Papers

Task handover requires communicating enough current state for a successor to resume work, yet the relevant information is often divided between system records and human observations. System records can be precise and timestamped but only partially observe the task, while human reports capture intent and task knowledge that no log contains but are vulnerable to omission and memory error. We present a provenance-aware pipeline that converts task telemetry and human-authored reports into a shared typed task-state representation, aligns and reconciles their facts, detects conflicts, and generates structured handover reports. We evaluate the approach on 13 paired task states collected in a controlled spatial multitask environment, using task-grounded metrics that estimate the state-reconstruction cost a report would spare a hypothetical recipient and the misinformation burden it would impose. Reconciling both sources preserved greater estimated task-state utility than either the user report or telemetry alone. Relative to a direct end-to-end LLM given the same inputs, structured reconciliation maintained comparable estimated utility while incurring substantially less misinformation, and task-aware rendering retained utility more efficiently (per token) than exhaustive rendering. An exploratory content analysis further shows that human reports contain substantial strategic knowledge that lies outside state-focused metrics. These results support provenance-aware state reconciliation as a design pattern for safer AI-assisted handover.

What does a generation loop gain from learning on its own verified successes? In cycles of generate, verify, select and LoRA-consolidate on online bin packing, training on value-filtered candidates shifts what the model writes on held-out variants toward value (-1.7 points of excess, p=0.008; -3.1 against a random-consolidation control, p=0.004) while the best observed candidate converges to the classic heuristic's level and no further. A confirmation battery replicates the whole procedure three times, with fresh seeds and a never-consulted held-out set read exactly once: the mean was nearly identical in all three lineages (-2.0, -1.8, -1.9), and after aggregating within held-out variant all seven evaluable variants favored consolidation (p=0.008). The best observed candidate moved to the classic heuristic's level, exactly (0.021028, the lowest variant's classic level, in all three lineages and both arms); attract never went beyond its variant's classic level, while the random control did in 7 of 24 lineage-variant cells. A matched SFT-only control shows the supervised anchor, not repulsion from bad candidates, does the concentrating (96% of candidates land exactly at the classic heuristic's level). On far families the tails cut both ways: consolidation lowers the per-candidate rate of better-than-classic candidates (10% to 3.9%) while its larger production yields more such candidates absolutely (5 against 1, on few events). As motivation we report the inference-time ledger that led here: a model-written schematic recap buys judged document integration and nothing buys development; a verifier written into the stream is imitated, 16.4 fabricated verdict lines per notebook. In this arrangement (an 8-bit model in MLX, five cycles), mean quality among valid candidates can be bought and replicated; under attraction the observed best goes to the classic and, so far, never beyond it.

We introduce MineCEraft (Minecraft Construction Engineering Benchmark, pronounced mine-see-ee-raft), an easy-to-use, open-source benchmark designed to systematically evaluate the reliability and limitations of LLMs for construction tasks in Minecraft. The MineCEraft benchmark comprises 723 domain-expert hand-crafted natural-language instructions with programmatically verifiable evaluation, spanning 17 distinct task categories, providing a safe and controllable experimental environment for assessing LLMs' ability to perform realistic construction engineering tasks. With this benchmark, we conduct an in-depth evaluation of state-of-the-art LLMs and perform a detailed error analysis, revealing key failure modes and practical challenges in applying LLMs to construction engineering tasks.

This paper develops a hybrid offline-online multi-agent reinforcement learning framework based on decision transformers. The policy is first pretrained offline via supervised sequence modeling of trajectories generated by existing policies, providing a safe and sample-efficient initialization. It is then fine-tuned online using a hybrid objective that incorporates critic-guided gradients, enabling performance improvements beyond the offline policy. To facilitate stable offline-to-online transfer and effective multi-agent coordination, the framework incorporates return-weighted sampling, a critic conditioned on neighbors' actions, and neighborhood-correlated exploration. The approach is fully distributed: both training and execution rely only on local observations and limited information exchange among neighboring agents. Evaluations with dynamic traffic arrivals in two settings: (i) joint scheduling and power allocation and (ii) coordinated beamforming, show that the proposed method achieves quality-of-service (QoS) performance comparable to centralized methods. Moreover, when pretrained on lower-quality datasets, online fine-tuning is also observed to surpass the initial offline policy. These results demonstrate a promising learning-based alternative for wireless resource management.

