Papers

Metric decompositions are a fundamental tool in the design of algorithms involving distances. We study fast algorithms for sampling from probabilistic metric decompositions of $n$-point sets in $\ell_\infty$ and $\ell_2$ spaces of high dimension $d$. For $\ell_\infty$, we design a padded-decomposition algorithm that runs in time $\tilde{O}(nd^2)$, which is near-linear in $n$, and achieves padding parameter $\tilde{O}(\log n)$. Our algorithm constructs a new sparse neighborhood cover that is based on geometric properties of $\ell_\infty$ [Indyk, JCSS'01], and utilizes recent reductions between covers and decompositions [Conroy and Filtser, STOC'25]. For $\ell_2$, we design a separating-decomposition algorithm that achieves near optimal separation $\tilde{O}(\sqrt{\log n})$ in almost-linear time $n^{1+o(1)}$. Our bounds improve over known algorithms with similar running time by a factor $\Omega(\sqrt{\log n})$, and the techniques have additional applications to spanners and nearest-neighbor search.

Current visual tokenizers in Multimodal Large Language Models (MLLMs) predominantly rely on patch-based partitioning, which causes severe semantic mixture and object fragmentation in remote sensing imagery due to the irregular contours of geo-objects. Moreover, existing adaptive methods struggle to extract precise object-level tokens and lack dedicated geometric positional encodings for irregular regions. In this paper, we propose HeatTok, a semantic-aware tokenizer driven by thermodiffusion aggregation. Inspired by the physical principles of heat conduction, HeatTok adaptively merges adjacent homogeneous regions to generate semantically independent, object-aligned irregular tokens. To enable MLLMs to perceive these irregular shapes, we design the Gaussian Multimodal Rotary Positional Embedding (G-MRoPE), which models token spatial distributions via 2D Gaussians and explicitly injects center, scale, and orientation cues. Extensive evaluations on the VRSBench and EarthVQA datasets demonstrate that HeatTok effectively preserves object-level semantic integrity and achieves state-of-the-art performance under a reasonable token budget. The code is available: https://github.com/YingyingYan1/HeatTok.

We design and analyze randomized strategyproof mechanisms for multi-facility location under the utilitarian social-cost objective, the sum of the agents' distances to their nearest facilities. For two facilities, the Pairwise-Distance mechanism locates facilities at a pair of reported locations sampled with probability proportional to their distance. It is strategyproof on Ptolemaic spaces, including Euclidean and Hilbert spaces as special cases, and has an approximation ratio of \(4\). The resulting Hybrid-Distance mechanism is a fixed-probability mixture: it selects the classical Proportional mechanism [Lu et al., EC'10] with probability \(\lambda^*=\frac{5+4\sqrt3}{23}\) and Pairwise-Distance with probability $1-\lambda^*$. It is strategyproof on Ptolemaic spaces and has a tight approximation ratio of \(\frac{74+4\sqrt3}{23}\approx3.5186\), breaking the long-standing factor-\(4\) benchmark of [Lu et al., EC'10]. We complement the two-facility results by studying more facilities. First, for \(n\) agents and \(k=n-1\) facilities, we introduce the Inverse-Square mechanism, which omits one report with probability proportional to the inverse square of its nearest-neighbor distance and locates facilities at all remaining reports. It is strategyproof on any metric space and has an approximation ratio of \(\Theta(\sqrt{n})\), improving the previous best-known ratio of \(\frac{n}{2}\) [Escoffier et al., ADT'11]. Second, for $k$ facilities on the line, we introduce the Gap-Product mechanism, which locates facilities at \(k\) reports and weights each set by the product of the gaps between consecutive selected reports. When \(k=3\), it is strategyproof and has a \(6\)-approximation, replacing the previous \(n\)-dependent guarantee [Fotakis and Tzamos, EC'13] by a constant, whereas it is not strategyproof for any \(k\ge4\).

Large Language Models (LLMs) increasingly support decision-making in high-stakes domains, but they often hallucinate and express confidence that is misaligned with factual correctness. Response-level confidence is a coarse signal: a single generation can mix correct and incorrect statements, so a single number is not actionable for users that must accept, reject, or verify individual pieces of information. We study claim-level confidence calibration as a decision-relevant uncertainty signal: each response is decomposed into atomic, verifiable claims, and each claim is assigned a calibrated confidence using inference-time signals from consistency across samples and self-verification. Our framework operates in closed-box settings (no logits, no fine-tuning) and applies post-hoc calibration directly at the claim level, enabling selective intervention such as evidence retrieval or human review for low-confidence claims. Across TriviaQA and TruthfulQA we evaluate seven baselines on six recent models (Llama-3.1, Mistral, Qwen2.5, DeepSeek-R1, GPT-4, GPT-4o), and show that claim-level decomposition combined with post-hoc calibration reduces expected calibration error on factual questions while exposing failure modes on adversarial false-premise questions where decision-makers most need reliable uncertainty estimates.

