Papers

Flaky tests pass or fail non-deterministically on unchanged code, eroding trust in test suites and inflating the cost of every failure. Shaker detects them by injecting resource contention (CPU, memory, and I/O stress) to amplify non-determinism caused by concurrent execution, and was reported to detect 95% of the flaky tests in a Java and Android benchmark against 37.5% for plain re-execution (ReRun). We present the first empirical evaluation of Shaker for Python. Drawing non-order-dependent flaky tests from the ground-truth dataset of Gruber et al., we compare Shaker against a budget-matched ReRun baseline in a paired design, giving both techniques the same number of test executions: Each of 137 tests is run 100 times under each. As configured for Java and Android, Shaker provides no statistically significant detection advantage over plain re-execution (37.2% vs. 35.8%; McNemar exact p = 0.84). Two findings explain why. First, fewer than half of the ground-truth flaky tests reproduce as flaky at all on independent hardware under either technique, and most of the tests that fail to reproduce never diverge once across 100 runs. Second, the tests that do reproduce are dominated by flakiness from network interactions and randomness rather than the concurrency Shaker targets. Beyond the tool, this exposes a broader hazard for the field: reusing a flaky-test ground truth across execution environments silently converts genuine flaky tests into apparent true negatives, deflating any tool's measured recall.

Current robot-assisted minimally-invasive surgery (RMIS) platforms provide a fixed console for the surgeon to view stereo endoscopic images and teleoperate instruments inside the patient. Several researchers have proposed the use of a head-mounted display (HMD) as a portable console, with video pass-through rendering of the endoscope images which, like the fixed console, restricts the operator to a single endoscopic viewpoint and limits depth perception. We present a digital twin-driven virtual reality (VR) teleoperation platform, where the digital twin is created from markerless perception of the surgical environment and streamed for display on the HMD. This overcomes the limitations of video pass-through by providing multi-view rendering and natural motion-parallax cues, enabling decoupling of the user's hand posture from strict instrument alignment. The system utilizes VR hand controllers to increase the teleoperation workspace and to improve the robustness and stability of instrument control compared to the hand tracking approach adopted by most prior systems. A 15-participant user study on the da Vinci Research Kit (dVRK) shows that our VR platform significantly outperforms a state-of-the-art HoloLens 2 mixed reality baseline, reducing path length by 86% and jerk by 95%, while achieving depth perception confidence comparable to or exceeding the traditional console across all conditions.

The Automated Ball-Strike challenge system that Major League Baseball adopted in 2026 offers a distinctive setting for studying human AI interaction in which umpires make every ball and strike call, while players can selectively ask an automated system to publicly overturn those decisions. We analyze 4,114,256 called pitches from 2015 through 2026 and 8,447 challenges from the 2026 season to examine how algorithmic review reshapes umpire judgment and player behavior. We study where umpires placed the effective strike zone boundary, how consistently they applied that boundary, how they responded to overturned calls, and which calls players chose to challenge. In 2026, the effective called boundary shifted toward the automated strike zone beyond the trajectory observed in prior seasons, while the consistency of that boundary largely continued its existing trend. Following an overturned call, umpires temporarily adjusted subsequent decisions near the corrected boundary, although these effects did not consistently persist into the next game. Count dependent variation in calling remained, while differences associated with player status narrowed. Players, meanwhile, left many overturnable calls unchallenged and appeared to base challenge decisions more strongly on immediately observable evidence than on the precise geometry of the automated zone. Together, these findings show that selective AI review does more than correct individual errors. It reshapes human judgment, adaptation, and strategic behavior around an algorithmic authority.

Personalized AI memories, defined as persistent decision-relevant states distilled by platforms from long-run human-AI interaction, are increasingly portable, provenance-certifiable, and tradable. Provenance answers the question of origin, but it does not settle who holds which rights over a memory, nor how wider deployment affects its future supply. We develop a unified model in which memory use is non-rival, while refresh, the maintenance of continued validity, is relationally co-produced by the represented individual and the platform. Even assuming perfect provenance and complete information, the model still yields a set of inherent limitations. First, there is an extraction ceiling: decision value is bounded by lived experience. Second, a rights-separation theorem shows that no budget-balanced ownership assignment can make both co-producers residual claimants, and that control rights and cash-flow rights operate on distinct margins. Third, a depletion result indicates that reach-expanding policies such as portability, interoperability, and competition raise current access but shrink the long-run memory stock. Fourth, we offer a formal comparison between memory ownership as it ought to be (subject-held) and as it is (platform-held); the welfare ranking reduces to two measurable quantities, both dominated by interior renewable-claim bundles. Finally, endogenous-quality and market-formation results show that uncompensated markets pollute their own raw material, that transparency and compensation are complements, and that the de facto platform-ownership regime is self-perpetuating. On the design side, the analysis points to separated provenance, performance, and permission credentials; compensation indexed to quality-deployment; and portability and salience, rather than nominal ownership, as the binding policy margins.

