Graphics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Thursday 13 August
This study introduces the spatial and analytical construction of the Quartic Overhauser (QOVR) surface generation framework designed to address boundary alignment and localized shape modification constraints. This framework implements a variable-parameter fourth-degree novel matrix formulation to achieve parameter tracking across orthogonal coordinate axes. The analytical pipeline integrates directional spline blending functions with symmetric spatial control matrices, ensuring that internal knot vector variations allow localized surface adjustments while maintaining the position convergence of global edge boundaries. Computational verification confirms that while the current formulation satisfies explicit C^{0} positional closure and C^{1} tangent continuity conditions across the internal and boundary interfaces without triggering global curvature propagation, with edge joint separation, or wave-like artefacts, C^{2} curvature continuity is not maintained. As demonstrated by three-dimensional mesh models and colormap visualizations, the spatial intensity fields condense strictly within the immediate neighborhood of the modified element, verifying that the displacement effect decreases progressively as the distance from the perturbed control points increases. This explicit decoupling aligns with structural symmetry and boundary alignment across adjacent geometric patches, satisfying manufacturing and reverse engineering sealing requirements.
Machine Learning
The Impact of Temporal Context Length and Encoding Strategies on Self-Supervised ECG Representation Learning
Self-supervised electrocardiogram (ECG) models are often trained on a few seconds of ECG signal and, increasingly, on discretized token sequences. It remains unclear whether these choices sacrifice information needed for rhythm inference and longitudinal consistency in real-world ambulatory recordings. We present a controlled study on the Icentia11k single-lead dataset that varies (i) the input horizon (16 seconds, 1 minute, 5 minutes, and 10 minutes) and (ii) the front-end representation (continuous convolutional patch embeddings vs. fixed vector-quantized tokens), while holding the Transformer backbone and training protocol constant. Representations are assessed by downstream abnormal rhythm detection and by patient-level retrieval that probes cross-session stability. Our results show that increasing temporal context beyond 16-second snapshots yields stronger transfer and higher retrieval accuracy, with the strongest performance achieved by the 5- and 10-minute models, indicating improved capture of slow-varying rhythm dynamics and individual-specific structure. Across all evaluated horizons, continuous patch embeddings outperform discretized tokens, suggesting that quantization can discard clinically relevant waveform detail. These findings motivate ECG foundation models that emphasize extended context and continuous encoders for clinical prediction and similarity-based applications. Our code and pretrained models are publicly available at https://github.com/muha-0/ecg-ssl-representation-learning.
Read the full abstract
Logic in Computer Science
Synchronous Observers Revisited for Runtime Verification of Lustre Using STL
Signal Temporal Logic (STL) is a popular formalism for the temporal safety properties of cyber-physical systems, most often used for runtime verification. In the synchronous family of languages, safety properties are instead expressed as synchronous observers, modules composed with a program for static verification, which are also runnable specifications suitable for runtime verification, though this use is rarely explored. We present a technique for compiling the synchronous fragment of STL (SSTL) into synchronous observers in the dataflow language Lustre. Unlike previous work, we allow arbitrary nesting of bounded SSTL properties via modular compilation, and admit a globally unbounded outer operator for online monitoring; the resulting observers serve both runtime verification and, as a by-product, static verification with the Kind2 model checker. We further contribute an interactive visualiser that renders a property's three-valued verdict over an editable trace, and evaluate on two case studies from the literature: a spring-mass system and a car-following cruise controller.
Read the full abstract
Numerical Analysis · Cross-listed from math.NT
Algebraic constructions of point sequences with quasi-uniform two-dimensional projections
Motivated by sequential space-filling designs for computer experiments, we study algebraic constructions of extensible point sets in the $d$-dimensional unit cube whose two-dimensional coordinate projections are all quasi-uniform. Our two constructions share a common parametrization in terms of finite configurations of distinct rational directions on the projective line $\mathbb{P}^1(\mathbb{Q})$. First, using a cubic number field, we construct explicit Kronecker sequences for which the mesh ratios of all two-dimensional coordinate projections remain uniformly bounded over every initial segment of length $N\ge 2$. Second, using a real quadratic field, a split prime, and a compatible $p$-adic embedding, we construct nested rank-1 lattice designs with the same uniform projection property at every nesting level. The proofs combine algebraic norm estimates with transference principles between simultaneous and dual Diophantine approximation, yielding lower bounds for the separation radii and upper bounds for the covering radii, both of optimal order in the number of points, uniformly over all coordinate pairs. We also investigate how the choice of rational projective coefficients affects the resulting mesh ratios. This leads to a minimax problem for finite configurations on $\mathbb{P}^1(\mathbb{Q})$, in which one seeks to minimize the maximum mesh ratio over all two-dimensional coordinate projections. These constructions provide extensible point sets with uniformly controlled bivariate geometry.