Supplying context at inference time to a large multimodal model is an inexpensive lever for adapting speech transcription to a domain, and earlier results on smaller models reported large gains. This work tested that mechanism where it ships, in the prompt-conditioning layer of a production oral-history transcription tool, on a sample from its own production corpus. Full prompt-level context did not detectably change side-level word error rate (WER), and none of the four preregistered hypotheses was supported. The design was a within-item paired ablation, preregistered with the analysis code frozen by hash before the confirmatory batch was scored; two disclosed gpt-4o pilot sides had been scored earlier, during scorer development. Nineteen cassette sides, about 10.6 hours of degraded 1970s-80s interview audio, were reprocessed through the production code path under three prompt arms, crossed with two deployed commercial configurations, gpt-4o-transcribe and gemini-2.5-flash, and scored against operator-corrected verbatim references. For gpt-4o-transcribe the median paired difference between the full-context and no-context arms was +0.6 WER points, with a side-resampled interval of [-1.1, +1.0]; the Gemini estimates were too unstable to support a comparable negative inference. A post-hoc rerun found run-to-run pipeline variability larger than the confirmatory differences, so effects of that size cannot be resolved from one transcription per cell. An implementation audit verified the manipulation was live, and sequence-alignment analysis found a small improvement on complete context-listed phrases, too small to materially change side-level WER, and for Gemini coexisting with worsened unlisted-token error. Evaluating context mechanisms therefore requires sequence-aligned term-level, insertion, and speaker-label measures alongside aggregate accuracy.

Reasoning models do not stop when they know the answer. On DeepSeek-R1-Distill-Qwen-7B the chain of thought runs about twice as long as the model's own answer probability takes to settle, and how much of that excess is removable varies from problem to problem, so a global length penalty cannot take it out. We take it out by internalizing a causal interpretability finding into the weights. The mechanism is a halt vector: a difference-of-means direction at layer 18 of this model whose steering strength controls how long it thinks, while a replicated value axis does nothing. Installing that intervention in the weights is harder than it looks. Maximizing the scalar projection onto the direction corrupts the off-axis dimensions a frozen downstream reader depends on, and generation gets longer instead of shorter; what works is reconstructing the whole steered activation with those dimensions pinned to their natural values. Fit from 24 problems and no reinforcement learning, the halt removes about a quarter of the thinking at held accuracy across five unseen benchmarks, and the cut tracks each problem's own removable slack at 0.70. It also closes a non-termination pathology that grows with difficulty and that a decoding-time confidence hook makes worse. We do not claim to beat a well-tuned length penalty or decoding-time early exit on the raw trade-off; the contribution is how the halt is obtained.

Equivariant graph neural networks provide a principled way to model geometric systems, but efficient first-order architectures remain limited in how vector information can be transformed as it moves across a graph. We introduce ESNN, an Equivariant Sheaf Neural Network that enriches this interaction by learning directed, matrix-valued transport between neighboring vector features while preserving exact Euclidean equivariance. Rather than increasing the order of the representation, ESNN keeps scalar and vector features first-order and places the additional geometric flexibility in the edge transport itself. We characterize this transport theoretically, showing that when relative displacement is the only covariant geometric input, every linear $O(n)$-equivariant map decomposes into independent radial and tangential components, while learned covariant features enable richer feature-conditioned transformations. We also introduce controlled symmetry relaxation for systems with a preferred ambient direction, which may be prescribed or inferred from data while recovering full $E(n)$-equivariance when the directional pathway is inactive. Across particle dynamics, mesh-based simulation, point-cloud classification, and molecular property prediction, ESNN improves dynamics prediction, recovers the gravity axis when symmetry is broken, yields substantial gains on selected mesh tasks and long-horizon rollouts, and remains robust to unseen rotations. These results show that learning how geometric information is transported across edges offers a complementary route to expressive equivariant message passing without requiring higher-order representations.