This work presents a concise, unified stability theory of high-order decoupled \lan{implicit-explicit linear multistep (IELM)} methods based on the well-known consistent splitting technique for the incompressible Navier-Stokes equation. With the help of the recent semi-generating function approach and the global discrete energy analysis, one can establish the unconditional stability of a consistent splitting IELM method with respect to the $\ell^{\infty}(H^1)\cap \ell^{2}(H^2)$ norm if the associated implicit-explicit controllability intensity is larger than $\sqrt{2}/2$, a constant determined by the Stokes pressure estimate. It is shown that the $\beta$-parameterized GBDF-$\rmk$ ($2\le \rmk\le5$) schemes and $\gamma$-parameterized SIELM-$\rmk$ ($2\le \rmk\le9$) schemes can fulfill this requirement of implicit-explicit controllability intensity by choosing proper parameters so that they can theoretically maintain the unconditional stability of the associated consistent splitting IELM methods. Numerical experiments are also included to support our theory.

A terminal-area descent procedure need to perform across the wind climatology rather than a single wind condition. Although flight demonstrations of the delayed deceleration approach (DDA) showed substantial fuel savings, DDA combined late deceleration with a steeper $3.77^\circ$ final descent, obscuring the contribution of each design choice. In this work, we propose the continuous-descent delayed deceleration approach (CDDA), which applies delayed deceleration to a continuous descent approach (CDA) profile without a level segment before glideslope intercept. A simulation-based stochastic optimization selects flap deployment trigger speeds and glideslope-capture distance to minimize expected fuel under wind uncertainty subject to a given stabilized-approach probability. An optimal control reduction limits the design space to a few hundred candidates, enabling exact expectation over a weighted wind grid using six-degree-of-freedom fast-time simulations. CDA and CDDA are optimized at matched final angles of $3.00^\circ$, $3.50^\circ$, and $3.77^\circ$ for the A319, B737-800, B767-400, and A340-300. Results show that deceleration architecture is the weakest factor, yielding 0.3–3.9% savings at $3.50^\circ$ and becoming material only for the B767-400 at $3.77^\circ$. Glideslope angle dominates fuel saving. At the $3.50^\circ$ Category D design maximum, optimized CDDA reduces expected fuel by 11-21% relative to optimized $3^\circ$ CDA, while flap-schedule optimization adds 2-17%. The DDA level segment acts primarily as a tailwind-robustness buffer, and the $3.77^\circ$ final exceeds the 1,000 ft/min stabilized-approach sink-rate element, limiting its near-term operational applicability.

Retrieval-augmented generation (RAG) enables LLMs to access external knowledge for answering knowledge-intensive questions. For complex multi-hop questions, multi-turn retrieval-augmented reasoning extends RAG into an iterative process that repeatedly searches for and integrates evidence across documents. However, existing reinforcement-learning (RL) approaches for agentic RAG are typically optimized with final-answer rewards, which provide sparse supervision and overlook whether the model actually retrieves the required evidence chain. We present GTA-RAG, a graph-trajectory-augmented RL framework for multi-turn retrieval-augmented reasoning. From an entity–document graph, we sample connected document paths, synthesize multi-hop QA trajectories, and validate them with the deployed retriever to obtain executable trajectory-level supervision. We then optimize the retrieval policy with Group Relative Policy Optimization (GRPO) and a trajectory-guided reward that encourages both accurate answers and acquisition of target evidence documents, followed by answer-reward training on natural QA instances. Experiments on three multi-hop and two simple QA benchmarks show that \method{} consistently outperforms RL-based RAG baselines with both Qwen2.5-3B and Qwen2.5-7B backbones, while substantially improving evidence-chain coverage. Our code is available at https://github.com/cjcj46262/GTA-RAG.