Read the full abstract
Computer Vision
Mr3D-VL: A generalist vision language foundation model for Multiparametric 3D Magnetic Resonance Imaging
Multi-parametric magnetic resonance imaging (mpMRI) is a cornerstone for brain tumor diagnosis and treatment, yet current AI models face critical limitations: their lack of natural language interaction and interpretability impedes spatial information integration and cross-modal reasoning required clinically. Key challenges arise from significant physical meaning differences across modalities, spatial misalignment due to scan intervals, and the need for complex multi-feature interpretation in tasks like glioma grading. While visual-language models (VLMs) show promise in cross-modal understanding, existing methods focus mainly on 2D image modeling, neglecting direct perception of 3D volumetric space. Although 3D VLMs have been proposed for report generation and feature alignment in 3D CT imaging, mpMRI applications demand collaborative inference across multiple imaging modalities-a requirement unmet by current solutions. To address this, we introduce Mr3D-VL, a dedicated visual-language foundation model for multi-parametric 3D MRI. With 4 billion parameters, it employs an unsupervised pre-trained shared 3D encoder and 4D rotational positional embedding for dual modality-spatial integration. Its cross-modal projection layer uses a multi-resolution feature implantation strategy to enhance feature perception across resolutions. Experimental results show significant improvements over existing 4B/7B/30B domain-specific and general-purpose models in text generation tasks, achieving a BERTScore of 0.856 for report generation, with question-answering accuracy at 0.713 and multiple-choice accuracy at 0.912.
Read the full abstract
Machine Learning
Finding the Needle in a Haystack: Test-Time Analog Circuit Representation Adaptation for Bayesian Optimization
Bayesian optimization (BO) is a sample-efficient framework for analog circuit topology search, where evaluating each candidate topology can require costly simulation. However, representation-based BO methods typically treat circuit embeddings as fixed after encoder training. This creates a mismatch between representation learning and optimization: embeddings learned to encode or reconstruct circuit structure are not necessarily organized according to the figure of merit (FoM) being optimized. This paper introduces Test-Time Analog Representation Adaptation for Bayesian Optimization (TTARO), an online deep-kernel BO framework that adapts circuit representations throughout the search process. Starting from pretrained circuit embeddings, TTARO jointly learns a nonlinear feature transformation and a Gaussian-process surrogate using the FoM labels of the circuits evaluated so far. Following each new evaluation, TTARO updates the representation and surrogate before selecting the next candidate. We compare TTARO with conventional Gaussian Process-based BO over fixed embeddings and with Deep Kernel Learning (DKL), which learns the representation only from the initial evaluated designs and keeps it fixed throughout the remainder of the search. By continually incorporating newly observed FoM labels into representation learning, TTARO aligns the search space with the optimization objective as BO progresses. In our experiments, TTARO reduces regret AUC by 15.2% on average relative to BO and by 20.7% relative to DKL across 40 encoder/kernel/acquisition settings, outperforming prior art in most settings with reductions as large as 46.7%.
Read the full abstract
Hardware Architecture
Spec-Driven Hardware Evolution via Executable Contract Refinement and Proof-Guided RTL Update
Hardware development is inherently evolutionary: major revisions typically begin by changing intended behavior and then updating a previously validated implementation, rather than regenerating RTL from scratch. Yet most recent LLM-based hardware research still frames the task primarily as prompt-to-RTL generation, offering limited support for semantic version evolution of trusted legacy designs. We present spec-driven hardware evolution, a contract-centered formulation for RTL version iteration. Instead of treating a new feature request as a direct prompt for RTL generation, we refine it into a reviewed executable contract for the next version. This contract specifies what must hold at the externally visible transactional level through a behavior-level reference together with explicit observation and checking semantics, while leaving how the change is realized in RTL to the evolution process. Based on this formulation, we organize hardware evolution into four stages: Specify, Plan, Implement, and Validate. After contract approval, the remaining stages proceed automatically: Plan derives cross-version semantic deltas and localizes affected RTL regions, aided by mutation-based semantic probing; Implement and Validate then perform legacy-aware RTL update under proof-guided checking and iterative repair. We evaluate the framework on a controlled version-evolution case study of a representative TPU datapath block under data-format changes. The results support the feasibility of contract-driven hardware evolution and demonstrate that the proposed backend workflow can effectively drive validated legacy RTL toward next-version functional convergence under a reviewed executable contract. An anonymous artifact for reproducibility is available at https://anonymous.4open.science/r/SDHE-3A6C.
Read the full abstract
Robotics
FUSE: Active Functional Affordance Grounding through Adaptive Semantic-Geometric Evidence Acquisition
Embodied agents must often identify and interact with objects based on their function rather than their identity, requiring them to actively acquire observations that reveal discriminative functional evidence. Existing affordance grounding methods operate from fixed viewpoints and lack mechanisms for deciding where to look when functional cues are occluded or incomplete. We introduce Active Functional Affordance Grounding, a new task in which an agent sequentially explores a scene to identify and spatially ground an object satisfying a functional query. To address this problem, we propose FUSE, an adaptive semantic-geometric evidence acquisition framework that combines explicit uncertainty-driven exploration with a learned amortized planner to efficiently select informative viewpoints. We further introduce a Habitat-based benchmark for evaluating active functional grounding. Experiments show that FUSE achieves the highest observed non-oracle grounding performance while reducing computation by 1.33x relative to fully explicit exploration, and remains effective across multiple affordance knowledge sources.