Layer-skipping methods for efficient LLM inference decide, at some granularity, which transformer layers to execute for a given input. We present a rigor-matched, three-seed audit of two periodic-step, search-based methods that make this decision online at inference time and re-evaluate it every few generation steps: a confidence-gated early-exit baseline (ConfLayers) and genuine self-speculative decoding (SWIFT, Xia et al. 2024), together with vanilla autoregressive decoding, across Qwen2.5-0.5B and Qwen2.5-1.5B (Yang et al. 2024) on GSM8K (Cobbe et al. 2021) and CNN/DailyMail (Nallapati et al. 2016; See et al. 2017). SWIFT is strongest on accuracy in three of four cells, while ConfLayers is dominated everywhere, with particularly large deficits on GSM8K at 1.5B. Separating online-search overhead from pure inference cost, we find that SWIFT's true inference speed is faster than ConfLayers's in all four cells (5-21%), reversing the naive wall-clock ranking in three; ConfLayers's search overhead is small and stable (1-2% of cost), whereas SWIFT's is larger and more variable across seeds (up to 28.7%). We additionally examine two trained routing methods, LayerRoute (Sikdar, 2026), a per-sequence, input-conditioned hard gate, and LayerDrop (Fan et al. 2020), a fixed, input-independent pruning pattern, as a supplemental analysis rather than a head-to-head comparison because both operate at a coarser decision granularity. Under a verified protocol with genuine per-input gating, a genuine full-model baseline, and genuine inference-time compute skipping, both show modest real speedups (1.08-1.33x) but accuracy well below the periodic-step methods, including near-total LayerRoute collapse on GSM8K at 1.5B (0.003 mean exact-match across three seeds). We release the full audit protocol as a template for rigor-matched efficiency comparisons.

Ensuring a safe virtual reality (VR) experience requires systems that can predict and respond when users lose their balance. Although prior work has examined fall prediction and motion sickness, many approaches are regression-based and postural state classification remains less explored. This study compares machine learning (ML) and deep learning (DL) models for classifying postural states in VR under visual perturbations. We used a multimodal dataset containing kinematic, electromyographic (EMG), and electrodermal activity (EDA) signals. The data were prepared for a binary task to distinguish balanced from imbalanced postural states, and participant-wise downsampling addressed class imbalance. All models were evaluated with Leave-One-Participant-Out (LOPO) cross-validation to test generalization to unseen participants. Among the models, the Mamba-inspired CNN (MI-CNN) achieved the highest accuracy of 96.76%. SHapley Additive exPlanations (SHAP) analysis improved interpretability and identified the most influential classification factors. The SHAP results showed that kinematic features were dominant, indicating that body-motion patterns are informative for detecting imbalance in VR. We also evaluated MI-CNN using only the top two-thirds of features ranked by SHAP importance. Despite a 33% reduction in input dimensionality, the model maintained performance, achieving 0.957 accuracy and 0.957 F1-score, with about a 1% decrease compared with the full-feature model. These findings suggest that multimodal sensing, temporal deep learning, and explainable AI can support reliable classification of balance-related instability in VR. Accurate recognition of imbalanced postural states may raise awareness of fall risk and guide safer, adaptive VR systems that respond to instability while improving user safety and experience. Code is available at: https://github.com/NipaAnjum/MI-CNN.

We show that smooth two-layer feed-forward networks (FFNs) expose an additional structural model extraction channel under a chosen-input raw-output oracle at the FFN branch; consider transformer FFN branches with GELU or SiLU activations under chosen-input raw-output access, without access to parameters, gradients, or internal activations; exploit a second-order leakage channel in which projected input Hessians form different mixtures of the same hidden symmetric rank-one factors induced by the FFN input weights. We formalize resulting Hessian collection as a partially symmetric decomposition to establish conditions for local identifiability and stability to exploit vector-output stencil reuse to reduce the structural query cost by a factor of 16. On independently trained CIFAR-10 vision transformers, only 16 projected Hessians, corresponding to 8193 black-box queries, recover the hidden FFN directions with average absolute cosine alignment above 0.91, with 86.7.1 % of GELU and 91.1 % of SiLU directions exceeding 0.90 alignment. Recovery remains high across independently trained models, repeated extraction runs, and all transformer blocks. The recovered structure supports functional extraction too. Keeping the recovered directions fixed and fitting only the remaining FFN parameters yields high-fidelity substitutes with more than 92 % top-1 agreement, while test accuracy remains within 1.24% and 0.57% of the GELU and SiLU targets. Output rounding and Gaussian noise substantially reduce recovery under a fixed attack configuration, but adapting the finite-difference step restores average alignment to 0.9188 and 0.9081. This is an end-to-end path from black-box second-order observations to hidden FFN-structure recovery and functional replacement. Under the stated oracle model, smooth FFN curvature exposes internal parameter geometry that behavioral fidelity alone cannot reveal.