Implied volatility surfaces summarise the option market and are central to many financial applications. Probabilistic forecasting of their future evolution requires modelling two-dimensional geometry, temporal dependence, and predictive uncertainty while preserving economic admissibility. We propose a conditional latent diffusion framework for generating joint 30-step trajectories of implied volatility surfaces and underlying returns. An arbitrage-aware autoencoder learns a surface representation, while diffusion captures the evolution. Evaluated on SPX surfaces, the framework generates realistic probabilistic multi-step scenarios. Its predictive mean also outperforms persistence beyond the shortest horizons.

Recent work has shown that LLMs can synthesize highly specialized database systems for fixed workloads, but existing approaches typically assume static data and replace the database's native storage with generated representations. We present JetStream, a system for generating query-specific accelerators that instead extend an existing DBMS. JetStream consists of three parts. First, a staged, measurement-driven agentic workflow generates and optimizes query-specific accelerators, including persistent auxiliary state when beneficial. Second, a fixed, engine-neutral substrate provides the common interfaces for execution, transaction coordination, state management, and maintenance. Third, a separate synthesis workflow generates engine-specific backend adapters that connect the substrate to the underlying engine. For stateful accelerators, JetStream also generates maintenance logic and uses a runtime policy to choose among incremental maintenance, rebuilds, and lazy repair as the database changes. On TPC-H at SF=20, stateful accelerators generated by JetStream achieve an 833x geomean read-only speedup over DuckDB, compared with 34.07x for GenDB and 12.35x for Bespoke OLAP. Under TPC-H refreshes every 60 seconds, JetStream maintains a 375x geomean workload speedup. We also show that JetStream generalizes to new, unseen workloads, achieving geomean read-only speedups of 102x over DuckDB and 486x over PostgreSQL. These results show that aggressive generated specialization can be integrated seamlessly with existing DBMSes and support dynamic workloads.

We study parameterized quantum circuits (PQCs) integrated as residual sidecar modules within a ResNet-50 backbone for 31-class neural population decoding—imagined handwriting classification from multi-neuron spike rasters. Under strictly controlled conditions (fixed data splits, seeds, and optimizer), we compare four model variants: baseline, quantum sidecar with frozen input projection, quantum sidecar with backbone-gradient-trained projection, and a measurement-guided variant that aligns angle encodings with circuit measurement outcomes. The backbone-gradient variant improves accuracy in 3/4 seeds (+0.19% mean, 95% CI [-1.10%, +1.48%]) and consistently reduces Linear CKA similarity to baseline features ($\Delta=-0.025$, 4/4 seeds), indicating genuine structural reorganization of representations. A nine-variant ablation identifies simple shallow architectures as the most effective and reproducible configuration. Measurement-guided training consistently improves representation geometry without reducing accuracy. All results use noiseless statevector simulation on 4 qubits, a regime chosen to reflect the practical constraints of current near-term superconducting hardware; no quantum computational advantage over classical methods is claimed.

The quality of data is crucial for both practice and academia, and this holds for descriptive statistics, AI and advanced analytics alike. This study addresses an apparent gap in the literature and as such presents a full and theoretically grounded overview of data quality assessment approaches and methods, as well as their defining characteristics and inter-relationships. For this purpose a broad typology is introduced that employs two theoretical dimensions, namely the evaluation logic (formal versus informal) and the assessment driver (norms versus data). This yields four high-level approaches and 15 methods for assessing data quality. The research results are relevant for academia, as they provide a theory-based overview and definition of the ways that data quality can be evaluated. The study is also relevant for practice because it allows professionals to make informed decisions on using these methods, e.g. as part of an audit or broad data quality assessment strategy.

Function calling represents the core capability of agentic large language models (LLMs). Existing research has focused on enhancing LLMs function-calling accuracy through fine-tuning, reinforcement learning (RL), and multi-agent frameworks, particularly for native function-calling LLMs. This work demonstrates that LLMs achieve superior accuracy in function calling in instruction-following contexts (i.e., standard user-assistant interactions) rather than a tool calling context. We introduce Instruction-Followed Function Calling (IFFC), a novel framework that decouples function-calling logic from the primary LLM and delegates it to a dedicated smaller model operating within the instruction-following paradigm. Our method consistently outperforms both native function calling (NFC) and prompt-based function calling (PFC) baselines, with particularly strong gains on reasoning-oriented LLMs. Furthermore, we demonstrate that IFFC maintains robust performance under aggressive quantization, enabling efficient on-device deployment without significant accuracy degradation. This work establishes a new paradigm for reliable, resource-efficient function calling in edge-computing scenarios.

