, homogeneous multi-agent systems face an irreducible, population-independent “causal floor” on achievable error, removable only by hierarchical (nested-loop) organization. Using a controlled disturbance-rejection testbed with an exactly computable optimum, we show this conclusion is too strong and replace it with a quantitative resource model built on three resources: population width $N$, per-agent internal-model memory $d$, and prediction across the observation delay $\tau$. We establish three claims. (i) The achievable floor is governed not by architectural hierarchy but by per-agent internal-model content: a flat, homogeneous swarm whose agents carry a matched internal model of the disturbance matches or beats a designed two-loop hierarchy at equal per-agent memory – so temporal depth can be dynamical (recurrent memory), not architectural (nesting). (ii) The three resources are not mutually interchangeable; we chart the exchange rates and the hard non-exchange boundaries on an explicit width$\times$memory map, including a strict equal-total-state-budget comparison. (iii) A residual floor is set by the observation delay and the environment's unpredictability over that horizon, which we verify against the optimal controller. We quantify the price of replacing oracle knowledge of the disturbance spectrum with online learning, provide a preliminary robustness check against a mild bounded nonlinearity and a spatially-extended plant, and distill four design rules for practitioners.

Efficient long-sequence modeling remains a central challenge for large language models, as self-attention scales quadratically with sequence length. Mamba offers a linear-time alternative through selective state space recurrence, but its predominantly diagonal state transitions restrict explicit interactions among state dimensions. We propose Motif-Mamba, a structured state space model that augments Mamba with a motif-constrained low-rank recurrent pathway. Inspired by the dynamics of three-node network motifs, the proposed pathway projects hidden states into a compact dynamical subspace, imposes motif-guided interactions, and maps the resulting dynamics back to the original state space. This design enhances cross-dimensional communication while preserving the linear-time recurrent structure of Mamba. Experiments on long-sequence extrapolation, language modeling benchmarks, and brain–computer interface decoding show consistent improvements over Mamba backbones, suggesting that motif-guided low-rank dynamics provide an effective structural prior for long-range sequence modeling.

Most conferences rely on peer-review of submissions, but as generative AI makes it easier than ever to prepare submission materials, some conferences are seeing an overwhelming surge of submissions. We wanted to see if generative AI could help our conference's volunteer reviewers by pre-reviewing abstracts for certain criteria. The Bioinformatics Open Source Conference (BOSC) was well-positioned to experiment with this, as we already had a detailed rubric used by reviewers to evaluate submitted abstracts on multiple criteria, including openness (public availability of the code or other content associated with the project), valid open source license, and "runnability" (how easy it is to download, build, and run the project - an important measure of reusability). For BOSC 2026, we built bosc-pre-review, an agentic skill that assessed six review criteria, and Runabilly, which builds and tests each project in a disposable Docker container for safety. The AI only gathered evidence to present to the reviewers; humans made every decision regarding the acceptance of the abstracts. After the review period, we surveyed the reviewers to determine how useful they found the pre-review. Most of those who responded said they found it useful, but they preferred to check the AI's conclusions against their own, rather than accepting the AI results unquestioningly.

Retrieval-augmented generation services over mutable enterprise documents repeatedly execute semantically equivalent analysis requests. Answer reuse can remove GPU-bound generation work, yet response caches require dependency consistency when filings, evidence chunks, and tool outputs change. FinCacheServe treats each generated answer as a serving object indexed by enterprise intent and guarded by document versions, evidence fingerprints, tool fingerprints, model identity, and decoding configuration. A vLLM implementation evaluates SEC-derived financial-document workloads with Qwen2.5 models. On a 2,230-request hosted 7B trace, FinCacheServe skips 53.27% of LLM calls with zero observed dependency-stale outputs. Across three hosted 32B operator-suite seeds, it skips 53.31% of 544 requests, compared with 38.97% for versioned semantic caching and 22.43% for grounded-style reuse. Capacity, backend, and SLO replays show oracle-bounded cache management, 100k-entry transactional metadata behavior, and 44.30% lower estimated Wh per dependency-fresh 2s-SLO success than versioned semantic caching.