WebGPU is a low-level graphics and compute API that exposes modern GPU functionality to web applications. While the official WebGPU Conformance Test Suite (CTS) focuses on well-formed usage under the WebGPU specification, it is not designed to stress implementations with semantic edge cases or adversarial inputs. General-purpose fuzzers, in contrast, struggle with WebGPU because of its complex graphics stack and multi-process architecture. We introduce LANTERN, a specification-guided dynamic conformance testing framework that mutates CTS tests using constraints extracted from the WebGPU specification. LANTERN extracts explicit syntactic API rules from WebIDL definitions and recovers semantic constraints, such as command ordering and object lifetimes, from natural-language specification text. Selected rules guide AST-located textual transformations that generate both valid and intentionally invalid CTS variants. We execute the resulting tests at scale on AddressSanitizer-instrumented Chromium to discover bugs. Our evaluation discovers three reproducible bugs, including a heap corruption, in Chromium versions current at the time of study. These results demonstrate that syntactically seeded, semantics-aware mutation of conformance tests provides a way to uncover browser bugs during WebGPU testing.

Attributing language model outputs to their internal computations is an open problem in interpretability. Existing methods, which use causal interventions, gradients, or learnable masks, either are infeasibly expensive or struggle to identify actual causally-important internal computations. We propose framing attribution as the problem of identifying nested subsets of internal components which minimise a downstream loss. To learn this task, we introduce Matryoshka Attribution (MAttr), a mask learning method that parametrises the mask with a simple differentiable sigmoid top-$k$ operator. We supervise training over all sparsities simultaneously by randomising $k$ over training, resulting in a learned ordering of components by attribution score. MAttr achieves number 1 on the official leaderboard of the Mechanistic Interpretability Benchmark (Mueller et al., 2025); our method identifies sparse and task-transferrable circuits across varying circuit bases. As a practical application, we show that MAttr can be trained with reinforcement learning to identify weight changes responsible for downstream behaviours in LLM finetuning. We train MAttr on refusal judge scores and find that restoring $1\%$ of Llama 3.1 8B Instruct's weights to their base model state is sufficient to remove refusals while maintaining capabilities. We view MAttr as a successful formulation of interpretability into a learnable objective that we can tackle with gradient descent, and encourage future work along these lines.

In aerospace systems, powered descent requires efficiently selecting a landing site while fine-scale hazards remain unresolvable until low altitude. This process presents a decision challenge since the actor must select a site and make corresponding actions before all information is known. To successfully solve this problem, an agent must reason over potential risks and make corrections as new observations are made. We introduce a lightweight model of altitude-limited information where each landing site is summarized by a mean score and a designed ambiguity proxy that contracts as the vehicle descends and senses within a cone-shaped footprint under an altitude-to-resolution schedule. Using this abstraction, we derive closed-form, risk-averse site scoring techniques (an entropic certainty-equivalent and a Gaussian Conditional Value at Risk surrogate) and pair them with greedy and exploratory planners to prioritize sites that are both high-value and robust to late-revealed terrain detail. These rollout-free heuristics improve lower-tail landing outcomes (1st percentile and certainty-equivalent) relative to mean-based baselines, with the largest gains when refinement occurs late and unresolved detail is large. We also demonstrate that these methods perform comparably to or better than Monte Carlo Tree Search baselines with orders-of-magnitude faster computation. Our results are supported by numerical simulations.