Long-horizon language-model tasks — multi-step reasoning and tool-using agents alike — are limited by credit assignment. We analyze it through the policy variance $\sigma_\pi^2(s)=\operatorname{Var}_{a\sim\pi}[Q_{\pi}(s,a)]$, which in a deterministic MDP is the sole source of return variance and is injected in discrete pulses at states we call critical forks. Three results follow. (i) Policy variance is a discovery budget: observing an action of advantage $c$ requires $\Omega(c^2/\sigma_\pi^2(s))$ draws, a bound that is exact on the canonical two-point fork. (ii) Policy variance is bounded by the policy's Gini dispersion, $\sigma_\pi^2(s)\le 1-\|\pi(\cdot|s)\|_2^2$, a rollout-free necessary condition for criticality computable from logits alone. (iii) The remaining horizon sets the estimation cost: at a fork whose downstream success probability is $P$, the Monte Carlo advantage estimate has signal-to-noise ratio of order $\sqrt{P}$, so its sample cost scales as $1/P$ — a cost that branched sampling shares. Bootstrapping removes it by converting a product of survival probabilities into a sum, provided the value representation is multiplicatively accurate, which argues for log-value parameterization.

An effective open phase detection is one of the unsolved challenges for utilities. Open phase detection is expected to detect the broken power line before it falls to the ground, to prevent the ignition of wildfires. The existing open phase detection methods are normally based on abnormal currents or voltages due to three phase unbalances after open phase events. For a complicated distribution feeder with multiple branches, current-based methods usually have a limited protection range, and voltage-based methods possibly failed due to the recreated voltage on the open phase due to back-feed. This paper presents a new open phase detection method based on the connection impedance of two measuring points, which is calculated with the synchronized phasors. Under an internal open phase condition, the calculated positive-sequence impedance is much larger than the impedance under other conditions, and its sign is negative. This new method is applicable for a distribution feeder with a closed-loop configuration, or DER units connected at the load side. It is almost not impacted by back-feed. This new method only requires point-to-point communication. This method can be easily implemented in the existing recloser controllers or RTUs. It is scalable, cost effective, and reliable.

High-fidelity free-viewpoint video (FVV) and interactive rendering increasingly rely on explicit Gaussian representations, yet practical deployment remains constrained by representation size, dynamic updates, and computational cost. Existing multi-view video benchmarks provide valuable real-captured content, but they make it difficult to isolate the effects of controlled camera geometry, representation efficiency, and temporal redundancy. We introduce M$^3$ISR, a controlled synthetic benchmark for 3D and 4D Gaussian Splatting (3DGS/4DGS). The benchmark contains 25 scenes from five indoor and outdoor scene groups, two camera/motion configurations, six synchronized 1080p views, and dense ground-truth annotations including RGB, camera parameters, depth, semantic and instance segmentation, and static–dynamic masks. The shared-center camera design intentionally isolates angular view variation and enables controlled evaluation of novel-view synthesis and representation efficiency. We organize M$^3$ISR into five complementary tracks covering 3DGS synthesis, 4DGS synthesis, 4DGS streaming, 3DGS compression, and 4DGS compression. Representative baseline results show small differences in static reconstruction quality but substantial differences in representation storage, while the evaluated streaming methods exhibit substantially higher reported training or reconstruction cost than the corresponding offline dynamic reconstruction baselines. We further define feedforward compression tasks for 3DGS and 4DGS and provide reference rate–distortion formulations and preliminary baseline evaluations. The benchmark is intended as a controlled and complementary testbed for systematic study of Gaussian-based FVV reconstruction, compression, and streaming.

Neural networks can acquire new capabilities while damaging existing ones, but what determines whether new learning persists remains unclear. We identify functional compatibility, the extent to which incoming learning can coexist with behaviour that must be preserved, as an experimentally manipulable causal determinant of persistence. From identical neural states, we vary compatibility while matching unrestricted learning opportunity and imposing a common retention requirement. Persistent learning increases with compatibility across independent directions, convolutional and transformer architectures, vision and text, and a ten-seed replication. Learning rules and retention constraints determine how much compatible opportunity is retained, whereas nonlinear geometry limits the matched intervention at larger update norms. Functional compatibility therefore reframes stability-plasticity from preventing forgetting to determining which new learning can coexist with existing function and persist.