Access to specialist clinical expertise remains severely limited across sub-Saharan Africa, where physician-to-patient ratios can fall below 1:25,000 in rural settings. Existing AI-assisted diagnostic tools predominantly require reliable internet connectivity and high-specification hardware, rendering them impractical for frontline healthcare workers in district hospitals and health centres. This paper presents Aletheia, an offline-first clinical decision support system designed for low-resource healthcare contexts across sub-Saharan Africa. Aletheia is built upon Qwen2.5-3B-Instruct, fine-tuned using Quantised Low-Rank Adaptation (QLoRA) on a curated dataset of 27,000 clinical reasoning samples spanning 50 disease conditions with elevated prevalence in East Africa. Evaluation demonstrates a Top-1 diagnostic accuracy of 80% (8 of 10 cases; 95% CI: 49.0-94.3%), Top-3 accuracy of 100% (10 of 10 cases; 95% CI: 72.2-100%), BERTScore-F1 of 0.909, and METEOR of 0.467. These diagnostic figures are computed over a deliberately small set of ten representative clinical case categories, one case each, and are therefore indicative rather than statistically robust; the wide confidence intervals should be read alongside them. The system achieves an Expected Calibration Error (ECE) of 0.275 and passes the Africa Deep Tech Challenge 2026 (ADTC 2026) memory budget constraint of 7168 MB, achieving a peak inference RAM of approximately 3630 MB on the standardised benchmark laptop. These results demonstrate the feasibility of deploying large language model-based clinical reasoning at the primary care level in resource-constrained settings without cloud infrastructure.

Artificial intelligence systems increasingly mediate how science is found and credited. We asked whether missing metadata prevents AI systems from crediting work. As a boundary test, an AI system citing without access to task-relevant paper lists often produced out-of-list identifiers, some fabricated. We then tested the mechanism in real scholarly infrastructure by using OpenAlex records to hide or restore author, institution, funder, reference, and text-access links while holding works and tasks fixed. Restoring the relevant link made the corresponding attribution possible; restoring the wrong kind did not, with 0 correct answers across 469 completed mismatched tests. Thus, in these tasks, one metadata facet did not substitute for another. Missing links led to invented answers, refusals, or tool-budget exhaustion, and web search did not recover hidden author links. In sum, AI systems credited work only when record connections were visible or recoverable. This motivates Nexus-Score, a record-level check for metadata gaps, to guide repair and help prepare the scholarly record for AI-mediated use.

The traditional "one drug, one target" paradigm of structure-based drug design (SBDD) frequently proves inadequate for treating multifactorial diseases such as cancer and neurodegenerative disorders, owing to compensatory signaling pathways and the emergence of drug resistance. While polypharmacology offers a synergistic therapeutic strategy, the rational design of ligands capable of simultaneously satisfying the geometric constraints imposed by multiple targets remains a major computational bottleneck. This review positions geometric deep learning (GDL) as a powerful integrative approach to overcome these limitations. We systematically survey GDL architectures ranging from invariant graph neural networks to SE(3)-equivariant diffusion models that harness non-Euclidean molecular data to capture intrinsic three-dimensional (3D) structural interdependencies. We critically analyze GDL applications across three core dimensions, including the characterization of shared binding pockets via geometric embeddings, multi-target bioactivity prediction through heterogeneous graph fusion, and de novo generation of dual-target ligands. Particular emphasis is placed on emerging structure-conditioned generative algorithms that integrate diffusion models with reinforcement learning to autonomously resolve complex geometric conflicts between competing binding sites. Furthermore, we evaluate the pivotal role of multimodal omics integration and specialized geometric benchmarking infrastructures in validating these models. By synthesizing these methodological advances, this review elucidates the paradigm shift in drug discovery from serendipitous exploration to rational, structure-driven polypharmacological molecular engineering, thereby providing a clear, structured guide for navigating the complexities of next-generation therapeutics.

