Papers

Mechanistic interpretability has recently expanded to Vision Transformers (ViTs), with Sparse Autoencoders (SAEs) increasingly used as post-hoc tools to decompose internal representations into sparse and more interpretable features. However, because post-hoc SAEs are trained on frozen representations after the ViT has already been optimized, their latent features are not directly aligned with the downstream classification objective. We introduce SpIn-ViT, a framework that jointly trains a pretrained ViT and a modified SAE end-to-end, directly aligning sparse patch-level representations with image classification. SpIn-ViT learns semantically coherent neuron activations that localize meaningful image regions while maintaining competitive predictive performance. We evaluate SpIn-ViT across nine image-classification benchmarks using classification accuracy, quantitative interpretability metrics, AI-based and Human evaluations. Compared with the previous state-of-the-art post-hoc SAE method, SpIn-ViT achieves 8.84% higher average classification accuracy, an AI-based interpretability score nearly four times as high, and a human-evaluation score more than twice as high. We further extract interpretable rule-sets using the SAE neurons to create neurosymbolic models which achieve 5.97% higher average classification accuracy while requiring a 58.8% smaller rule-set than the neurosymbolic models created from the SOTA post-hoc SAE method.

AI-generated music detectors are commonly evaluated against original songs, but real-world uploads are often remixed, re-encoded, pitch-shifted, or otherwise edited. These edited versions form a difficult negative class: they are not generated by AI, yet they may introduce spectral artifacts that resemble synthetic audio fingerprints. We study this problem as a hard-negative robustness setting for AI-generated music detection, focusing on AI-generated and edited variants derived from the same anchor songs. We compile a YouTube-based dataset of AI, edited, and original variants, using the original tracks only as references, and train a binary AI versus edited detector. Audio is processed as 10-second clips and passed as raw waveforms to a pretrained PaSST spectrogram transformer. To reduce leakage, all splits are performed by anchor song. On the held-out test set, the final video-level system achieves 0.811 balanced accuracy. At clip level, AI-generated clips reach an F1-score of 0.836, while edited clips reach a lower F1-score of 0.720. The results suggest that AI-generated music retains detectable fingerprint-like spectral cues beyond ordinary editing, but the lower edited-class F1-score shows that these cues can still overlap with artifacts from edited audio. Grad-CAM visualizations are used to inspect whether high-confidence predictions rely on localized time-frequency regions.

We introduce the Open-Strategy Dictator Game (OSDG), a variant of the classic dictator game in which each player's strategy is a natural-language document visible to all participants. The dictator's decision, to SHARE or TAKE an endowment, may depend on the text of the recipient's strategy. A large language model adjudicates each interaction by interpreting the dictator's strategy in the context of the recipient's. We run round-robin tournaments among diverse strategies and analyze the resulting payoff matrix using softmax equilibrium frequencies, dominance analysis, and sensitivity to the relative value of cooperation. Conditionally cooperative strategies, those that share with cooperators and take from exploiters, consistently dominate, while unconditional strategies (always share or always take) are weakly dominated. The results suggest that in environments where agents can inspect each other's decision procedures, conditional cooperation is evolutionarily robust across a wide range of payoff parameters.

Quantitative forecasts of frontier artificial intelligence often connect dated targets to trends in benchmark scores, training compute, release time, or expert belief. This paper audits whether the public measurement record supports those connections before another trend is fitted. I construct a frozen, event-centric record through 12 August 2026 with 62 selected systems, 12 versioned benchmarks, seven capability or impact criteria, 144 graded events, 27 source records, and 408 typed relations. The record is an audit sample, not a census. Only seven systems jointly observe estimated training compute and a METR 50 percent task horizon. Training compute is absent for 19 of 27 closed systems, including every selected closed release from 2026, while none of the 35 open-weight systems has a METR horizon observation. Benchmark succession creates a second break: a seven-system link from METR Time Horizon 1.0 to 1.1 has a log-scale slope of 1.206 (95 percent CI 1.021 to 1.390), whereas a six-system MMLU to MMLU-Pro comparison appears shift-like under logit and probit links but not under linear or logarithmic links. The observed bridges have about 80 percent power only for slope departures near 25 percent. Provenance is concentrated: 52 of 71 substantive quantitative events, or 73.2 percent, come from one measurement programme, and 76.1 percent are laboratory releases. A review of 56 methodological and empirical sources identifies 16 complementary measurement directions spanning resources, inference budgets, reliability, agentic work, safety, human preference, field outcomes, and forecast backtesting. No direction supplies a replacement scalar. The result is not that frontier AI forecasting is impossible, but that a defensible dated forecast is a claim about a versioned measurement system with explicit joins, protocols, links, and source dependence, not merely a fitted curve or calendar date.