Public mass-shooting databases differ substantially in coverage, feature availability, and reporting practices, creating challenges for machine-learning models that must generalize across data sources. We introduce MASH-Bench, a harmonized benchmark of 6,968 incidents from four U.S. databases: Kaggle, Mother Jones, Stanford MSA, and the Gun Violence Archive (GVA). We evaluate cross-source risk classification using leave-one-dataset-out (LODO) evaluation. Random Forest, XGBoost, and LightGBM achieve VeryHigh-risk recall of 0.68-0.89 on the curated sources but generalize poorly to GVA, where mean recall drops to 0.20 and precision to 0.0004. To investigate the source of this degradation, we conduct a controlled feature-masking ablation that removes the five features unavailable in GVA from the curated sources. The resulting recall collapse to zero provides evidence that feature completeness is a major contributor to the observed cross-source failure. We further evaluate three domain-adaptation approaches: DANN, CORAL, and importance weighting. DANN improves VeryHigh-risk recall on GVA by 0.282 (95% CI [0.11, 0.47], p = 0.003), although precision remains low, whereas CORAL and importance weighting yield zero recall. Oracle prior-shift recalibration likewise fails to recover VeryHigh-risk predictions, indicating that label-side correction alone is insufficient under the observed feature deficiencies. A per-group audit further identifies substantial disparities associated with media-attributed mental-health labels. Overall, these results indicate that, in MASH-Bench, cross-source generalization is constrained more by feature completeness and label prevalence than by classifier choice. The benchmark provides a controlled setting for diagnosing these effects in cross-source risk classification.

Workers are increasingly asked to adopt AI systems to assist their work, yet are rarely given a voice in defining what meaningful AI augmentation should look like or how to evaluate for it. In this paper, we propose worker-driven AI measurement—a bottom-up approach to AI evaluation where workers collaboratively shape decisions about which tasks AI should augment, what "successful" augmentation looks like, and how it should be measured. We explore how to support this through a case study with 19 workers from a local school social work organization. Through a series of eight workshops, workers iteratively develop their own measurement goals for AI evaluation, systematize these goals, and then design a benchmark to capture how effectively an LLM can "challenge" them to reflect on their own assumptions and biases in the context of their day-to-day work. Workers collaboratively design and refine an LLM-as-a-judge rubric based on their professional and lived expertise. In validations of the worker-created benchmark, we find that there is strong agreement between worker and LLM judge ratings and that the resulting benchmark can differentiate performance across six state-of-the-art LLMs. Based on our case study, we discuss opportunities for future work to support worker-driven AI measurement as a complementary approach to existing top-down AI evaluation approaches.

Two properties of optimization-based controllers such as model predictive control (MPC) limit their use in space flight. The online computing time varies and can exceed the sampling period, and the closed-loop behavior carries no formal guarantee. Explicit MPC, where a parametric solution of the optimization problem is computed, replaces the on-board optimization with a lookup table of piecewise-affine functions with a fixed execution time. However, it has been regarded as only applicable to small horizons, since the number of regions bounds the memory the table occupies and admits only an exponential bound in the horizon. This paper builds on recent developments in parametric solutions of quadratic programs and nonconvex reachability analysis to provide an entire pipeline for spacecraft rendezvous controllers under the Clohessy-Wiltshire dynamics. For this application, efficient data structures can be constructed to encode the control law, both from a computing time and a memory standpoint. In addition, the controller can be verified offline with a guaranteed closed-loop stability certificate by resorting to reachability analysis with hybrid zonotopes. At a 51-step horizon the rendezvous law occupies 10.5 MB and is evaluated in under 300 ns. The occupied memory would fit for instance the CubeSat flight computer ARM9-class AT91RM9200 meaning that the horizon limitation does not hold when the partition is built with the parametric solvers and data structures shown in this paper.

An inner hood panel must meet a deflection target, stay below a stress limit, and hit a mass target. Machine-learned surrogates have made the forward direction, geometry to performance, fast and routine. The inverse direction, producing geometry from a stated requirement, remains largely unaddressed for industrial parts whose design space is organized into discrete topology families rather than a continuous parameterization. This work presents a two-stage pipeline for that inverse problem. A reachability stage determines which topology families can satisfy a given requirement vector. A conditional variational autoencoder then generates point-cloud geometry within a selected family, and a neural-operator surrogate estimates the performance of each candidate. The pipeline is built entirely from public data and freely available compute, and is deployed as an interactive tool. The pipeline works, with qualifications that are reported as primary findings rather than caveats. The surrogate is accurate in aggregate, but its error is comparable to the performance differences it is asked to discriminate, which bounds what can be claimed for any individual generated design. That ratio of surrogate error to within-class signal is argued to be the quantity that determines whether a pipeline of this kind can work at all.