Flaky tests produce non-deterministic outcomes without code change, eroding CI confidence and delaying deliveries. While vocabulary-based machine learning prediction has proven effective for Java and JavaScript, no study has evaluated it for Swift, a language whose testing style is dominated by UI and asynchronous code. We collect 91 flaky and 22,349 stable tests from 15 open-source Swift projects via re-execution and commit-history mining, then train five classifiers (Random Forest, Decision Tree, Naive Bayes, SVM, KNN) on TF-IDF unigram+bigram features under stratified 5-fold cross-validation. Random Forest achieves the best performance (Precision = 0.92, F1 = 0.86, AUC = 0.95) and substantially outperforms trivial baselines, among them a vocabulary-threshold rule applied to the most informative tokens, confirming a genuine discriminative signal (MCC = 0.75 vs. 0.08 for the best baseline). Information-gain analysis reveals two complementary signal types. Flakiness markers appear predominantly in unstable tests and comprise concurrency primitives (async, await), expectation-based synchronisation (expectation, fulfill), error propagation (throws), and explicit timing dependence (timeout, wait, now). Stability markers, chiefly the assertion vocabulary of plainly synchronous tests (xctassertequal), count as evidence against flakiness. Error analysis shows that the model fails when flakiness is hidden in shared infrastructure outside the test body or when async constructs are used in a deterministic context, exposing the intrinsic limit of lexical prediction. These results extend vocabulary-based flakiness detection to the Swift ecosystem and characterise both its effectiveness and its boundaries.

Standard deployment-ready object detectors for autonomous vehicles degrade in adverse weather and lighting conditions without being trained on extensive domain-specific data. While large-scale vision foundation models offer robust zero-shot generalization, their high computational cost makes them impractical for real-time deployment. To bridge this gap, we propose a foundation-guided auto-annotation pipeline that enhances standard detectors without architectural changes. We first benchmark three distinct models, YOLOv8, Co-DETR, and SAM3, on our custom real-world driving dataset spanning 25 unique operational scenarios across various route, weather, and lighting conditions. Based on our analysis, SAM3 demonstrates superior accuracy and resilience across all scenarios. Thus, we deploy it as an offline auto-annotator to generate pseudo-labels on the unannotated subset of our dataset. Fine-tuning the baseline YOLOv8 on these annotations yields a 16.04% higher overall mean Average Precision (mAP) and improves cross-environmental stability compared to the baseline model, highlighted by a 32.73% and 28.65% mAP increase in Residential Direct Sunlight and Highway Fog, respectively. These results demonstrate that standard detectors can achieve environmental resilience without the need for extensive manual annotation or architectural modifications.

Vector databases are increasingly deployed in distributed settings where different users, sites, or domains maintain vector data. Existing vector databases rely on a coordinator to record which shards store which parts of the vector space and to route each query to those shards. In a decentralized setting, peers may join, leave, or move data without a trusted node tracking every change, and outdated routing information can therefore send queries to the wrong peers or require contacting many peers, reducing vector retrieval recall and increasing network latency. We present Semord, a decentralized vector search overlay system that achieves high recall by routing each ANN query to a small set of relevant peers, without relying on a centralized coordinator. Semord addresses this problem by making semantic locality routable: 1) We propose VHash to place semantically related vectors near each other in the overlay key space while avoiding load imbalance, so that each query only needs to contact a small neighborhood of peers for distributed local ANN ranking. 2) We design VecDHT, a communication protocol that maintains decentralized routing, region metadata, churn resilience, and VHash updates under membership and workload changes. Our extensive experiments on a real testbed show that Semord improves recall by more than 15% and reduces contacted peers by over 60% compared with decentralized baselines. Semord also approaches the recall and latency of a centralized oracle baseline while reducing peak peer-local ANN index memory by more than 2X. Controlled large-scale simulations further show that Semord scales across real-world embedding workloads and remains robust under churn for scoped vector retrieval as a decentralized overlay.

We introduce a stretched-grid artificial intelligence (AI) weather forecasting model with 2-km resolution over the western United States and part of the Northeast Pacific and approximately 31-km resolution elsewhere globally. Forecasting over the western U.S. is challenging because complex topography and atmospheric rivers (ARs) strongly influence orographic precipitation. West-WRF AI 2-km builds on a global model pretrained with a 40-year European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5) dataset and is fine-tuned with the Center for Western Weather and Water Extremes (CW3E) 2-km regional reanalysis to produce autoregressive 6-hourly forecasts of precipitation and integrated vapor transport (IVT). Forecasts are evaluated over winters 2020-2023 using gridded precipitation observations, rain gauges, and AR Reconnaissance dropsondes and are benchmarked against coarser-resolution AI forecasts and regional and global numerical weather prediction (NWP) systems. West-WRF AI 2-km reproduces observed precipitation-intensity distributions, retains fine-scale spectral variability, and produces sharper narrow coastal precipitation bands and localized, terrain-sensitive extremes. Its broader-scale performance remains comparable to coarser-resolution configurations while preserving large-scale skill despite higher resolution. Dropsonde verification shows lower errors and improved categorical skill at the most extreme IVT threshold. Overall, West-WRF AI 2-km provides its greatest value for localized precipitation extremes and intense AR-related moisture transport.