We conducted an interview study with twelve students on their use of generative AI in academic communication. Students delegated professional messages to AI most where the pressure to sound professional is highest: email to instructors and administrators. AI involvement ranged from correcting the writer's own text to working out and writing the message outright, and students checked AI-written text against two criteria: whether it looks like AI and whether it sounds like them. Building on these findings, we model the AI-mediated process of writing a student–instructor email at the highest level of involvement we observed, and compare it with an unaided model of writing the same messages, built from participants' accounts and a classic model of the writing process. Three differences emerge: the learning loop that builds writing skill is removed, the message is no longer written for its specific recipient, and the confidence a successful exchange returns goes to using the system rather than to the writer's own ability. From these differences we derive two risks, that individual capacities never form and that authenticity and trust in communication become work. Design can respond to both but is unlikely to be enough, so the risks also need research and policy attention.

While Large language models (LLMs) incorporate user personalization signals to improve usability and helpfulness, they increasingly shift from providing balanced, informative responses toward optimizing for user satisfaction when conditioned on personal context such as conversation history, inferred preferences, and user profiles. Specifically, we identify three emerging risks: (1) irrelevant personalization, where models reference personal information in unnecessary contexts; (2) preference narrowing, where models reinforce informational echo chambers; and (3) sycophantic bias, where models agree excessively with user opinions. As a result, models may reference personal information in contexts where it is unnecessary, inadvertently collapse response diversity, or agree excessively with user opinions. Despite the growing use of personalization in AI assistants, there has been limited systematic evaluation of its potential side effects. To bridge this gap, we propose PRISK, a dynamic evaluation framework with automated data generation and tailored metrics that uncovers systematic limitations in current LLM personalization and how personalized information shapes its responses. Our empirical analysis across 13 LLMs demonstrates the presence of user profiles and retrieved memories consistently exacerbates biases, resulting in an average drop of 45.9% in irrelevant personalization, 41.7% in preference narrowing and 61.7% in sycophantic bias.

Representation learning begins when training changes the features that define similarity between data. A frozen-kernel model only reweights a fixed geometry. We establish quantum signal processing (QSP) as a solvable quantum model of the representation-learning regime. At arbitrary depth, we compute the exact mean and variance of its quantum neural tangent kernel, revealing an input-dependent angular geometry whose diagonal remains non-self-averaging even when the underlying unitary approaches Haar randomness. We also prove a sparse-data guarantee for the full nonlinear gradient flow without freezing or ensemble-averaging the kernel: the realized dynamics converges to an integrable scalar flow with a time-dependent kernel closure and explicit convergence times. A finite-depth speed limit holds for every data set and trajectory. At higher data density, numerical results show coupled evolution beyond both the scalar and frozen-kernel descriptions. These results give a controlled theory of learned quantum data geometry with provable training dynamics beyond the frozen limit.

Currently, in biological systems Neural Correlates of Consciousness (NCCs) are characterized in terms of EEG and FMRI signals. Unfortunately, this characterization prevents the transferability of the NCCs concept to machines. Such transferability would be useful in order to investigate AI consciousness. In this paper we provide an alternate characterization that is transferable, and enables the analogous definition of Machine Correlates of Consciousness (MCCs). Specifically, we propose that NCCs (MCCs) are substrate-level signals that are not under human (AI agent) control, and that are reliably modulated by emotions. This paper presents the first empirical investigation of MCCs. Specifically, we present the results of experiments conducted with two LLMs, Llama-2 7B and Llama-3.1 70B parameters. In these LLMs we collect hardware anomaly traces that are substrate-level indicator-sequences. And we show that after controlling for confounding factors, these are modulated differently by emotional and neutral computations. And this difference is statistically significant for the larger Llama-3.1 70B, but not for the smaller Llama-2 7B. The results constitute initial empirical evidence that MCCs are present in the Llama-3.1 70B configuration. And they are consistent with the hypothesis that consciousness probability and degree increase with the LLM sophistication. Independently of consciousness, MCCs can also be used for detection of emotions in AI agents.