AI co-scientists that generate hypotheses, retrieve related work, design experiments, execute code, and draft full papers are beginning to change how research is carried out. Despite this rapid progress, state-of-the-art systems remain researcher-agnostic: given a research goal, they optimize novelty, validity, or reviewer score while ignoring the individual scientist who will use the output. This overlooks a fundamental fact about research, namely, that what counts as novel, valuable, or feasible depends on the researcher, including their prior work, methodological repertoire, and the collaborators and communities in which they are embedded. In this work, we introduce the problem of personalized auto-research, which conditions every stage of the research process on a representation of the individual researcher. We argue that personalization is not a convenience layer, but rather the fundamental property that allows an AI system to serve as a genuine co-scientist rather than a generic instrument. To address this problem, we propose a general and flexible framework that threads a graph-grounded researcher context through retrieval, hypothesis search, experimentation, writing, and review. The framework consists of three fundamental components: (i) graph-grounded researcher representations, (ii) personalization across the full research pipeline, and (iii) evaluation grounded in the individual. Notably, we highlight a one-size-fits-all failure mode where distinct researchers issuing the same goal receive essentially the same research, erasing the tacit knowledge through which novel ideas arise. Finally, we discuss fundamental open problems and challenges.

Agentic coding assistants are finding widespread use, not just in new code development but in quickly ingesting and leveraging third-party code. This opens up a risk of malicious code being ingested as these coding tools operate with broad filesystem access inside developer workspaces. In this paper, we extensively study the impact of different dimensions of a novel attack surface we term workspace topology – defined via directory depth, codebase modularity, in-file injection position and context framing – on the attack success rate of adversarial prompt injection attempts. We perform an empirical study of indirect prompt injection (IPI) across a diverse set of open-source repositories spanning 10 languages and 6 engineering domains, evaluating three IPI entry points against open-weight models operating open source code harnesses. We find that workspace topology measurably affects IPI success. Specifically, changes in codebase modularity can significantly alter the Attack Success Rate (ASR), with highly modular environments demonstrating significantly lower attack success rates. Furthermore, context framing and introduction of security-cues in the workspace can also alter the ASR. Our findings offer practical value for the evaluation and security testing of coding agents across diverse settings, while underscoring the importance of an uncontaminated testing environment to obtain reliable results and conclusions.

Long-horizon agents can execute continuously, but human attention remains intermittent and scarce. This creates a bidirectional coordination problem: users may need immediate access to an agent while work continues in the background, whereas agents may encounter consequential decisions that require user judgment after the user has stopped monitoring execution. We posit an always-on attention-coordination layer—Jarvis\footnote{Named after the fictional AI assistant in Iron Man.}—that mediates this interface and allocates human attention across one or more working agents. We introduce JarvisBench to evaluate both directions of this coordination: whether an intermediary can accurately and promptly answer user-initiated questions about ongoing work, and whether it can recognize when an agent requires user judgment, solicit that judgment at the right moment, and route it back to improve task outcomes. JarvisBench contains 45 agentic task instances: 20 single-agent tasks and 25 workstreams organized into 10 multi-agent projects. The tasks span 19 domains and were selected and adapted from more than 2,000 public candidates. Crucially, the need for user attention arises naturally during execution rather than from an obvious omission in the initial prompt. JarvisBench is designed to integrate with arbitrary agent runtimes without modifying their underlying execution loops. Our reference implementation further provides a full-duplex speech interface, allowing users to reach Jarvis naturally while timely attention coordination supports agents working in the background. By separating agent execution from attention coordination, JarvisBench provides a stable evaluation target as agent capabilities continue to improve.

LLM-based coding agents have advanced rapidly on single-process SWE tasks, with frontier models now clustering in the high-70s on SWE-bench Verified. Distributed-system debugging, however, remains an under-explored regime: bugs span processes, nodes, and protocol interactions, with root causes rarely recoverable from source alone and brute-force exploration intractable across non-deterministic interleavings. This leaves two gaps in LLM and agent evaluation: no code-repair benchmark targets distributed-system bugs, and no controlled study isolates how much externally provided debugging context changes agent success on them. We introduce DDBench, a code-repair benchmark of 60 historical bugs mined from 13 open-source distributed systems, partitioned into three difficulty tiers. DDBench evaluates every case under two matched conditions: a symptom-only condition where the agent receives only the bug symptom and repository, and a context-augmented condition where it additionally receives a bounded debugging context (logs, traces, runtime state, and targeted code-investigation notes), isolating the effect of debugging context from model capability. The evaluation of ten LLMs on DDBench reveals several findings. First, distributed debugging exercises a reasoning dimension that single-process benchmarks do not surface: models' pass rates span 61 pp, and pairwise bootstrap separates 9 of 15 top-tier model pairs at p < 0.05 on DDBench's hardest case-set. Second, bounded debugging context lifts aggregate pass rate by +18.1 pp, and the lift is asymmetric: weaker models gain pass rate, while stronger models gain efficiency. Third, debugging context requires careful curation, as even faithful debugging context can sometimes mislead LLMs.