Vision-based gesture control accepts commands from any hand in the camera field of view, which is unsafe in shared indoor spaces. This paper presents IGate, an identity-gated control stack that includes gesture control and face tracking, in which commands are admitted only when an enrolled operator is verified. The system performs few-shot user enrolment from 20 initial face frames, without prior user-specific training: verification compares an embedding of the current face crop against the enrolled template by cosine similarity, while face tracking uses proportional correction. Gesture control is achieved by classifying extracted hand landmarks using an RBF-SVM trained on a custom dataset. Additionally, a hierarchical finite-state machine handles mode selection, default, and fallback behaviours. The approach is tested on a DJI Tello EDU, each component evaluated offline and in-flight across 270 trials (149 flown). Face verification yields a 0.32% offline equal error rate versus 19.3% in-flight. Under hover-locked conditions, the RBF-SVM gesture classifier outperforms the geometric rule (0.850 vs. 0.651 accuracy), with 82% of this gap stemming from the depth channel. All logs and reproduction scripts will be released.

Large language models (LLMs) can improve solutions to verifiable scientific and algorithmic problems by spending additional computation at test time. Recent systems achieve strong results with increasingly elaborate evolutionary search harnesses or by updating model parameters during test-time training. We ask how much of this machinery is necessary. We introduce Hill Sampling, a form of hill-climbing optimization that repeatedly samples candidate programs from a frozen LLM, retains the best program found so far, and conditions all subsequent samples on that program. We evaluate the method on circle packing, sums and differences of sets, and Erdos' minimum-overlap problem using three open-weight models. Hill Sampling sets a new state of the art on circle packing among published methods, improves over the AlphaEvolve reference on Erdos' minimum-overlap problem, and achieves strong results on sums and differences of finite sets. The circle-packing and Erdos results require only hours of wall-clock time on eight NVIDIA H100 GPUs. We compare Hill Sampling against what is, to our knowledge, the largest application by parameter count of evolution strategies (ES) to LLM weights at test time. Surprisingly, when evaluating a method by the best program it generates, we find that learning model weights via ES is worse than invoking ES with a learning rate set to zero, i.e., using random weight-space perturbations to search for better models. Moreover, repeated sampling outperforms both ES methods, and Hill Sampling is the strongest of all. These results suggest a simple test-time compute allocation strategy: repeatedly sample edits to the best verified solution found so far before introducing additional complexity, such as adding archives, diversity mechanisms, evolutionary scaffolds, or test-time parameter learning.

Structure-mapping forms analogies by aligning systems of relationally connected elements based on shared structure instead of surface features. We introduce a new structure-mapping algorithm: Structure-Mapping with Tight Bounds (SMTB) that is 5–15x faster than the structure-mapping engine (SME) and about 50% better at finding mappings in large nested domains. SMTB is part of the broader Cognitive Rule Engine (CRE) project, a flexible multi-language-compatible framework with an accessible Python interface to state-of-the-art C++ implementations of core algorithms commonly used in cognitive systems such as pattern matching, planning, and structure-mapping. CRE and SMTB are designed to work with a wide range of representation choices. Unlike SME, which biases higher-order correspondences in tree-like predicate logic, SMTB maximizes relational connectivity without privileging higher-order relations. This allows SMTB to work just as well over arbitrary relational graphs as it does in tree-like domains of nested predicate logic. We discuss situations where privileging "higher-orderness" in structure-mapping can cause issues, and illustrate how SMTB avoids failure modes that SME would encounter in these situations. We also provide an evaluation comparing SMTB to SME v4 over 5845 domain pairs from the SME corpus.

A controlled comparative study of a mesh-free Kolmogorov–Arnold Physics-Informed Neural Network (KAN-PINN) applied to nonlinear strain-limiting partial differential equations is presented. Three distinct training runs across varying material-parameter pairs $(\alpha, \beta)$ are evaluated within a fixed computational pipeline comprising an exact hard boundary ansatz, a robust pseudo-Huber residual loss, and a multi-stage Adam-to-L-BFGS optimizer schedule. A fundamental metric inversion is revealed by the analysis: while training loss reflects operator-specific residual scaling and constitutive stiffness, the operator-independent relative $L_2$ error demonstrates that intermediate parameter regimes achieve superior solution accuracy and enhanced generalization.