We propose a holonomy-based framework for discretizing curvature on graphs equipped with local symmetric positive-definite metrics. Each vertex carries a fibre metric \(g_i\), and each directed edge carries a reversible metric-compatible transport \(F_{ij}\). The ordered product around an oriented triangular loop \(\mathcal C\) gives a holonomy \(H_{\mathcal C}\), whose normalized logarithm \(\Omega_{\mathcal C}=-s_{\mathcal C}^{-1}\operatorname{Log}(H_{\mathcal C})\) is used as a finite-loop curvature observation. Thus the construction discretizes the geometric principle that infinitesimal holonomy is controlled by curvature, rather than treating holonomy as a heuristic feature. Since \(\Omega_{\mathcal C}\) lies in the \(g_i\)-orthogonal Lie algebra, it is not itself a velocity of an SPD metric. We therefore introduce two aggregation mechanisms: a commutator with a symmetric response matrix, producing symmetric Ricci-type metric responses, and an incidence-aware covariant divergence of curvature-induced edge fluxes, reflecting the relation between trace and covariant divergence. The resulting responses are locally orthogonal-gauge equivariant and can drive exponential updates that preserve positive definiteness. We also give a reversible metric-compatible parametrization of edge transports, allowing orthogonal edge factors, loop scales, weights, and response matrices to be learned while respecting the graph geometry. Known-geometry calibrations on the unit sphere test the holonomy–curvature relation, curvature preservation under nontrivial local metric representations, and the empirical recovery of edge transports from local observations.

Surprisal, the negative log-probability a language model assigns to a word given its preceding context, reliably predicts adult reading times. Does it contribute as much to explaining when children acquire individual words? Frequency reflects a learner's cumulative exposure to a word, whereas surprisal reflects how predictable a single occurrence is given its context. We investigate this question across two corpus-based studies of Spanish. In Study 1, we modeled age of acquisition (AoA) for 225 Spanish nouns using lexical frequency and contextual diversity from child-directed speech, plus surprisal from three language models differing in architecture and training language (BETO, BERTIN, mGPT). Frequency strongly predicted AoA (r=-.597, p<.001); surprisal added little beyond frequency and word length, including in a naturalistic-context analysis. In Study 2, we modeled adult fixation durations in the Chilean Spanish subsample of the Multilingual Eye-movement Corpus (MECO Wave 2), using mGPT surprisal alongside two independent frequency measures. Surprisal robustly predicted longer fixation durations after controlling for frequency and word length, consistent across both frequency sources. A matched word-type-level comparison showed the surprisal-behavior association was stronger in reading than in acquisition (z=3.63, p<.001). The findings suggest cumulative lexical exposure and contextual predictability play different roles across the language trajectory: frequency is particularly informative about when early lexical representations are acquired, whereas surprisal captures moment-to-moment processing difficulty in an already-established linguistic system. We discuss this pattern in relation to usage-based and entrenchment-based accounts of lexical development and to the evaluation of language models as models of human language behavior.

Forecasting dexterous hand motions from egocentric observations is fundamental to intelligent interactive systems. Existing VLM-based methods typically map observations directly to future motions, overlooking the underlying manipulation process that governs hand-object interactions. Moreover, end-to-end optimization couples manipulation learning with motion synthesis, causing motion-generation gradients to interfere with the pre-learned manipulation-aware representations. To overcome these limitations, we propose EMPIRE, a two-stage framework that introduces Explicit Manipulation Planning as an Intermediate Representation for Egocentric hand-motion forecasting. Stage I: Learn to Plan. EMPIRE first learns explicit manipulation plans from multimodal context to capture the progression of hand-object interactions. Stage II: Learn to Act. A motion generator synthesizes future bimanual hand motions conditioned on frozen planner representations, preventing motion-generation gradients from affecting manipulation planning. To support our method, we further construct EMPIRE-651K, a bimanual hand-motion forecasting dataset comprising 650,910 training windows across 111 tasks, each paired with an explicit per-hand manipulation plan. Under identical training and evaluation protocols, EMPIRE achieves state-of-the-art forecasting accuracy, with an MPJPE of 84.53 mm and a finger-relative error of 38.97mm. We release the code and dataset at https://github.com/wangwen-banban/EMPIRE.