Read the full abstract
Machine Learning
Demand Transfer Estimation at Scale via Restricted Logit Modeling
Item demand forecasting is an integral component of store assortment optimization. Existing literature focuses on learning a suitable customer choice model and using this model to determine the value of an objective function (i.e. expected demand) with respect to an assortment proposal. However, for large item universe with many categories, this approach can prove inefficient, needing a separate demand forecast for every possible item assortment. An alternate approach exists whereby we combine the efficiency of forecasting item demand independently, while at the same time applying adjustments to the independent forecasts that account for the relations between item demand and the availability of other similar items on the shelf. Central to this approach is the estimation of Demand Transfer (DT) coefficients. These DT coefficients represent the percent of a particular target item's (item that the customer walked in the store to buy) demand that is redirected to each other item in the universe should the target item be removed from the shelf. We introduce an approach that allows us to compute these DT coefficients on large item universes (assortments having 1 million+ items). Experiments on data as well as historical transaction data for multiple locations within categories demonstrate that when certain reasonable assumptions about substitution behavior are satisfied, our procedure is able to accurately estimate underlying DT coefficients and lead to improvements in demand forecasting.
Read the full abstract
Artificial Intelligence
Beyond the Best Guess: Improving LLM Solution Coverage with Evolution Strategies
Large Language Models (LLMs) are increasingly deployed in discovery domains such as math and science. The usual approach is to present the problem to the model and use its answer as the proposed solution. However, beyond this best guess, discovery can be enhanced by increasing test-time compute. In a process called pass@k, the model is allowed to explore the solution space and generate diverse candidate solutions. Unfortunately, the standard approach to post-training LLMs through Reinforcement Learning (RL) may limit pass@k: the model's output distribution narrows around high-reward outputs, causing the solution coverage to collapse. The alternative is to use Evolution Strategies (ES), a population-based, gradient-free post-training method that optimizes directly in weight space through random perturbations. As this paper shows, ES achieves consistently higher pass@k than RL and produces a broader output distribution with greater solution coverage. This coverage in turn makes it possible to achieve better results in e.g. standard math benchmarks. Thus, ES provides a better foundation for post-training in discovery problems and other domains where diverse solution coverage is critical.
Read the full abstract
Discrete Mathematics · Cross-listed from math.CO
On the Gap of Finite Posets
Let $P$ be a finite nonempty poset with $n$ elements, let $f:P\to\{1,\ldots,n\}$ be a uniformly random order-preserving bijection, and put $h_P(x)=\mathbb{E}[f(x)]$. Aires and Kahn (2025) introduced $\operatorname{gap}(P)$ as the largest difference between consecutive values in the ordered list consisting of $0$, $n+1$, and all the expected ranks $h_P(x)$. Write ${w}(P)$ for the largest size of a pairwise incomparable subset. We prove three results. First, we prove a weighted strengthening of an ideal inequality conjectured by Kahn and obtain the explicit gap-width bound $\operatorname{gap}(P)\le 2 {w}(P)-1$. Second, for every $L>0$ we construct a width-two poset such that the expected-rank list of every maximal chain has a gap of at least $L$, with $0$ and $|P|+1$ added as endpoints. Finally, for every $r\in\mathbb{N}$, we construct a poset $P_r$ for which the relative order induced on every nonempty selected set $X$ has base-two entropy below $3|X|$, while $\operatorname{gap}(P_r)\ge(3/2)^r$. Thus the gap can be arbitrarily large while the induced order on every selected set has relatively small entropy. The key ideas behind all three results were found by ChatGPT 5.6 Sol.
Read the full abstract
Artificial Intelligence
The Role of Natural Language Understanding in Multimodal Video-Based Dengue Diagnosis
Detecting infection-related behavioral changes in mosquitoes from video data is challenging because mosquitoes are small, move rapidly and irregularly, and are affected by environmental factors such as background, lighting, and shadows, which can make reliable feature extraction difficult. In this study, a YOLO- and Contrastive Language-Image Pre-training (CLIP)-based vision-language framework is proposed to classify mosquito flight frames of uninfected and Dengue virus serotype 2 (DENV2)-infected mosquitoes. First, YOLO is used to isolate mosquito regions from the background. Then, visual features extracted from video frames are aligned with biologically meaningful textual prompts in a shared embedding space. The multimodal model was fine-tuned using supervised bidirectional contrastive learning and evaluated through frame-level image-text similarity-based classification. The results show that the proposed method achieved 98.54% accuracy and 99.91% sensitivity at the frame level. After temporal aggregation of frame-level information, the model achieved complete video-level performance. The ablation results showed that fine-tuning and CLIP-based representations were essential for this domain, while the textual branch provided semantic image-text alignment rather than an accuracy advantage over the vision-only model. These findings suggest that vision-language models can provide a useful framework for analyzing infection-related biological behaviors from video data.