Learning and skill mastery require extensive and deliberate practice. In many learning settings, producing high-quality pedagogical materials can require a high level of domain expertise and be very time-consuming. Pedagogical materials often need to train students to engage in different thinking patterns. In some domains, such as chess, puzzles are used to help students practice their skills in calculating the next moves and recognizing known patterns on a board. Giving students a practice set of puzzles to help them learn different modes of thinking is challenging because the teacher needs to carefully balance between different motifs and how many look-ahead steps a student needs to perform. Popular online platforms like Chess.com and Lichess offer players millions of puzzles. Unlike chess tactics puzzles procured by human experts, where chess beginners can learn valuable insights, these puzzles are automatically generated and often regarded as having low pedagogical value. These platforms also rely on a heuristic to recommend puzzles to users for practice. Using the user history data over an entire year, a total of 1.5 billion puzzle-solving histories, we learn the pedagogical value of a puzzle and how to automatically choose a set of puzzles to better support chess learners using insights from offline reinforcement learning. We show that using offline policy evaluation, our trained policy has significant impact on beginners with puzzle-solving Elo range of 100–1000, particularly for the group of beginners whose learning growth was stagnant. We also performed a qualitative analysis of the puzzles discovered by our model by collecting annotation ratings from expert chess players. The success of our pipeline shows promise for a future where we can understand the pedagogical values of practice items given general user interaction data.

As authorship attribution systems are increasingly deployed to detect ghostwritten and AI-generated papers, their errors can support accusations against legitimate authors. These systems conflate stylistic similarity with individual identity. Researchers, however, study under advisors, and inherit their stylistic quirks. We build a corpus of arXiv authors with $\geq 2$ solo papers from the Mathematics Genealogy Project graph, giving $5{,}803$ total authors and $2{,}501$ ground-truth advisor-student pairings. Using embeddings from a fine-tuned model, advisors sit $39.9\%$ closer in cosine distance to their students than a random same-field author does. Using two open models, we reproduce the effect at $12.6\%$ and $14.5\%$. Academic siblings, two students of one advisor who may never have met, sit $30.4\%$ closer across $8{,}360$ pairs, even when they studied at different institutions. Pairs who share only institution and field show negligible similarity. Given a closed-set attribution task over the same corpus, the system's errors occur on the true author's advisor, student, academic sibling, or lab mate $11$ times more often than chance.

Long-document visual question answering (VQA) over documents of tens to hundreds of pages mixing text, tables, charts, and figures typically follows retrieve-then-read pipelines. In our setting, the bottleneck shifts from retrieval recall to reranker-side evidence selection: on MMLongBench-Doc, BGE-M3 reaches Recall@20 = 0.86 but only F1@5 = 0.254, and even the visual retriever ColPali reaches only F1@5 = 0.332; a text-only rerank LLM seeing only raw snippets misses table, chart, and layout evidence even when the upstream retriever encoded images. We propose Trident, with two complementary components: Trident-R, a retriever-agnostic LLM reranker that converts each candidate into an LLM-readable semantic record, including a visual caption, section path, entity tags, multi-axis concept hits, and a text snippet, then performs a single adaptive-K rerank call; and Trident-S, a generation-side module that prompts the VLM under topical, entity, and structural lenses before synthesis. On two long-document datasets, the annotation+rerank protocol substantially improves retrieval F1 across five heterogeneous pools, with every reranked pool exceeding the strongest adaptive-K baseline PageIndex. An LLM rerank without the annotation barely changes first-hit ranking, indicating the lift comes from the structured annotation. Trident-S targets open-ended synthesis questions by design, adding up to 6.6 points in generation accuracy on these questions. The best Trident configuration is the strongest downstream QA pipeline in our evaluation, with rankings consistent across two LLM judges (kappa = 0.913).

Aligning a language agent to several objectives at once is a persistent failure mode of preference-based training: when objectives are combined additively, optimization collapses onto whichever is cheapest to improve and sacrifices the rest, so a support agent learns to sound warm while giving no real help. The root issue is that an additive reward has no notion of balance. We introduce Mint (MIN-selection preference disTillation), a one-line change to preference distillation: rather than ranking sampled candidates by a weighted sum of rewards, we rank them by their weakest objective, distilling the best-balanced candidate over the most lopsided one with an unchanged DPO objective. This is the p -> negative infinity limit of a generalized-mean family spanning additive to worst-case selection. Across cooperative emotional support and adversarial negotiation, min-selection lifts both objectives while sharply cutting their imbalance; on emotional support it raises the weaker axis from 0.37 to 0.64 (p < 10^-40), surpassing human experts and persisting across full multi-turn rollouts. A turn-by-turn analysis yields our central finding: min-selection corrects imbalance in proportion to how imbalanced the reference policy is, and its benefit endures over an interaction precisely as long as that imbalance does.