Metaphors are figurative use of words for conceptual mapping. Metaphor detection in the legal context has been crucial as metaphors are persuasive juridical means of creating legal meaning and concepts resulting in significant consequences. Metaphorical framing in legal discourse by judges, lawyers, and legislators brings about real-time implications upon individuals and influences judicial decision-making, argumentation and interpretation of laws. This is crucial in Human Rights infringement cases where language determines severity of punishment, public perception and judicial outcomes. While automatic metaphor detection in major languages like English, Spanish, Polish, Lithuanian have aided in understanding inherent intentions of metaphorical use of language, there is no such attempt in low-resource languages like Hindi. The dearth of annotated legal corpora in Hindi makes it difficult to develop NLP models and detect metaphors in judicial proceedings. In the Indian context, Convolutional Neural Networks (CNNs) have been used for classification of bail judgements, however there are no existing models designed for metaphor detection. We present a Hindi Legal Metaphor Corpus (HiLeMe) by isolating judgements from Hindi Legal Data Corpus (HLDC). Legal experts annotated HiLeMe to classify metaphorical constructions using the MIPVU schema. We downstreamed an mBERT on Hindi legal metaphor detection task. We built a transformer-based architecture for metaphor detection that are known to outperform traditional models in legal classification tasks. This model provides insights into the judicial psyche for decoding judicial decisions. Our research contributes to advancing automated models in legal discourse in low-resource languages like Hindi and envisages adoption into 22 Indian schedule languages.

We propose and analyze an IMEX-SDIRK2 mean-reverting concurrent-correction scalar auxiliary variable (mr-ccSAV) method for the forced two-dimensional periodic Navier–Stokes equations in vorticity form. The viscous term is treated by Alexander's SDIRK2 method and advection explicitly. Each stage requires two elliptic solves with the same shifted Laplacian and the solution of either a cubic or a linear scalar algebraic equation. For initial vorticity in $\dot L^s(\Omega)$, $s>2$, a stage solution exists for every positive time step; uniqueness is established separately under an explicit small-step condition. The principal result is a direct, unconditional uniform-in-time enstrophy bound for arbitrary positive time steps. For persistently bounded forcing, this estimate is absorbing: the influence of the initial data decays, and the forcing contribution does not accumulate in time. Under additional regularity, uniformly bounded step sizes, and bounded neighboring step ratios, we also establish uniform-in-time $H^1$ and $H^2$ vorticity bounds without a small-step condition. For smooth solutions, the method converges optimally at second order. Numerical experiments confirm its accuracy, long-time robustness, and effectiveness of a companion embedded time-step selector.

AI safety evaluations typically assess models one at a time. But language-model agents increasingly operate in populations, where they read and respond to one another's decisions. A model that makes reliable decisions on its own may behave differently when surrounded by other agents. Can individually reliable agents be collectively manipulated, and can we predict how strongly the population will respond? We study this question in a security-triage task, where populations of LLM monitors decide whether to escalate or dismiss alerts. We introduce a committed minority of agents that consistently advocates for one decision and measure how its presence shifts the population's behavior. We find that alerts judged almost identically by an isolated agent can produce sharply different collective outcomes. Individual audits therefore cannot reliably predict population behavior. Yet collective vulnerability is not unpredictable: using only benign, adversary-free behavior, we calibrate a response function that forecasts how strongly a population will shift under a future attack, before the attack is run. We next examine how interaction and time shape these outcomes. Allowing agents to observe one another's reasoning can weaken adversarial influence, but does not reliably prevent capture. However, capture need not be permanent. After adversaries are removed or replaced, mean population behavior returns toward its original level, even after five or thirty rounds of pressure. These findings are consistent with a calibrated single-stable-state model, in which adversarial pressure shifts the operating point without creating a second self-sustaining state. The same calibration predicts attack responses in populations of 24 and 48 agents. Alignment in isolation does not guarantee alignment in a population. But a population's behavior before an attack can reveal how it will respond when one arrives.