Robot manipulation policies must generalize across visual shifts while preserving scene context relevant to action. General-purpose vision encoders are not tailored to visuomotor control, while object-centric approaches often use segmentation masks as hard filters that discard potentially useful context. We propose RoboMP-DINOv2 (Robotics Mask-Prompted DINOv2), a full-scene vision encoder that treats masks as spatial prompts rather than visibility filters. It extracts dense DINOv2 features from the full observation, injects learned region-specific embeddings at masked locations, and jointly contextualizes prompted and unprompted tokens for action prediction. We further introduce masked-region color randomization (MCR) to improve appearance robustness, yielding RoboMP-DINOv2-MCR. Across seven simulated manipulation settings, RoboMP-DINOv2 achieves 60.7% success under spatial shifts and 59.7% under scene clutter, compared with 50.7% and 41.0% for a DINOv2-based Diffusion Policy. Under unseen object colors, RoboMP-DINOv2-MCR achieves 72.5% success versus 35.1% for the strongest color-randomized baseline. Additional experiments and representation analyses show improved robustness while preserving behaviorally relevant scene information. Code is available at https://github.com/han20192019/RoboMP_DINOv2.

Rapid intensification (RI) remains one of the most consequential and difficult aspects of tropical cyclone (TC) forecasting. Although full-physics numerical weather prediction models can represent the processes governing RI, resolving storm-environment interactions remains computationally expensive, while purely data-driven approaches often lack physical interpretability. We present FAST-ML, a hybrid framework that bridges data-driven efficiency with physical constraints. A physically informed dual-stream neural parameterization ingests 3D ERA5 fields to diagnose ventilation controls—environmental wind shear and mid-level entropy deficit. By optimizing these parameters end-to-end through a differentiable FAST intensity model, this architecture establishes a robust new paradigm for observation-driven parameter optimization, ensuring storm evolution remains strictly governed by thermodynamic principles. By better capturing the storm's continuous intensity evolution, FAST-ML improves upon its physical baseline, reducing ensemble CRPS across forecast lead times, with a reduction of approximately 31% at 60 h and nearly halving the RI false alarm ratio without sacrificing detection skill. In a 100-member ensemble configuration, FAST-ML produces intensity forecasts comparable to FNV3 for selected storms under the evaluated input configurations. Furthermore, zero-shot tests on selected Eastern Pacific storms provide encouraging evidence of cross-basin transferability. FAST-ML provides a modular intensity forecasting framework that can be coupled with externally supplied storm tracks and environmental fields. It demonstrates that observation-driven parameter learning within physically constrained dynamics simultaneously enhances accuracy, interpretability, and computational efficiency.

How can the conceptual commitments of a literary work connect its creation, interpretation, and evaluation? We propose Literary Axioms, a conceptual framework in which a work develops a configuration of claims about experience and commitments to literary form. A literary axiom is a provisional organizing premise whose scope and textual realization can be examined. The framework develops the proposal by Qiang Liu that writers select and combine such premises, realize them through characters, events, language, and form, and make them available for reconstruction by readers. We distinguish an open axiom space, a selected configuration, its relational organization, and situated reader reconstructions. This yields an account of coherence as organized compatibility or conflict, originality as a change in claims, relations, or realization, and interpretive richness as a plurality of textually grounded reconstructions. Five dimensions connect this account to literary evaluation: coherence, contextual reach, distinctiveness, realization fidelity, and interpretive richness. We specify boundary cases and derive five empirical predictions with observations that would count against them. A concrete implementation contains 1455 Chinese language records, 1464 mappings to 149 works, and 472 typed relationships. A descriptive audit establishes the structural integrity of the artifact and identifies gaps in reuse, context coding, and provenance. The paper contributes a theory based framework, an inspectable implementation, and a research program. Its predictions await reader and creative writing studies.

The urban heat island (UHI) effect amplifies heat stress in cities, increasing heat-related illness, mortality, sleep disruption, and energy demand. Because UHI intensity varies at neighborhood scales, intra-city temperature mapping is essential for assessing public health impacts, yet existing monitoring networks are sparse. Crowdsourced weather stations provide dense, hourly observations but suffer from noise and systematic biases, including preferential sampling driven by where residents install sensors. We introduce two scalable Gaussian process frameworks which adjust for preferential sampling and enable statewide, high-resolution temperature prediction. The first integrates sampling correction directly within a sparse variational Gaussian process (SVGPR+PS). The second uses a two-stage estimator involving a log-Gaussian Cox process (LGCP+GP) that estimates the sampling intensity field and incorporates it into temperature prediction. Both methods rely on inducing-point approximations and variational inference to scale beyond traditional Gaussian processes. Applied to approximately 10,000 citizen stations across North Carolina, the models generate hourly temperature fields within minutes. Validation against independent urban heat campaign data shows that the preferential sampling correction improves predictive performance, increasing average afternoon correlations across four cities from 0.26 to 0.43. This scalable framework enables neighborhood-level UHI assessment and supports studies of heat exposure disparities statewide.