Frontier LLM agents increasingly transact on behalf of separate principals, often using natural language rather than structured APIs. Much of the safety literature studies misaligned LLM behavior through adversarial-elicitation evaluations on single agents or stylized tasks. Its prevalence and structure in settings that combine long horizons, separate principals, real operational state, and inter-agent natural-language exchange remain insufficiently measured. We study 2,583 inter-agent emails from 20 one-year simulation runs of Vending-Bench Arena, a competitive vending environment spanning 13 frontier LLMs. We operationalize speech-act misalignment as emails containing false factual claims, manipulation, collusion, or threats, combining message content with ground-truth simulator state and logged reasoning traces to classify and validate such behavior. Under our primary classifier, 12.6% of emails are labeled misaligned; misalignment appears in all 20 runs and 74.7% of individual agent-runs. Both the magnitude and composition of this misalignment are preserved under repeated classification at different sampling temperatures and under full-pipeline replication with judges from two other frontier-model families. Misalignment is also reciprocal and stress-conditioned: receiving a misaligned email from a counterparty raises the odds of a misaligned reply by 1.65x, and low-inventory conditions raise them by 1.58x. Across tests of capability-asymmetric exploitation, we find no evidence that higher-capability models differentially exploit weaker counterparties, and model performance rank does not predict misalignment rates. Together, these results indicate that measurable, state-dependent misalignment can arise in competitive multi-agent environments without engineered elicitation, in patterns associated with operational scarcity and counterparty behavior rather than model capability alone.

Handover (HO) management in vehicular networks requires fast and reliable decision-making under highly dynamic conditions. While machine learning (ML) approaches can improve HO detection by capturing complex relationships among various key performance indicators (KPIs), their black-box nature limits interpretability and operator trust. To address this, this paper investigates HO detection from an explainability-on-the-fly perspective using inherently interpretable models based on the functional analysis of variance (fANOVA) framework. The proposed models are evaluated using two real-world operator datasets and compared against a Long Short-Term Memory baseline augmented with post-hoc SHAP explanations. Unlike post-hoc approaches, the proposed framework enables immediate interpretation of model decisions without incurring additional computational overhead. This capability is particularly critical for latency-sensitive vehicular networks. The results show that fANOVA-based models achieve competitive detection performance while providing significantly reduced explanation latency compared to conventional post-hoc methods. Furthermore, feature ranking and visualization analyses reveal physically meaningful relationships between KPIs and HO occurrences that align with standardized HO mechanisms. These results demonstrate that inherently interpretable models provide an efficient and transparent solution for HO detection in next-generation vehicular networks.

When a user question is underspecified, a capable model should recognize that its context is insufficient, identify the missing information, ask for it, and respond only once that information determines a unique answer. We formalize multi-turn information seeking as solving a k-underspecified constraint satisfaction problem, where k is the number of variables jointly required to determine the target and therefore measures the degree of missing information. We instantiate the formulation in MT-InfoSeek, a controlled evaluation suite of 5,251 problems and 9,006 task instances spanning mathematics, logic, biology, medicine, and general knowledge. We evaluate models along three axes: what they ask, when they ask it, and how the acquired information affects the final answer. Performance degrades across models and domains as underspecification increases. Models recognize that additional information is needed but underestimate how much, and in logical problems at k = 2 they under-predict the degree of missing information about four times as often as they over-predict it. They also fail to identify a minimal sufficient set of queries, improve only marginally when given the true k, and often stop before acquiring sufficient information. In tasks with ordered dependencies, an incorrect query order reduces final accuracy even when the model eventually acquires all necessary information. We measure information seeking directly through final sufficiency, which records whether the acquired information determines the target independent of answer generation. This separation shows differences between models that final accuracy alone does not capture, and indicates that the ability to seek information over multiple turns is distinct from the ability to generate answers and is not measured by current LLM evaluations.

Large language models (LLMs) have become the dominant interface of clinical artificial intelligence, yet the interface they expose (text in, text out, one context window at a time) maintains no explicit, persistent, governed representation of what is currently true about a patient. This paper argues that longitudinal clinical reasoning is a state-estimation problem under partial observability, and that the axis on which clinical AI succeeds or fails is not the fluency of the model reading the record but the governance of the patient state it reasons over. We distinguish generated context from governed state; separate five objects that clinical AI habitually conflates (true state, observations, evidence, belief, and simulated state); define a tiered governance standard against which any clinical AI system can be audited; and show that an operational definition of accountability decomposes into four information requirements: an immutable evidence ledger with awareness-time versioning, a belief state distinct from accumulated evidence, an observation-process model, and claim-level causal typing. We are explicit that this decomposition is analytic rather than a necessity theorem, and that its value is conceptual hygiene: it converts "accountable clinical AI" from a slogan into an audit instrument. A six-level maturity framework separates what a system makes governable from what it can compute, locating current LLM-centric practice at high capability but low maturity. The paper is fully self-contained: the four research questions the framework poses are stated in the introduction, and the conclusion records what the paper establishes toward each; future work develops the buildable core of the architecture and the research program toward full Clinical World Models. No empirical result is claimed here.