Read the full abstract
Artificial Intelligence
Privacy-Preserving RAG by Concealing Sensitive Information from External LLMs
Retrieval-Augmented Generation (RAG) is widely used to improve the performance of Large Language Models (LLMs) in answering user queries. Existing privacy research on RAG has focused on preventing unauthorized users from accessing sensitive data. However, another important problem that is often overlooked in RAG privacy research is that external generators have access to the query and the retrieved documents, which may contain confidential information that could potentially be misused or accessed for unintended purposes. In this paper, we introduce the Sensitive Entity Alias Generator (SEAG), a privacy-preserving framework that empowers users to utilize powerful third-party generators without disclosing sensitive information. SEAG introduces a lightweight model that locates sensitive entities, generates corresponding aliases, and constructs an entity replacement table. The table is used to replace sensitive words in the user's query and in the retrieved documents before they are forwarded to an external generator. For this purpose, two datasets were constructed: one for fine-tuning SEAG models to generate entity replacement tables, and another for evaluating the entire SEAG framework. The experimental results demonstrate the success of the SEAG framework. As for the User metric, which measures the ability of the model to provide a correct response to the user while hiding sensitive information from the external generator, all SEAG models achieved over 80% accuracy. Additional analysis further evaluated the ability of SEAG models Qwen-3, LLaMA-3.2, and Phi-4 to hide all sensitive entities within given documents. The results show good performance with total accuracies of 77.83%, 76.73%, and 74.91%, respectively.
Read the full abstract
Artificial Intelligence
Lines and Ladders: A Context-Aware Multi-Agent Framework for Large-Scale Retail Price Taxonomy
Maintaining price consistency and executing an Every Day Low Price strategy is critical for global retailers. However, with catalogs spanning millions of active items, manual governance of price relationships is infeasible. Inconsistent pricing across item variants distorts customer value perception and cannibalizes sales. To address this, we present a scalable, context-aware Multi-Agent Framework designed to automate the construction of "Lines and Ladders" pricing taxonomies. Our framework employs specialized LLM agents to construct these coherent pricing structures by identifying key attributes, extracting multi-modal values, and applying hierarchical grouping logic. Evaluated on real-world enterprise data and deployed in production, our 3-Agent system achieves an F1-score of 0.83 for Lines, outperforming single-agent baselines by mitigating cognitive overload. The system achieves >90% precision and >75% recall in Food & Consumables, and 80.2% assignment accuracy in the unstructured General Merchandise catalog.
Read the full abstract
Artificial Intelligence
On the Expressive Power of Transformers
Multi-layer transformers form the critical component of essentially all large language models (LLMs) in use today. Because of their ubiquity and computational capability, there is a rapidly growing body of work that aims to precisely calibrate the expressive power of transformers as language recognizers by comparing them against standard models of computation studied for decades by the theoretical computer science community. In this endeavor, circuit complexity has by and large emerged as the "correct" branch of computational complexity to analyze the expressive power of transformers; the reason is that parameterizing transformers by the various resources they use, such as attention and precision, leads to direct comparisons with different classes of circuits parameterized by resources such as type of gates, size, and depth. Here, we present an overview of selected results that delineate the expressive power of transformers using concepts and methods from circuit complexity.
Read the full abstract
Artificial Intelligence
Designing AI Pipelines for Decision-Ready ITSM Intelligence
IT service management (ITSM) systems accumulate large volumes of heterogeneous ticket data that are difficult for sales and executive stakeholders to convert into actionable intelligence. This paper presents a sociotechnical AI pipeline, designed and evaluated following design science research principles, that transforms raw ITSM exports into a multilevel decision-support artifact. The pipeline combines LLM-based schema normalization, HDBSCAN sub-topic clustering, and hierarchical agglomerative clustering to generate executive-facing Main-topics and granular Sub-topics. A stakeholder evaluation across six artifacts and five raters from Sales Engineering and customer success roles shows that all four decision-support metrics, interpretability, actionability, trust, and likelihood of use, on average exceed 4.0 out of 5.0, with trust as the most consistent signal. The findings position ITSM analytics as an Information Systems (IS) problem of transformation, abstraction, and human-centered design.