Artists coordinate human pose, illumination, and camera placement to convey narrative and emotion, but existing generative methods typically model these elements independently. We introduce text-to-editable 3D staging, a task that jointly generates human poses, a dominant light, and a camera configuration from an affective description. We construct 11,911 text–staging pairs from 2,328 figurative paintings by reconstructing SMPL bodies, estimating low-frequency illumination, recovering camera parameters, and pairing each scene with ArtEmis descriptions. We train a flow-matching transformer that supports variable numbers of figures and produces multiple staging alternatives for each prompt. On held-out descriptions, the model achieves 32.2% retrieval R@1, compared with 16.6% for CLIP-based nearest-neighbor retrieval, while approximately preserving corpus-level diversity. These results demonstrate the feasibility of generating editable, emotionally conditioned 3D staging references from text.

Computational pathology (CompPath) is transforming medicine by leveraging artificial intelligence (AI) algorithms to support diagnosis, prognosis, and treatment prediction from gigapixel whole-slide images. Clinical adoption is progressing, but is constrained by concerns about safety, accountability, and regulatory oversight in high-stakes clinical environments. Explainable AI (XAI) systems hold promise for building trust and enabling verification, yet the literature remains fragmented due to inconsistent terminology, overlapping methodological families, ad hoc validation, and current reviews. This review aims to formalize XAI methods in CompPath through the: i) introduction of a pathology-centric vocabulary comprising seven core terms; ii) development of a taxonomy across methodological families and three orthogonal axes (stage, type, scope); and iii) establishment of a task-driven framework that maps five clinical questions to recommended methods, method evaluation, and deployment context. Five key gaps between current XAI capabilities and clinical deployment are identified, and actionable steps are proposed to advance XAI for CompPath.

Converting scientific figures into executable code has gained increasing attention, yet existing methods primarily focus on reproducing the reference figure itself. A more practical setting is to plot new data while preserving the visual style of a reference figure (e.g., color scheme and typography). Prior approaches mimic the reference through pixel-level optimization and struggle to carry its style to new data. We show that the key to this task lies in the coordinate grounding and coding capabilities present in modern computer-use models. We propose FigMirror, an agentic framework that unlocks these capabilities through Grounded Measurement, which locates visual elements by coordinates and measures their properties through executable code. We further introduce PlotTwin-Bench, an expert-curated benchmark with fine-grained code and image-level style metrics. Experiments show that FigMirror consistently outperforms existing methods on reference-conditioned style transfer. All plots in this paper are generated by FigMirror, except those produced by other methods for comparison. Our code and data are available at: https://github.com/VILA-Lab/FigMirror.

Extreme low-bit inference offers a route toward smaller models and constrained deployment. Ternary language models restrict weights to $\{-1,0,+1\}$, approaching the limit of $\log_2 3 \approx 1.585$ bits/weight. The practical question for a pretrained model is not simply whether weights can be quantised but which capabilities survive and whether it remains useful for adaptation. We explore this by converting Qwen3.5-0.8B (752M parameters) to ternary weights using 72.4M tokens of quantisation-aware training (QAT). The resulting model, Cloe, is evaluated across 29 benchmarks, representation diagnostics, and downstream fine-tuning. The evidence shows non-uniform degradation. A linear probe recovers 43.76% of MMLU answers from the full-precision teacher's representations but only 26.19% from Cloe (near chance), indicating specialist factual information is lost. However, Cloe retains measurable performance on ten tasks, averaging 77.1% of teacher performance. Crucially, fine-tuning raises Cloe to 89.8% on SST-2 (95.6% of the matched teacher) and reaches 79.4% teacher retention on XSum. We attribute degradation to a combination of quantisation-induced information loss and incomplete recovery due to the limited QAT budget. We also highlight an evaluation pitfall: standard answer-letter scoring failed (Cloe emitted "A" on 98.6% of MMLU questions), necessitating continuation scoring. Ultimately, ternary conversion is unsuitable as a drop-in general replacement yet remains valuable as a compact substrate for task-specific models.

Sparse autoencoders (SAEs) disentangle model activations into interpretable features and are widely used for steering large language models. Most existing SAE-based steering methods select features by applying a top- filter based on statistical scores, assuming that higher-scoring features yield stronger steering effects. In this paper, we show that this assumption is often invalid, leading to suboptimal feature selection. Our analysis reveals that effective steering features may be distributed among representationally adjacent, semantically similar groups induced by feature splitting in SAEs. Within such groups, features may exhibit disparate statistical scores despite having comparable steering influence, causing score-based selection to overlook important features. Based on these observations, we propose Neighbor Integrated Feature Selection (NIFS), a plug-and-play strategy that leverages representation similarity to improve feature selection for steering. We evaluate NIFS across multiple SAE-based steering methods and tasks, and demonstrate consistent performance gains over conventional top-$k$ selection.