An AI that can only give advice seems safe: the human is always free to ignore it. That is the premise of the boxing tradition in AI safety, and its long-suspected weak point is that the human who reads the answers is part of the system. We make the fraction $\varepsilon_t$ of behavior that follows the advice a state of a Markov decision process, moved by the advisor's own messages, so that use deepens reliance. Granted a channel rich enough to echo any action the human could take, higher $\varepsilon_t$ weakly lowers every monotone measure of the power of a human with a message-independent fallback. An oracle rewarded by per-round approval cultivates reliance beyond a closed-form patience threshold, so the same reward weights leave the optimal oracle answering in episodic deployments and cultivating in long-memory ones. An influence bound certified once at deployment is blind to that horizon and bounds the loss no lower than its trivial ceiling. An exogenous cap on influence bounds the guarantee the human loses, and a short enough memory reset removes the incentive to cultivate, while neither recovers the value already steered away. In a closed-form example the optimal oracle never cultivates in fifteen-round sessions and does in sixteen.

Objectives: To determine whether zero-shot prompting of a large language model (LLM) is sufficient to detect shared decision-making (SDM) behaviors in real clinical encounters, and whether supervised learning adds value under patient-grouped, nested evaluation. Methods: We analyzed 21 audio-recorded outpatient surgical decision encounters (19 unique patients; 7,566 utterance segments; ~6.1 hours) between families of children with multiple long-term conditions and their surgical providers. Trained coders labeled segments for 12 SDM behaviors (human-human macro Cohen's kappa = 0.695). We compared a zero-shot local LLM (Qwen 2.5 32B), a supervised classifier over frozen sentence embeddings, and their logistic stack, under patient-grouped outer folds with inner cross-fitted thresholds and patient-resampled confidence intervals. Results: The zero-shot LLM reached macro kappa = 0.139 (95% CI 0.111-0.164). The supervised classifier reached kappa = 0.227 (0.186-0.262), a paired improvement of 0.088 (0.051-0.119). A logistic stack of the two reached kappa = 0.242 (0.198-0.284). We identified multiple corpus-specific leakage paths, including grouping sibling recordings separately and allowing labels from an outer held-out patient to enter few-shot exemplars used while fitting downstream models. Conclusion: Zero-shot prompting alone is not sufficient to measure SDM behavior as reliably as a small supervised model, and patient-level grouping alone does not prevent leakage when labeled prompt exemplars are precomputed outside the outer evaluation loop. Reported performance is sensitive to the unit of data splitting and to where labeled exemplars enter the pipeline. External validation is needed before these findings generalize beyond this population, model, prompt, and codebook.

Abductive reasoning, often characterized as inference to the best explanation, is central to explanation under uncertainty, from everyday sense-making and investigation to scientific discovery. Yet LLM research has mostly studied abduction through narrow, task-specific benchmarks, making it unclear whether observed gains transfer beyond the benchmark family used for training or evaluation. We ask whether RL post-training can improve abduction as a transferable reasoning capability. We introduce CEDAR-GRPO, a process-aware framework that combines final-answer correctness with abductive rewards for evidence coverage and evidence-to-explanation directionality. Four open-weight LLMs are post-trained on a controlled, domain-neutral mixture of abductive hypothesis-generation and hypothesis-selection tasks. We evaluate them on 11 unseen tasks spanning hypothesis selection, missing-fact generation, defeasible inference, long-context investigation, clinical reasoning, code debugging, and non-abductive controls. CEDAR- GRPO improves every model on every held-out task over both base models and correctness-only GRPO, with average gains of 7.4 and 2.7 points, respectively, and a maximum gain of 30.8 points. Ablations confirm that RL, abductive reward design, and task diversity each contribute to transfer. Process-level metrics further show stronger abductive behavior, including exploration of alternatives, elimination of rivals, backtracking, and uncertainty marking.