Read the full abstract
Computers and Society
From Fair Representation to Just Recognition in Generative AI
The fair AI/ML literature has long distinguished distributive fairness, concerning how automated systems allocate resources and opportunities, from representational fairness, concerning how they shape the ways individuals and social groups are perceived, understood, and accorded social status. Generative AI is rebalancing these normative dimensions. Unlike predictive systems, large language models (LLMs) and related technologies are fundamentally expressive: their primary function is to convey meaning rather than automate domain-specific decisions. Representational harm has also become central to value alignment, especially in research on what and whose values and perspectives AI systems should represent. Existing approaches to harms in the representation of social groups often appeal to descriptive accuracy, but this strategy has important limitations. For many social groups, no stable or bounded referent exists against which representational accuracy can be judged. It is also unclear who has the authority to decide what counts as misrepresentation, while even accurate representations can reproduce harmful social patterns. The underlying problem, we argue, is therefore not simply misrepresentation but misrecognition. Drawing on political theory, especially Nancy Fraser's account of participatory parity, we show how moving from representational fairness to recognitional justice provides better conceptual and normative tools for governing central fairness challenges in generative AI.
Read the full abstract
Wednesday 12 August
Human-Computer Interaction
Short-Form Video Exposure and Academic Strain: A Dose-Response Analysis and Acquiescence Diagnosis
Short-form video platforms occupy a substantial share of undergraduate leisure time, raising concerns about sustained attention, deep learning, and academic well-being. However, much existing evidence relies on single-source Likert instruments with limited attention to response-style bias and measurement validity. This study analyses 863 university students who completed a 67-item, seven-construct battery assessing short-form content addiction, attention fragmentation, notification overload, cognitive fatigue, deep learning engagement, academic burnout, and self-regulation, alongside categorical exposure measures. A layered analytical protocol was implemented in Python, including descriptive screening, dimensionality assessment, reliability, convergent validity, careless-response diagnostics, and known-groups criterion validity. Daily exposure and in-study checking frequency showed consistent monotonic associations with maladaptive outcomes. Short-form content addiction exhibited the strongest effect ($\eta^2=0.117$, $\rho=0.273$, $p<0.001$), supported by nonparametric tests. However, the measurement analyses revealed substantial validity concerns: parallel analysis retained a single factor per construct, average variance extracted did not exceed 0.31, and a general agreement factor explained 59–89% of construct variance. Controlling for this factor reversed the association between addiction and deep-learning engagement from $+0.55$ to $-0.49$. These findings indicate that short-form video exposure is associated with cognitive and academic strain, while response-style contamination can substantially distort theoretically meaningful relationships. The results highlight the need for balanced item wording, embedded response-quality controls, and explicit response-style modelling in behavioural-addiction research.
Read the full abstract
Artificial Intelligence
MCPGuard: Finding Conformance Gaps in MCP SDKs via Dual-Track Analysis and Dynamic Adjudication
The Model Context Protocol (MCP) connects large language model agents to external tools and services through official SDKs that implement message validation and session logic. Specification constraints omitted by these SDKs may propagate to downstream applications, highlighting the importance of SDK-level conformance testing. Yet existing MCP conformance tools mainly exercise running endpoints and reveal little about whether SDK models enforce specification requirements. We present MCPGuard, an automated LLM-based framework for detecting specification-to-SDK conformance gaps. Its structural track aligns JSON Schema constraints with normalized SDK models and produces constraint-specific evidence. Its behavioral track extracts normative prose rules, retrieves bounded function-level slices, and audits them with LLMs for candidate omissions. An evidence-aware dynamic adjudicator then inspects complete source context and builds PoCs to validate candidates. A shared kernel with language-specific adapters enables analysis across SDKs. On the official Python 1.26.0 and TypeScript 1.29.0 SDKs, MCPGuard confirms 13 and 19 findings, respectively, none of which are exposed by two black-box baselines. Tests on open MCP servers further show that these gaps can systematically propagate downstream. By making SDK conformance of specification constraints directly testable, MCPGuard complements conventional endpoint-oriented conformance suites.
Read the full abstract
Human-Computer Interaction
Do the Numbers Match the Plot? Statistical-Visual Verification for Automated Exploratory Data Analysis
Automated exploratory data analysis (EDA) systems can execute statistics correctly while producing misleading interpretations. For example, correlations may be driven by influential observations, and mean differences may arise from skewness, overlap, or sparse subgroups rather than broad distributional shifts. We present EDAgent, a sequential framework that performs targeted visual verification before report synthesis. For each finding, the system records the statistical result, generates a visualization request identifying relevant distributional features, creates the corresponding chart, and uses the observed pattern to retain, qualify, or reconsider the interpretation. We call this process statistical-visual verification. EDAgent includes VisPlan, an 8B-parameter model that converts visualization rationales into detailed chart plans. We evaluate 87 end-to-end reports on EDA-Bench using a six-dimension rubric and GPT-5.1 multimodal evaluation. Maximum-score proportions reached 87.4% for factuality and visual consistency, 93.1% for statistical-visual cross-validation, and 96.6% for analytical narrative. On 166 open-ended VisEval tasks, VisPlan improved the mean overall score from 4.55 (Qwen3-8B) to 4.62, with the largest gain in aesthetics. These descriptive results indicate that visual verification can make automated EDA reports more transparent about how statistical findings relate to underlying data structure.