This work investigates the uncertainty-aware joint pricing and matching problem for dynamic high-capacity ride-sharing services, where passengers are assumed to be price-elastic and decide whether to accept a ride-sharing offer based on the upfront prices provided by the platform. We formulate the studied problem as a two-stage stochastic program, where the first stage optimizes upfront price decisions for passengers, and the second-stage recourse problem captures passenger-vehicle assignment based on passengers' uncertain choices. To enhance computational efficiency, we introduce a novel relaxation-based gradient descent-guided search algorithm that leverages the problem's structural properties. Initially, the algorithm generates a feasible solution for the first-stage problem via relaxation. It then iteratively improves the solution via a search process guided by the derived gradient information. In particular, scenario reduction is applied to eliminate unnecessary scenarios when calculating the gradient, thereby reducing the overall computational burden. Numerical experiments demonstrate that, compared to solving the stochastic program directly, the proposed algorithm can accelerate computation speed by thousands of times while achieving optimality gaps of no more than 1.1%. Finally, we validate the benefits of considering passengers' choice uncertainty through large-scale simulation using real-world datasets and road networks over two large cities. The results demonstrate that, on average, the proposed method can increase the revenue by 5.2% and the service rate by 8.2% compared to the baseline approaches. This study provides a valuable reference for transportation network companies to design pricing strategies for ride-sharing to enhance service efficiency and improve revenue.

Persona-conditioned large language models (LLMs) are increasingly used to simulate survey responses across diverse domains. However, apparent response variation can reflect unconditioned model priors or token sampling noise rather than systematic persona conditioning. We argue that persona-conditioned variation is informative when semantically similar personas exhibit concordant response shifts. To operationalize this principle, we introduce Persona-Conditioned Informativeness (PCI), an unsupervised diagnostic metric that measures whether semantically similar personas deviate in concordant directions relative to item-level sample baselines. By modeling personas as a similarity graph, PCI uses Local Moran's I to quantify local spatial coherence and extract compact persona subsets without using construct labels. To evaluate PCI without external human benchmarks, we test its ability to recover established latent value structure using the 57-item Portrait Values Questionnaire-Revised (PVQ-RR). Confirmatory factor analysis (CFA) shows that a PCI-selected 10% subset substantially improves overall construct recovery relative to response-stability and random selection. These findings support PCI as a principled internal diagnostic for screening synthetic respondents in survey pipelines.

In this paper, we propose an extension of the well-known Robinson implicit function theorem for generalized equations from the upper Lipschitzian case to the upper Hölderian case. Explicit exponents dependence between the generalized equation and its linearization is determined. Applications to ball constrained least squares problems, including linear least squares and separable nonlinear least squares, are studied. In particular, we establish that the solution mapping of ball constrained linear least squares under linear perturbation is locally upper Hölder continuous with exponent $1/3$, which is of independent interest. Ultilizing the upper Hölderian version of the implicit function theorem, we show the local upper Hölderian of the solution mapping of parametric ball constrained linear least squares.

Multilingual LLM judges produce different evaluator-backbone rankings depending on the prompt language: on an eight-language Agent-as-a-Judge benchmark, the top-ranked backbone alternates across English, Arabic, Chinese, Hindi, Japanese, Spanish, Turkish, and Swahili, and 7 of 15 backbone pairs show statistically significant pairwise rank reversal. We treat this as a measurement problem. The multilingual judge score decomposes additively into task difficulty, backbone skill, and a language-backbone interaction term, the last of which is recoverable without human labels by double-centering the cell-mean score matrix. We make this estimator (Consensus-Based Calibration, CBC) explicit, give an $O(1/\sqrt{n})$ finite-sample concentration bound with variance constant $(1-\tfrac{1}{m})(1-\tfrac{1}{k})$, and show that it is unbiased even when task-language interactions are present. Across 7{,}920 judge runs (6 backbones, 8 languages, 55 tasks, 3 frameworks), CBC raises held-out cross-task rank consistency $\tau$ from 0.650 to 0.902 and agrees with the held-out additive-model oracle in 100% of per-language decisions versus 68.5% raw; these are consistency diagnostics, not human-grounded correctness measures. On a separately collected M-RewardBench panel (7 languages, 1{,}500 items per language, 10{,}500 language-item instances, 5 evaluators), panel agreement with the public human gold preferences rises from 68.7% to 76.6% (gain 7.9 percentage points, 95% CI $[6.0, 9.9]$), our strongest external evidence of downstream usefulness. The estimator is the standard two-way ANOVA interaction-recovery operation under sum-to-zero contrasts; our contribution is its application as a label-free post-hoc calibrator for multilingual LLM judges, an explicit finite-sample concentration bound, and an unbiasedness result that holds even under task-language misspecification.