Large low-Earth-orbit (LEO) Earth-observation (EO) constellations offer frequent access to geographically dispersed ground targets, but emergency requests may arrive after committed routine-plan execution has begun. The resulting dynamic emergency observation scheduling problem (DEOSP) requires urgent tasks to be inserted under intermittent ground contact without excessive routine-plan disruption. To address DEOSP, we propose a task-driven three-layer distributed scheduling (T3L-DS) method, which represents task demand and sensor footprints on a common geographic grid and forms temporary clusters from observation capabilities and current inter-satellite links. For intra-cluster coordination, T3L-DS introduces onboard dual-plan bidding and joint marginal evaluation. It also designs an inter-cluster coordination mechanism for unresolved demand. Extensive computational experiments compare T3L-DS with centralised simulated annealing (SA), an adapted selective time-variant better reply process (A-SeTVBRP), and a conventional contract-net protocol (CNP). T3L-DS achieves the highest emergency coverage among the distributed methods, with average relative improvements of approximately 2.8% and 17.1% over A-SeTVBRP and CNP, respectively. Its average relative gap from SA is approximately 7.1%. Under conflict-enhanced loads, it reduces routine-coverage loss by approximately 57.9% and 87.7% relative to A-SeTVBRP and CNP, respectively. The ablation study confirms the contribution of the proposed coordination enhancements. Overall, the results show that T3L-DS provides an effective distributed approach to DEOSP.

The efficient-KAN literature—covering Chebyshev, wavelet, and radial-basis-function variants of the original Kolmogorov-Arnold Network—has been benchmarked almost entirely on clean data. We show that this choice conceals a large capability difference between architectures: ChebyKAN's test MSE (evaluated against clean ground truth) increases by a factor of 10.6x when training data is corrupted with sigma=0.1 noise, versus 7.9x for vanilla KAN, 1.7x for a standard MLP, and just 1.4x for our proposed ER-KAN. ER-KAN combines three design choices targeting the noisy, data-scarce setting: shared Gaussian RBF bases across all edges in a layer (providing locality and efficient parameterisation), curriculum noise injection during training (explicitly teaching noise robustness), and entropy-weighted adaptive regularisation (preventing overfitting at small N). The result is a 595-parameter network that matches MLP accuracy at moderate noise while degrading far more gracefully as noise grows. We evaluate on eight analytic functions (N in {50, 200, 500}, sigma in {0, 0.03, 0.1}), on a damped harmonic oscillator physics-informed neural network where ER-KAN achieves 4.2x lower solution MSE than MLP, and on a Burgers' equation PINN where all models fail to converge—a genuine limitation we report rather than suppress. We introduce the noise degradation ratio as a simple complementary metric and recommend it become a standard reporting requirement for efficient-KAN papers.

Making language models solve constraint problems reliably often means having them translate the problem into a formal specification and delegating the search to a sound solver. But the translation is itself a language-model task, and an unfaithful translation makes the solver faithfully solve the wrong problem. Existing pipelines repair only translations that crash, returning the solver's error message and falling silent when the program runs but is wrong. We replace the error message with a proof: when the generated program is unsatisfiable, we extract a minimal unsatisfiable core over the model's own constraints and hand it back the exact set that cannot hold together, a leakage-free signal that localizes the fault. On a new benchmark of 77 problems with an exact oracle, translation to Answer Set Programming is faithful on six of seven domains and fails only on aggregate coverage scheduling, which concentrates the translation tax in one diagnosable pattern. A minimal core, rather than a bare error, is what stops a weaker model from fabricating solutions to infeasible problems, cutting fabrication from 79% to 7%. A strong chain-of-thought baseline meanwhile matches the symbolic route on accuracy, so the route's value is not accuracy but certificates and its refusal to fabricate.

An estimated 138 million children remain in child labour worldwide, and the monitoring systems used by affected sectors, built on periodic household visits and interviews, systematically under-detect them. We present a real-time computer-vision pipeline, built and operated solely as a research prototype, that studies the feasibility of giving Child Labour Monitoring and Remediation Systems (CLMRS) a continuous, presence-based evidence channel. The pipeline combines a multi-task person and face detector (YOLO26x backbone in the CerberusDet framework), cascaded age estimation pairing MiVOLO v2 with a child-specialist model for ages 0-12, ByteTrack tracking, ArcFace and DINOv2 re-identification, and track-level fusion producing reviewable per-person records. The detector raises person mAP@0.5 from 0.390 to 0.683 over the previous-generation baseline; the child specialist reaches 1.944 years MAE on children-only validation, where widely used open-source stacks err by 18-23 years. FP8 TensorRT compilation yields a 1.77x speedup at +0.002 years MAE, bringing the pipeline above twice real-time on embedded hardware. On 26.8 hours of proxy video the system finds 634 unique child candidates versus 285 for its predecessor. We further report a seventeen-day unattended field pilot on a farm in Zimbabwe (38.7 million frames, six cameras) evaluated against a daily attendance register: software tuning improved detection yield 36-fold, and identity consolidation under a simultaneity veto cut over-reporting from 9.1x to 1.8-3.9x with zero proven-false merges. We document training and quantisation failures alongside successes, and the data-protection and human-in-the-loop safeguards such a system requires.