Read the full abstract
Artificial Intelligence
The Cost of Saying Hello: An Empirical Study of Handshake Overhead in AI Agent Communication Protocols
Every interaction between AI agents begins with a handshake: the initiating agent must locate its counterpart, learn what it can do, and establish a session before the first useful message can be sent. Surveys of agent communication protocols name handshake overhead as an efficiency criterion, yet none report measurements. This paper presents the first empirical, phase-decomposed benchmark of the handshake in three major agent protocols: MCP (in both its 2025 initialize and 2026 discover eras), A2A, and ACP. Using each protocol's official SDK, we instrument every request for round trips, per-phase latency, wire bytes, and token cost across cold-start, warm-repeat, capability-mismatch, and degraded-metadata scenarios (N=50 per cell, on loopback and under an emulated 50 ms RTT). Round-trip count dominates handshake cost. A legacy MCP handshake takes three round trips and 230 ms at 50 ms RTT; A2A needs one round trip and 61.0 ms. Round trips are not the whole bill: when an LLM reads the counterpart's metadata, handshake cost scales with the counterpart's capability inventory at 24-107 tokens per capability, so a 200-capability counterpart spends 11% of a 200k-token context window before the agent reads its own task. MCP's own 2026 revision cuts its handshake from three round trips to two and, with era pinning, to zero on warm reconnects, protocol evolution that corroborates the cost we measure. We release the full harness, raw data, and analysis at https://github.com/pvarshh/handshake-benchmark.
Read the full abstract
Information Retrieval
TAGR: Task-Adaptive Geo-Semantic Retrieval across Heterogeneous Urban Delivery Tasks
As urban residents demand higher-quality instant delivery services, demand forecasting, estimated time of arrival prediction, and next-stop prediction have become critical tasks in urban delivery systems. However, the three tasks rely on historical evidence with substantially different structures and granularities. Meanwhile, their prediction targets differ in output form, making it difficult to integrate retrieved evidence through a single fusion mechanism without causing mismatch or negative transfer. These issues can increase prediction errors and subsequently lead to imbalanced resource allocation, where some regions experience unmet demand. In contrast, others suffer from underutilization, ultimately degrading overall efficiency and service quality. To address these challenges, we propose TAGR, a Task-Adaptive Geo-Semantic Retrieval framework across heterogeneous urban delivery tasks. TAGR constructs an LLM-based geo-semantic shared prior and task-specific retrieval banks to share regional knowledge while preserving heterogeneous task-specific evidence. The LLM-derived regional representations encode geographic context and semantic relations among regions, providing transferable knowledge for all three tasks without requiring online LLM inference. It further introduces a task-adaptive retrieval augmentation module to accommodate heterogeneous output structures, thereby mitigating information mismatch and negative transfer caused by a uniform retrieval-fusion strategy. Experiments on three city datasets demonstrate that TAGR consistently outperforms state-of-the-art baselines across multiple tasks.
Read the full abstract
Human-Computer Interaction
From Community Values to AI Design Constraints: A Mixed-Methods Study of a Proposed AI-Driven Wildfire Risk Assessment Tool in Los Angeles County
AI-driven hazard tools are increasingly proposed for resident-facing risk communication. However, residents are often asked for feedback only after decisions about data use, privacy, and explanation have been made. This study takes an earlier approach by examining residents' values, concerns, and practical expectations for a proposed AI-based risk assessment tool before a predictive model or functional interface is developed. The proposed smartphone application would use property photographs to estimate parcel-level wildfire risk and provide an interpretable risk score, uncertainty information, and mitigation recommendations. Using a convergent mixed-methods design, we surveyed 30 Los Angeles County residents, seven of whom also participated in a virtual town hall. We analyzed the data using descriptive statistics, exploratory FDR-adjusted Spearman correlations, and inductive thematic coding of open-ended and town hall responses. Participants were cautiously receptive: 66.7% said they would be very likely or likely to use the tool. They linked fair risk assessment to whether the tool considered relevant property and neighborhood conditions. Privacy and data security were the most common concerns (64.3%), while cost was the main barrier to acting on recommendations (65.5%). Likelihood of using the tool was strongly associated with likelihood of saving or sharing risk results ($\rho = 0.752$, $p_{\mathrm{FDR}} < .001$). We translated the findings into the AI Value Map, a community-grounded artifact with nine value dimensions. The Map connects participant findings to provisional design requirements and possible consequences if they are not addressed. This study provides exploratory evidence on residents' responses to a proposed AI wildfire tool and offers a pre-model procedure for translating early community input into traceable design constraints for resident-facing AI hazard tools.