Traditional controllers are designed for specific systems and do not transfer across different system orders and dynamics. We present a Generalist Controller, a learning-based controller capable of controlling systems of varying orders and dynamics. The approach introduces a novel dynamic state-space representation using attention mechanisms with masking, enabling a single neural network, trained in one shot, to handle systems with different dimensions without architectural modifications by assigning a system tag to each system. We generated 314,630 demonstrations from 25 diverse systems, including stable, unstable, minimum-phase, and non-minimum-phase dynamics, spanning linear and nonlinear systems from autonomous underwater and aerospace vehicles to mechanical systems and chemical processes. The model learns cross-system control strategies through multi-scale temporal processing and a mixture-of-experts architecture. Simulation results demonstrate that the proposed generalist controller achieves comparable performance to system-specific LQI controllers across all tested systems, including challenging cases such as non-minimum-phase and unstable dynamics, whilst generalising to unseen operating conditions including actuator saturation, noise, disturbance, and reference trajectories not encountered during training. This work represents a significant step towards generalist control policies within a defined family of dynamical systems, demonstrating effective control across a range of single-input single-output (SISO) systems of varying order and dynamics using a single learned policy without system-specific tuning.

Vision-language models (VLMs) often answer visual questions using learned language and category priors rather than grounding their predictions in the image itself. Counterfactual images provide a natural diagnostic setting for this failure mode: when visible evidence contradicts what is usually true, a grounded model should answer from the pixels, while a prior-following model will produce a canonical but visually incorrect response. However, existing counterfactual benchmarks mainly ask whether such prior-following behavior exists. In this paper, we ask a further question motivated by the rise of tool-augmented and agentic vision systems: can additional visual evidence views help VLMs reason against their priors? We introduce PriVE-Bench, a Prior-vs-Visual Evidence Benchmark that uses paired original and counterfactual images to distinguish visually grounded answers from prior-consistent errors. We further introduce PriVE-Tools, a controlled agentic-vision-inspired extension that evaluates whether tool-derived visual evidence – including bounding boxes, crops, zoom panels, and contours – improves grounding under the same counterfactual conflicts. Across open- and closed-source VLMs, we compare raw, paired-image, and tool-conditioned inputs using accuracy, prior-following error rate, and other-response rate. Our results show that visual evidence tools can help in some settings, especially when models can use localized evidence effectively, but they are not a universal remedy: several models continue to follow language and category priors even when relevant visual evidence is explicitly provided.

Cross-domain few-shot semantic segmentation (CD-FSS) has predominantly been formulated as learning domain-invariant representations or improving support-query correspondence. Nevertheless, large domain shifts still make prototype matching unreliable: inconsistent hierarchical responses corrupt the support representation, deterministic prototypes cannot express boundary and appearance ambiguity, and treating prototypes with different reliability equally during optimization weakens foreground-background separation. We therefore propose DAUPNet, a unified framework that reformulates cross-domain prototype matching as uncertainty-aware prototype discrimination. DAUPNet first harmonizes hierarchical support-query features to provide stable evidence, then represents foreground and background prototypes probabilistically, and finally uses their estimated uncertainty to regulate contrastive optimization. On four standard target domains, DAUPNet achieves 72.6% and 76.7% average mIoU in the 1-shot and 5-shot settings, respectively, including substantial gains on the two medical domains. These results demonstrate that modeling prototype uncertainty and incorporating it into optimization provides a robust and interpretable approach to CD-FSS under severe domain shift. The code is available at https://github.com/madness-Lei/DAUPNet