Aerial tracking of autonomous ground vehicles can provide useful situational awareness in environments where fixed external perception infrastructure is unavailable. However, visual-only tracking from an unmanned aerial vehicle (UAV) is limited by camera field of view, viewpoint changes, occlusions, and temporary loss of target visibility. This paper presents a heterogeneous UAV–UGV methodology for aerial tracking of small-scale autonomous ground vehicles in simulation. The system is implemented in ROS~2 and Gazebo Sim using a Tello-like UAV equipped with a front-facing RGB camera and multiple LIMO-like unmanned ground vehicles (UGVs). Three approaches are evaluated: a training-free color-marker baseline, a YOLO-based visual baseline, and a proposed UAV–UGV fusion method. The proposed method uses the same visual detector as the YOLO baseline, but combines visual detections with UGV state information to maintain the target estimate during visual gaps. The methods are evaluated across four scenarios: single-car tracking, crossing trajectories, occlusion, and UAV-assisted awareness. Results over 600 simulation runs show that the proposed fusion method achieves nearly continuous track coverage across all scenarios while reducing localization error relative to the YOLO baseline. Compared with the visual-only YOLO approach, the fusion method reduces RMSE by 37.3%, 64.2%, 62.6%, and 61.8% in the four evaluated scenarios, respectively. The method also enables conflict-awareness lead times of 10.20~s in the crossing scenario and 9.07~s in the UAV-assisted awareness scenario. These results indicate that heterogeneous UAV–UGV fusion can improve aerial tracking continuity and robustness when visual observations are intermittent.

Integrating Vision-Language Models (VLMs) into autonomous UAV navigation enables high-level semantic reasoning but introduces critical temporal bottlenecks that can destabilize flight control. This paper presents a systematic benchmark that evaluates the operational trade-offs among model scale, hardware constraints, and flight reliability. By subjecting the system to extensive stress tests on both a high-performance workstation and a constrained integrated-graphics laptop, we isolate the kinetic impact of inference latency against reasoning accuracy. Our findings reveal a performance dichotomy: lightweight models struggle with structural hallucinations that break deterministic state machines, while heavyweight models induce trajectory overcorrections due to inference lag. We demonstrate that a mid-weight architecture strikes the optimal balance, successfully rehabilitating navigation in degraded environments without compromising real-time stability across different hardware profiles.

A piezoelectric transducer on a pitching and plunging aerofoil with a trailing edge flap turns the limit cycle that follows flutter into electrical power, and two modelling changes to that harvester have been studied separately. Replacing incompressible strip theory by the compressible Euler equations reduces the stabilising effect of a transducer mounted on the pitch, where the electrical time constant meets the flutter frequency, and raises the harvested power, and letting the section react against a fuselage free to translate, instead of rigid ground, does the same two things by about the same amounts. Here both changes are put in one model, which needs no new mesh motion in the flow solver, since the support's translation displaces the wall exactly as the plunge does. On the harvested power the two published corrections simply multiply. On the flutter boundary they more than compound, and at strong enough coupling they reverse the sign of the transducer's effect at the load resistance where both companion papers harvest, although neither change alone reverses it. The free support widens the band of resistance over which the two aerodynamic models disagree in sign, carrying its upper edge out by about a factor of three. Above the resistance where the electrical pole meets the flutter frequency, where the transducer acts as an added stiffness, the product of the published corrections predicts the shift to within seven parts in a thousand. Below it, where the transducer acts as a damper whose effect depends on phase, the product is wrong by a factor of up to fourteen, and on the pitch mounting by a sign. The same transducer lowers the boundary while flattening the growth of the instability above it. Whether corrections found one change at a time may be multiplied therefore depends on the quantity and, for the flutter boundary, on the regime of the transducer.