Read the full abstract
Human-Computer Interaction
Speech-muscle coupling at jaw and ear electrodes: a volume-conductor model with orientation and electrode-count controls
Objective. We compared how strongly speech-muscle sources couple to electrodes on the jaw and around the ear, and tested sensitivity to source orientation, electrode count, and anatomical detail. Approach. Using reciprocity in MIDA (116 labelled compartments, 500 micrometres), we evaluated 22 electrode positions against ten segmented articulators. Muscle was modelled as both source and conducting tissue. Comparisons used equal electrode counts and a hemisphere orientation sweep, with anatomy-derived fibre models for temporalis and lateral pterygoid. Results. Orbicularis oris, buccinator, mentalis, depressor anguli oris, and platysma favoured the jaw at every sampled orientation and tested electrode subset. Temporalis showed a 2.57 dB ear advantage under the uniform sweep, reduced to 1.15 dB with the derived fibre field. Its jaw-minus-ear interval [-1.45, +5.46] dB crossed zero; 50.5% of four-site ear subsets favoured the ear. The derived lateral-pterygoid interval [-5.33, -1.12] dB favoured the ear at the fixed jaw basis. Both derived constructions have unresolved anatomical limitations. Simplifying non-muscle soft-tissue conductivities preserved montage assignments while changing gaps by up to 3.27 dB. Significance. In this head model, jaw placement better captures lip and chin activity. Ear-side preferences depend on source assumptions and electrode selection. Targeted four-site placement improved lateral-pterygoid coupling by 1.07 dB over the median random ear montage. These coupling predictions require experimental testing.
Read the full abstract
Computers and Society
Perceived and Actual AI Deepfakes: The Case of Sudan
Sudan's AI deepfake risk currently appears driven more by demand-side vulnerabilities than the volume of AI deepfake content. While AI-generated disinformation in Sudanese feeds remains limited, perceived (alleged) AI deepfakes and general AI skepticism worsen the situation and contribute to general uncertainty in multimedia content. Within the analyzed cases, we found that the prominent stance was distrust driven more by motivated reasoning and contextual reliance than by durable AI detection skills. In practice, rejection or acceptance of AI deepfakes often reflects belief-driven judgment more than technical assessment. These weak defenses leave the information environment vulnerable to future sophisticated AI deepfake campaigns and risk normalizing truth indifference. Short-term interventions should prioritize modality-agnostic demand-side interventions such as pre-bunking, AI literacy, scaled fact-checks, and tiplines that strengthen public resilience across all forms of disinformation. More technical AI- specific and supply-side interventions can be developed as institutional capacity and technology allow.
Read the full abstract
Machine Learning
Fixed Points Without Fixed Diffusion: Implicit Neural Sheaves for Convergent Test-Time Computation
Implicit Graph Neural Networks (IGNNs) define node representations as fixed points of graph neural operators, enabling effectively infinite-depth propagation, iteration-independent parameterization, and flexible test-time computation. Yet these benefits depend on the equilibrium being unique and reached by fixed-point iteration. Existing constructions often impose constraints on recurrent updates to obtain these guarantees, limiting the transformations available at equilibrium. This raises a central question: can IGNNs gain expressiveness through richer, edge-dependent transformations while retaining the inherent strengths of their equilibrium formulation? We introduce SheafDEQ, a subhomogeneous deep-equilibrium architecture with adaptive neural-sheaf propagation. Its learned, matrix-valued sheaf restriction maps can align, mix, or reverse neighbouring representations. Under mild regularity conditions, we prove that SheafDEQ admits a unique equilibrium reached globally by fixed-point iteration from any positive initial state. Contractivity further guarantees convergence under bounded communication staleness. We evaluate SheafDEQ on distributed-inference tasks requiring repeated nonlocal aggregation and on community detection whose rewiring increasingly favours cross-community interactions. SheafDEQ improves over implicit baselines on Sums, MNIST Terrain, and Coordinates, and on community detection as connectivity becomes increasingly heterophilic. SheafDEQ remains competitive with recurrent models without equilibrium guarantees. Continued-iteration diagnostics show decreasing residuals and retained test accuracy through 100 iterations, unlike the finite-horizon control, while delayed-update experiments show low sensitivity to bounded communication staleness.
Read the full abstract
Machine Learning
Why Clipping Matters in AdaGrad? Toward a High-Probability Theory under Generalized Smoothness
We analyze the original same-step coordinate-wise AdaGrad under generalized smoothness and heavy-tailed noise with bounded variance. In this setting, local curvature may grow sub-quadratically with the gradient norm, and stochastic gradients are assumed to have only bounded conditional second moments. We show that unclipped AdaGrad can become anisotropically miscalibrated: under heavy-tailed noise, the adaptive denominator can learn the geometry of rare noise shocks rather than the local curvature of the objective, leading to a persistent directional distortion that blocks finite-horizon Euclidean progress. We then prove that clipping repairs this failure mode. Our main result is a finite-horizon high-probability guarantee for the original non-lagged AdaGrad update, yielding $\frac1T\sum_{t=0}^{T-1}\|\nabla f(x_t)\|^2=\mathcal{O}\left(\frac{d\big(\sqrt{\log T} + \log \frac{1}{\delta}\big)}{\sqrt{T}}\right),$ and hence $\widetilde{\mathcal O}(\varepsilon^{-2})$ complexity. This shows that, for AdaGrad under heavy-tailed noise, clipping is a structural stabilizer of the adaptive geometry rather than merely a robustness heuristic.