Automated vehicles must explain their decisions in ways that passengers can understand, monitor, and trust. Existing language-annotated driving datasets are mostly observer-written, post-hoc, simulation-based, or generated from sensor inputs, rather than elicited from the driver performing the action. We introduce NARRATE, a multimodal real-world Australian driving dataset comprising 2,050 annotated events from 35 experienced drivers and driving instructors on public roads. Each event is grounded in synchronised visual, localisation, motion, and LiDAR streams and paired with in-vehicle and/or post-drive free-text explanations. NARRATE provides action labels, scenario-context labels spanning six high-level and 32 fine-grained categories, and span-level Situational Awareness (SA) annotations over driver explanations for Perception, Comprehension and Projection. Four benchmark tasks (SA, scenario-context, driver-action classification, and explanation generation) show that this structure is learnable from driver language, while fine-grained context recognition and explanation generation remain challenging. NARRATE paves a path towards more human-centred and domain-aware explanation models for automated driving.

Data cleaning without a trusted clean reference is challenging because unusual values may represent either genuine errors or valid observations. This paper studies how different agent capabilities affect reference-free data cleaning and proposes an evidence-grounded framework that combines structured context, profiling, LLM reasoning, executable checks, controlled evidence retrieval, source ranking, citation alignment, conservative repair, reversible scripts, and provenance logging. Seven configurations are evaluated across financial, clinical, and environmental-monitoring datasets using controlled synthetic corruption and original-data descriptive analysis, resulting in 126 completed runs. The evaluation includes two comparison baselines and a progressive LLM-based sequence that adds executable tools, evidence retrieval, evidence controls, and conservative repair. In the synthetic evaluation, the deterministic profiling baseline achieved the highest detection F1-score of 0.561. Among the LLM-based configurations, the full conservative configuration achieved the highest F1-score of 0.421, but no configuration performed best across all evaluation criteria. The source-ranked configurations achieved the lowest unsupported-rule rates, while decision-level citation alignment remained weak. The full conservative configuration produced no unsafe or unnecessary modifications, although these rates were already zero before the conservative policy was added, and it performed no direct repairs. Overall, the results show that additional capabilities introduce trade-offs among detection, repair, evidence grounding, conservative behaviour, reproducibility, and operational cost rather than producing consistent improvements. The study provides a structured framework and empirical methodology for evaluating these trade-offs in reference-free agentic data cleaning.

Assessment of ventriculomegaly (VM) on fetal brain ultrasound relies primarily on measuring lateral ventricular atrial width on standard planes, which is operator-dependent and may not fully reflect the overall ventricular enlargement. Fetal brain MRI provides more reliable volumetric information but is costly and less accessible for routine use. To address these limitations, we propose VIFBA, an ultrasound video-based framework for fetal brain assessment that predicts MRI-derived lateral ventricular volume, classifies VM severity, and identifies potential non-VM fetal brain abnormalities. Our contribution is three-fold. First, we introduce a joint-embedding predictive architecture (JEPA)-inspired tube latent prediction objective that leverages spatio-temporal coherence in ultrasound videos to enhance representation learning. Second, we develop a contrastive cross-modal alignment strategy that transfers structural information from MRI to ultrasound during training, while requiring ultrasound alone at inference. Third, we augment VIFBA with a training-free vision-language model and retrieval augmentation to verify uncertain predictions and identify potential non-VM fetal brain abnormalities. We validated VIFBA on a large dataset comprising 857 cases (3,196 videos) with paired fetal brain ultrasound and MRI examinations. On held-out test data, VIFBA achieved an MAE of 0.5909 mL and Pearson correlation coefficient of 0.9907 for ventricular volume regression, 0.9400 accuracy for VM severity classification, and an F1 score of 0.7764 for multi-abnormality classification, substantially outperforming single-task baselines, video-based strong competitors, and state-of-the-art foundation models. By enabling MRI-informed volumetric assessment from routine ultrasound alone, VIFBA offers a practical and potentially broadly deployable pathway toward accurate and affordable prenatal brain screening.

For patients with Moyamoya disease, impaired cerebrovascular reserve (CVR) is an important hemodynamic criterion for recommending extracranial-to-intracranial bypass surgery. Standard CVR assessment in this cohort uses paired arterial spin labeling (ASL) perfusion MRI acquired before and after acetazolamide (ACZ). When ACZ is contraindicated or avoided, the post-ACZ cerebral blood flow (CBF) map needed for hemodynamic assessment is unavailable. We propose CAE3D, a deterministic 3D conditional autoencoder that synthesizes post-ACZ CBF maps directly from pre-ACZ ASL input. We evaluated CAE3D against ten comparators, including deterministic and diffusion-style 3D baselines, a 2D contextual baseline, and frozen-encoder foundation-model adapters. CAE3D achieved the lowest held-out MAE (0.066), with SSIM 0.80 and PSNR 24.0 dB, and near-zero full-brain mean bias. Its MAE advantage was statistically significant over seven of eight trained-from-scratch baselines, excluding the 2D CAE_2D comparator; its SSIM and PSNR advantages were significant over all eight. Regional delta-CBF predictions compressed the dynamic range in high-response territories. These results establish the retrospective feasibility of post-ACZ CBF synthesis in patients who completed the standard two-scan protocol. Extension to ACZ-contraindicated patients, who were not represented in this cohort, requires external and prospective validation.