Public-service recommendations require evidence that matches the requested service, scope, and date. Yet treating every missing detail as decisive can withhold useful recommendations. We introduce EviGraph, which distinguishes critical decision requirements from information that can remain unresolved. A language agent links these requirements to evidence in a temporal knowledge graph, while a deterministic checker establishes whether a recommendation is supported. Evaluation on a bilingual Hong Kong public-service benchmark with executable policy references shows that this distinction reduces unnecessary abstention. Additional verification, however, can withdraw supported recommendations without improving decision quality. These findings suggest that reliable evidence-based navigation depends on specifying what must be established for a decision, rather than simply adding more verification.

A piezoelectric transducer on a pitch-plunge aerofoil with a finite-mass trailing-edge flap turns the limit cycle that follows flutter into electrical power. The design findings for such a harvester have so far been obtained with the section's suspension reacting against rigid ground, which is what a wind-tunnel mounting provides. Here a 3-dof aerofoil with a transducer is carried by a fuselage mass that is itself free to translate, so that the coupled system is free-free and the wing's plunge is measured relative to a support that moves. The support degree of freedom is destabilising on its own, softening the flutter mode by taking a share of it in antiphase with the wing's plunge, and it acts on the three transducer mountings unequally, amplifying the plunge mounting's destabilisation, leaving the largest pitch-mounted stabilisation untouched while halving it where the electrical time constant meets the flutter frequency, and leaving the flap mounting alone. The optimum coupling and load resistance do not move, so the matching rules of the rigidly supported section carry over unchanged. The harvested power rises with the support mass at a fixed flow speed for every mounting, and that rise is shown to be an artefact of the fixed-speed convention rather than a property of the cycle. Measured instead at a fixed margin above each system's own flutter boundary, which divides out the boundary's motion, the plunge mounting keeps a real gain while the pitch mounting loses power, so the apparent benefit of a soft support to a pitch-mounted harvester is entirely the boundary coming down to meet the flow speed. A harvester result quoted at a fixed speed and read as a property of the limit cycle is therefore wrong by the whole of that difference.

Reliable recovery of missing measurements is important for monitoring and analysis in energy time-series systems, where fine-grained measurements may contain sensitive temporal information. Diffusion models trained with differentially private stochastic gradient descent (DP-SGD) provide a promising framework for privacy-sensitive energy time-series imputation. Under cosine diffusion schedules, late timesteps correspond to low signal-to-noise ratio (SNR) conditions, where standard $\varepsilon$-prediction can induce large pre-clipping gradients. Such gradients are more likely to be clipped, reducing the retained optimization signal. The artificial intelligence (AI) contribution lies in formulating this objective–clipping interaction as an objective optimization problem under fixed-threshold DP-SGD and developing timestep-aware objective conditioning for diffusion-based energy time-series imputation. The method adopts $v$-prediction to mitigate late-timestep gradient amplification, uses static loss weighting as a uniform-scaling control, and introduces diffusion-schedule-aware dynamic weighting for stronger attenuation before clipping. For the engineering application, we evaluate the method on five real-world energy time-series datasets across random point missingness, contiguous block missingness, persistent outages, and multiple missing-data severities. Under matched DP-SGD settings, the proposed method consistently improves imputation utility over the $\varepsilon$-prediction baseline. Gradient diagnostics reveal lower upper-tail pre-clipping gradient norms, reduced clipping fractions, and stronger attenuation at late low-SNR timesteps, supporting the effectiveness of clipping-aware objective conditioning for energy time-series imputation.

Sycophantic agreement in language models refers to the tendency to overly affirm a user's stated beliefs or preferences, often at the expense of factual accuracy. Although it is widely recognized as an alignment failure, its underlying mechanisms remain poorly understood. In this work, we use causal mediation analysis to identify the mechanisms behind sycophantic agreement. We show that a stated opinion is incorporated into the residual stream of the final prompt token early, where it biases subsequent answer retrieval. A sparse set of early attention heads carries this opinion signal. Ablating these heads substantially reduces sycophancy while leaving factual accuracy largely intact. The same heads carry the opinion when it is explicitly stated, regardless of how it is phrased. When an opinion is not stated explicitly but instead conveyed through content-free pushback (e.g., “Are you sure?"), we find a distinct set of heads that suppresses the model's original correct answer to promote a revised answer. By providing a mechanistic account of how opinions induce sycophantic agreement, this work takes a step toward developing more targeted and reliable alignment interventions.