Read the full abstract
Machine Learning
Cosine Similarity Is Not Evidence: Measuring the Noise Floor of Interpretability Transfer Under Quantization
A statistic reported without the quantity needed to interpret it is not evidence. We develop that thesis for a concrete practice in AI safety. Interpretability artifacts are calibrated on full-precision weights, deployed on quantized ones, and certified as surviving the change by scale-invariant statistics (cosine similarity, correlation, AUROC) that are reported without their noise floor. For the difference-in-means direction estimator, the split-half floor is governed by one dimensionless number, $\kappa = n\rho^2/d$. The closed form $\mathbb{E}[\cos] \approx (1+4/\kappa)^{-1}$ is classical; the missing input is the class separation $\rho$, which we measure on real activations; no compression-transfer study we know of reports it. On Qwen2.5-1.5B-Instruct, $\rho = 33$–$61$ across depth, so two independent runs of the estimator agree to $0.978$–$0.994$ by sampling alone. A published cosine of $0.996$ between full-precision and quantized refusal directions therefore cannot be read as preservation without the $n$ it was computed at, which is not reported. Where $n$ is known, we judge each low-bit cosine against the split-half null measured within that quantized model, because a full-precision null assumes the low-bit estimator has the same variance. That assumption is exactly what a null exists to test. The result is plain: at INT4 the direction rotated, and the deficit exceeds the estimator's own noise. At INT8 we detect no movement, which is not an equivalence claim. We also show that a scale-invariant statistic cannot distinguish translation from attenuation of a transferred decision variable, although the two call for opposite remedies. We close with reporting recommendations that cost one forward pass. Code, data, and a one-cell reproduction are released at https://github.com/pvarshh/quantinterp
Read the full abstract
Software Engineering
IaC-Guard-V: A Verification Framework for LLM-Generated Infrastructure-as-Code Repairs
Infrastructure-as-Code (IaC) misconfigurations are a leading cause of cloud security incidents, and Large Language Models (LLMs) are increasingly proposed as automated repair agents. Yet the trustworthiness of LLM-generated IaC repairs remains underexplored. IaC presents distinctive verification challenges: security scanners may behave differently across tools and configurations, infrastructure semantics cannot be validated through ordinary unit tests, and provider-specific rules create a fragmented verification landscape. We present IaC-Guard-V, a verification-centered framework that evaluates AI-generated IaC repairs through four dimensions: syntactic validity, target-issue resolution, regression safety, and patch minimality. We construct a benchmark of 70 real-world misconfigured Terraform and Kubernetes artifacts spanning 70 unique scanner rules and eight violation classes, and evaluate three repair strategies across three LLM families in 630 runs. Although all models achieve 100% syntactic validity, only 32-50% of single-shot repairs pass full verification. Verification-guided iterative repair significantly improves verified-fix rates to 68-92%. An open-source model with verification-guided repair outperforms the strongest commercial model without verification at one-twelfth the cost per verified fix. Kubernetes repairs reach near-perfect rates for commercial models but require verification-guided iteration for the open-source model. Structured prompting consistently reduces Terraform repair quality, challenging common assumptions about constrained LLM output. The benchmark and artifacts are released for reproducibility.
Read the full abstract
Computation and Language
Framing by Wording, Framing by Selection: A Large-Scale Two-Dimensional Audit of French News Headlines, 2022-2025
News headlines frame public issues both by what they select and by how they word it, yet computational framing work typically collapses these operations into a single score. We introduce a two-dimensional framework that separates salience framing, measured through four wording devices (loaded vocabulary, blame attribution, threat framing, rhetorical question), from selection framing, measured through outlet-level story-form and high-charge distributions. We build a 10,000-headline French supervision set using three LLM annotators with majority-vote resolution and human arbitration, validate the labels against two annotator-independent blind human studies, and apply the strongest classifier to 902,111 deduplicated headlines from 25 French outlets (2022-2025). Three main findings emerge. First, salience and selection divergence are positively correlated yet leave nearly half of outlet-level variance unexplained, populating interpretively distinct off-diagonal cells in a four-cell outlet typology. Second, default classification thresholds systematically inflate corpus-level salience estimates; a precision-floor recalibration protocol corrects this distortion. Third, group-mention analysis reveals sharply unequal salience contexts: headlines mentioning Jews, the Far-right, and Muslims carry the highest detected salience rates, which broad event-context composition does not fully explain (residuals are descriptive, not same-event causal estimates; per-group lexicon precision is reported alongside). To our knowledge, this is the largest framing-focused French headline audit to date; we release the supervision set, lexicons, and analysis code.