Vision-Language Models (VLMs) have achieved strong progress in multimodal understanding. However, scaling dense or sparse Mixture-of-Experts (MoE) models to improve performance limits deployment in resource-constrained environments due to the trade-off between high memory usage from full loading and increased latency from on-demand loading. Recently, the Per-Layer Embedding (PLE) architecture addresses this by scaling models with large external embedding tables stored in read-only memory (ROM) and performing lightweight lookup to retrieve relevant embeddings to enhance token representations. Nevertheless, existing PLE-style methods are primarily designed for text embeddings due to the convenience of ID-based retrieval, limiting their effectiveness in VLMs where multimodal embeddings contain richer information for visual tasks. In this paper, we propose LookME, the first framework that enables lookup-based enhancement for multimodal embeddings in VLMs while supporting partitioned storage and on-demand loading. To efficiently lookup arbitrary continuous multimodal embeddings from large-scale embedding tables, we propose a hierarchical two-level lookup method employing a coarse-to-fine strategy that performs lookups from the scene-level to the intra-scene primitive-level. Furthermore, we integrate the lookup method with a sparse injection strategy, which adaptively prioritizes critical embeddings over voluminous multimodal embeddings within layers, and facilitates embedding table reuse across neighboring layers, improving the trade-off among efficiency, model size, and performance. Experiments on multiple visual benchmarks show that LookME outperforms text-only PLE-style methods, validating the effectiveness of lookup-based multimodal embedding enhancement.

Medical Vision-Language Models (Med-VLMs) require reliable reasoning from fine-grained visual evidence, yet existing models can produce plausible clinical answers by relying on language priors or medical templates rather than truly attending to diagnosis-critical regions. On-Policy Distillation (OPD) offers dense token-level supervision on student-generated trajectories and provides a privacy-compatible means of capability transfer without requiring the redistribution of raw patient data. However, standard OPD uniformly distills all tokens, causing sparse evidence-dependent tokens to be diluted by abundant clinical narrative tokens. Inspired by the success of OPD in the large language model community, we propose Med-OPD, to our knowledge the first unified post-training framework that integrates on-policy distillation with medical evidence-aware supervision for Med-VLMs. We introduce Medical Evidence Advantage (MEA), a teacher-grounded counterfactual signal that uses an answer-aware hint to focus teacher scoring on evidence supporting the target diagnosis, and measures each token's dependence on medical visual evidence by comparing teacher likelihoods under the original and evidence-degraded imaging modalities. Based on MEA, Med-OPD redistributes the distillation signal at both the token and trajectory levels, emphasizing diagnosis-critical tokens and evidence-reliant rollouts. Experiments on OmniMedVQA subsets show that Med-OPD consistently outperforms SFT and standard OPD across CT, MRI, Disease Diagnosis, and Lesion Grading. These results demonstrate that evidence-aware distillation can better strengthen medical VLMs' reliance on key visual evidence and improve reliable multimodal medical reasoning. The source code and data is publicly available at: https://github.com/yunhang8658/MedOPD.git

Model-Based Systems Engineering (MBSE) relies on formal system models as primary technical artifacts for representing requirements, structure, and behavior across the system lifecycle. With the standardization of SysMLv2 as a textual language, interest is increasing in translating natural-language descriptions directly into executable models. For practical deployment, generated models must be accepted by industrial modeling environments, not merely satisfy grammar constraints. We present a conformance-checker-driven framework for reliable natural-language-to-SysMLv2 translation that enforces production-level acceptance as the termination condition. The system embeds a SysMLv2 conformance checker within a generate-check-repair loop. Each model is evaluated using the checker, and deterministic diagnostics are incorporated into revisions until zero conformance errors are achieved. Using the production checker as the oracle ensures the framework targets deployability rather than grammar plausibility. We evaluate the approach on the full SysMBench prompt set of 151 prompts across four large language model backends, yielding 604 prompt-model cases. Single-shot generation achieves 51.16% production-conformance acceptance, while our approach achieves 100.00% conformance. By elevating production conformance from a post-processing check to a control mechanism within generation, the framework converts probabilistic outputs into production-accepted SysMLv2 artifacts suitable for loading, visualization, and engineering use.