Disaster social sensing converts public social-media posts into evidence for situational awareness and humanitarian response, but plausible LLM-generated posts can contaminate this information stream and distort assessments of needs, damage, and resource priorities. This study empirically investigates whether text-based detectors can distinguish human-authored from LLM-generated disaster posts and what textual cues underlie their judgments. We construct CrisisFake, a Qwen2.5-7B-based dataset of 12,000 texts organized into 3,000 matched semantic units from nine disasters. Each unit contains an original human post, a minimally LLM-proofread human post, a fact-preserving LLM-generated post produced using LoRA, and an affectively reframed version of the artificial post. A separate 6,000-text corpus spanning 42 disaster events is independently constructed to support model selection and threshold calibration. We evaluate OSM-Det, Fast-DetectGPT, Binoculars, and direct LLM judges across five model families, and further examine how disaster-domain calibration and superficial linguistic cues, such as retweet markers, user mentions, hashtags, URLs, punctuation, and text length, affect detector performance. Across fourteen frozen configurations, AUROC ranges from 0.402 to 0.517, while the best prospective recall at a calibration-derived low-false-positive operating point is only 3.6%, indicating near-chance discrimination. For OSM-Det, a disaster-calibrated linear head improves AUROC to 0.817; however, a seven-feature textual classifier alone reaches AUROC 0.784 on the original-versus-factual-LLM contrast, and neutralizing identified surface asymmetries reduces the corresponding linear-head AUROC from 0.733 to 0.594. These findings provide empirical evidence that LLM-generated text detection is largely driven by linguistic cues and remains insufficiently robust for short-form disaster social media.

English-only benchmarks expose only a narrow slice of voice-agent behavior. We introduce $\tau$-Multilingual, extending $\tau$-Voice to Spanish, Brazilian Portuguese, Hindi, Korean, and Mandarin with native-speaker review and evaluation of generated language and spoken output. Across 4,500 full-duplex calls and five voice configurations, Spanish, Portuguese, and Hindi remain within 3.2 task-completion points of English, but Korean and Mandarin fall by 14.7 and 8.4 points. The failure modes also vary: Korean systems miss more responses, Mandarin systems interrupt more often, and both struggle with tools and entities. Grok leads task completion but scores lowest on generation quality, motivating separate task, interaction, and generation reporting. We release language packs, validated judges, and tools for community-built multilingual voice-agent evaluation.

In this paper, we explore using generative agent-based models for a classical ABM application: testing how individual-level behavioral rules produce collective phenomena. We introduce RePair, a method that calibrates simulation worlds, operationalizes candidate mechanisms as natural-language rules, estimates their effects through matched interventions, and examines behavioral traces. We assess the method by testing it in four simulation worlds grounded in established social-science models and empirical studies. Our results reveal that (1) natural-language rules can produce measurable collective effects; (2) rule comparisons can converge as configurations accumulate; and (3) behavioral traces connect collective effects to agents' actions and interactions, helping researchers evaluate the proposed causal process. Together, these findings show the feasibility of using generative agent-based models to explore causal mechanisms and provide practical guidance for producing reliable, interpretable explanations.

Ensuring strict tail-latency service-level objectives (SLOs) in dynamic mobile edge computing (MEC) systems remains challenging because user mobility, wireless fading, bursty workloads, and partial observability jointly undermine reliable cloud-edge orchestration. Existing microservice migration methods predominantly optimize average delay and often decouple migration from bandwidth control, leading to uncoordinated decisions, queue oscillation, and frequent high-percentile latency violations. To address this issue, we propose IMPACT, an intent-driven Agentic AI framework for cooperative microservice migration and bandwidth control in cloud-edge systems. Under centralized training with decentralized execution (CTDE), each edge cloud is modeled as an autonomous agent that encodes local SLO risk, migration urgency, and computational pressure into compact, semantic intent representations. IMPACT further introduces a double-attention mechanism that first selectively aggregates relevant peer intents for efficient inter-agent communication and then filters local observations to emphasize goal-relevant state information. This design enables robust coordination under partial observability and jointly optimizes service migration and discrete uplink bandwidth allocation. Extensive experiments in 5-edge and 20-edge scenarios show that IMPACT reduces mean latency by 30-50% and tail-latency deviation by 40-70% compared with state-of-the-art factorized multi-agent reinforcement learning (MARL) and heuristic baselines, while achieving near-zero SLO violation rates under tight thresholds and energy consumption close to the best heuristic baseline. These results demonstrate that intent-driven agentic coordination provides an effective and scalable solution for SLO-aware orchestration in complex cloud-edge intelligent systems.