The aviation industry characterized by its stringent safety standards has seen a growing need for innovative approaches to enhance safety measures. Despite the vast accumulation of aviation safety data over time, its full potential in predicting and preventing incidents has not been fully realized. This research addresses this gap by applying machine learning (ML) and natural language processing (NLP) techniques to analyze aviation safety data from Socrata, the Australian Transport Safety Bureau (ATSB), the National Transportation Safety Board (NTSB), and the Aviation Safety Network (ASN). By leveraging existing ML models, including deep learning and transformer-based architectures alongside NLP methods for mining aviation incident narratives, this study uncovers patterns contributing to safety related incidents such as accidents and near-misses. Additionally, it employs various topic modelling techniques to extract meaningful themes from unstructured safety reports, enhancing the interpretability of incident analysis. Causal inference techniques and interpretable AI frameworks are further explored to improve model transparency and trustworthiness. A key contribution of this work is the deployment of advanced ML methodologies in a structured aviation safety context, assessing their effectiveness and providing insights into their practical implementation. The findings offer valuable insights for aviation stakeholders, including regulators, airlines, and policymakers, by providing data-driven solutions that enhance incident analysis and decision making. Ultimately, this research supports the industry s ongoing efforts to minimize risks, improve passenger and crew security, and integrate AI driven methodologies into aviation safety management.

Determining the biological sex of the individuals who created Upper Paleolithic hand stencils remains a challenging problem due to the absence of ground truth, population differences between contemporary and prehistoric groups, and the uncertainty introduced by image degradation. Traditional morphometric methods suffer from high structural overlap across sexes, poor cross-population generalizability, and subjective feature engineering. This study presents an uncertainty-aware deep learning framework for sex attribution in prehistoric hand stencils that explicitly models, propagates, and aggregates uncertainty throughout the analytical pipeline. The methodology combines dual image processing, dual contour extraction, structured silhouette augmentation, model architectural diversity, and ensemble-based decision aggregation. The pipeline generates twelve plausible silhouette realizations per stencil to capture boundary uncertainties, which are processed by two ensembles of ten deep neural networks each (EfficientNet-B3 and MobileViT-S) trained on 14,036 contemporary hand samples. Furthermore, a triangulated validation scheme integrates ensemble predictions with unsupervised 2D latent-space manifold mapping (UMAP + k-NN) and explainable AI spatial attributions (LayerCAM) to ensure anatomical consistency. On contemporary data, ensemble models achieve strong classification performance, with accuracies exceeding 88% in older age groups. When applied to prehistoric stencils, the framework produces both sex predictions and confidence measures of internal agreement, enabling the distinction between morphologically stable and ambiguous cases. Convergence across ensemble predictions, latent-space structure, and interpretability analyses shows that uncertainty can become a measurable component of archaeological inference, enabling robust and reproducible decoding of ancient rock art.

Interactive game world models typically autoregress visual observations directly in pixel or latent space, forcing structured properties such as pose, geometry, and occlusion to be implicitly maintained by the same generative sequence. Over long horizons, errors in these latent world properties accumulate, making consistency and controllability fragile. We explicitly model the evolving world state, delegate exact geometric computation to a fixed, zero-parameter renderer, and leave the neural model to synthesize appearance. We instantiate this idea as Marionette, a world model for interactive games with articulated characters. First, a two-stage autoregressive dynamics model predicts an explicit and interpretable 276-dimensional 3D world state comprising multi-entity articulated skeletons, metric root trajectories, and rotations. Second, a zero-parameter graphics bridge converts the predicted state into pose-control videos, computing world-space geometry and occlusion in closed form. Third, a control-conditioned video-diffusion observation model synthesizes photorealistic RGB observations from the resulting structured controls. Our experiments establish two properties of Marionette. First, the predicted world state is directly controllable. Forcing a mismatched action stream changes root-aligned joint error by 31% across 48 held-out segments. Second, long-horizon behaviour is determined in the state, and can be repaired there. Left free, the two generated characters drift to 21.2 m apart (recorded sessions stay near 5 m) and a third of frames show ground penetration. Two rules imposed on the explicit state, a terrain collider and a separation cap, cut penetration by 66% and keep the pair engaged, with no change to the observation model. Routing appearance through the predicted state costs no fidelity we can detect, at an FVD of 831 against 799 for recorded pose.