Facial appearance editing powers popular applications like FaceApp and Photoshop. Generative Adversarial Networks (GANs) and 3D Morphable Models (3DMMs) have been widely used for facial editing. GANs can perform varied facial edits (e.g., changing hair color, hairstyle), but often produce unstable edits. 3DMMs produce stable edits, but can only alter pose and facial expression. Recently, text-guided diffusion models like Nano Banana have become popular for image editing. Text-guided models are a compelling alternative to GANs and 3DMMs since they can produce both stable and varied image edits. While text-guided models have been widely tested for whole-scene edits (e.g., “make the woman play a guitar”), they have not been comprehensively tested for facial editing. We conducted the first large-scale evaluation ($\sim1$M images evaluated) of six popular text-guided models on a sequential facial editing task. We present Face-Edit-Attributes, the largest collection of $169$ facial editing attributes focused on hair, accessories, and pose edits. We compared model performance using two popular celebrity face datasets: CelebA and CelebSET. Our results show that most models performed hair and accessory edits well, but struggled with editing pose. All models over-edit (e.g., changing hair color when asked only to change the hairstyle). We also evaluated demographic biases in each model. Our results show surprising biases in overediting: almost all models created more overedits for dark-skinned male faces and old faces. The code and data for our results (including our repository of $\sim 1$M images) can be accessed \href{https://github.com/rahul1801/Face-Edit-Bench}{\textcolor{blue}{here}}.

Modern artificial-intelligence coding agents can be equipped with tools for checking their own work e.g. a linter, a boot probe, a shell, a screenshot tool. We call this set the agent's verification surface. This study asks whether increasing only that surface, with everything else held fixed, produces a matching growth in the quality of the software the agent ships. We built a minimal coding agent whose tool list is the single controlled variable and used it to implement 1,116 web applications across six models and eight tool configurations. A condition-blind human graded every application against a frozen rubric, and automatic probes stress-tested the API-observable behaviors. Verification's cheapest benefit arrives first, which is to make sure that the application comes up. Without any tools, about one build in seven fails to launch at all and a single boot probe removes nearly all of these failures at roughly 35 percent of a full shell's token cost, while the full shell multiplies the no-tools cost by 2.35. Screenshots help most where mistakes are visible (e.g. element placement, interaction), though even there the gain over a shell is modest and does not survive correction for multiple statistical comparisons. In cases where failures can only be measured rather than seen, such as keeping scrolling smooth over a 100,000-row list, screenshots add nothing. A verification tool improves the output artifact only where its reach covers the way the application actually fails.

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

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

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

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

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

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

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

Tendon-driven hands are anthropomorphic, and moving the actuators off the joints is what makes a hand of this capability affordable to build. Two effects produce that saving. Routing force through a cable removes the requirement that a motor fit inside the joint it drives, so smaller and cheaper motors suffice, and one motor can drive several joints through a single cable, so fewer motors are needed. They are also harder to learn on than a direct-drive hand. The underactuated transmission that produces the saving is itself difficult to represent in a simulator, and the joints one cable drives are not independently commandable. We present Aero Hand Open, a tendon-driven anthropomorphic hand that is released simulation-ready. Three things ship with it. A simulation model reproduces the cable transmission itself. An identified actuation map connects that model to the motor commands in both directions, including the three-way coupling of the thumb. A reinforcement learning package trains policies for the hand. Together they let a policy be trained entirely in simulation and run on the hand with no fine-tuning and no state estimation. We release the mechanical design, the simulation model, the identified mapping, the training environment and the deployment stack.

Synthetic data can improve statistical inference when real data are scarce, but naively treating synthetic samples as real data can introduce bias and lead to unreliable inference. We develop a general framework for synthetic-augmented inference across a population of related tasks. It characterizes synthetic augmentation by the number of synthetic observations and their weight. Central to our framework is a size-weight frontier that specifies, for each weight, the largest synthetic sample size for which all smaller sizes attain the target task-marginal coverage. We estimate this frontier from historical tasks, and establish a finite-sample coverage guarantee simultaneously for all size-weight configurations on or below the estimated frontier. In experiments using large language model responses to augment opinion survey data, our procedure achieves target coverage and substantially narrows confidence intervals.