Papers

PLC developers reuse existing components and third-party libraries to reduce the cost of developing new applications, but this reuse can also introduce security risks. When components are distributed only in binary form, provenance tracing, vulnerability assessment, and security auditing become more difficult. To support these analyses, binary2source matching can retrieve the corresponding Structured Text (ST) source file for a PLC binary from a candidate source repository. However, due to the properties of PLC software, this task faces three challenges: cross-platform compilation heterogeneity, a binary-source representation gap, and inadequate representations of PLC semantics. Therefore, we present PLC-Bin2Src, a cross-platform binary2source matching framework that retrieves corresponding ST source files for binaries produced by CODESYS, GEB, OpenPLC v2, and OpenPLC v3. First, PLC-Bin2Src uses platform-aware binary frontends to recover user control logic across toolchains. Then, ST-to-C conversion and shared normalization are used to reduce syntactic differences between source- and binary-side representations. Next, control-data flow graphs (CDFGs) and function call graphs (FCGs) are constructed to capture program semantics, while recovered symbols provide complementary identity evidence. Finally, PLC-Bin2Src compares these representations and combines their similarity scores to rank source candidates. We evaluate PLC-Bin2Src on PLC-BEAD. The results show that PLC-Bin2Src achieves 95.89% Recall@1, 99.66% Recall@5, and an MRR of 0.9769 across four PLC platforms.

Class disentanglement–the separation of a representation's class-conditional point clouds along depth and over training–is measured by descriptive curves: the sentence such a study wants to write, layer l+1 is more disentangled than layer l, is an eyeball judgement with no null. We supply the inferential layer for a topological measurement of class overlap, the Intersection Euler Characteristic Profile: the Euler characteristic of the overlap of the clouds' ball unions as a function of scale, from one Alpha-complex sweep with no boundary-matrix reduction. Every number carries a test–exact permutation tests in both directions, a guarded separation certificate the invariant requires, and a paired sign-flip test for comparative claims. Building that test taught a lesson outliving this invariant: its statistic must be scale-free. On the raw profile mass, which has units of feature length, 12,375 paired tests return 5,633 significant steps of which every one at the first epoch points the wrong way, certifying feature-norm dynamics as disentanglement; the dimensionless statistic returns 2,707, with 2,026 decreases. Across 111 networks and 52,650 measurements, disentanglement is depth-graded and early, and interaction quotients rank class pairs by confusability (rho=0.83), on par with cheap separability statistics. In a 96-model factorial, augmentation is the one training choice that separates classes relative to chance; weight decay compresses the overlap without separating. Only a k-fold statistic can pose the structural question: the joint entanglement of a class triple sits below its strongest pair in 97% of triple-layer cells and 99.5% of deep cells, at median ratios far below a measured null floor, in vision encoders and frozen language models–a regularity, not a law. The unnormalized mass predicts test accuracy (R^2=0.94), the quotient does not, and neither beats a linear probe.

Personalized memory helps LLM agents deliver stable, tailored assistance by storing and reusing user-specific data across interactions. In multi-user scenarios, however, retrieval must consider not only semantic similarity but also whether the current requester matches the identity associated with the stored memory. We propose Bio-Memory, a biometric-aware memory architecture that conditions memory retrieval on both semantic similarity and biometric matching. Built on top of A-Mem, Bio-Memory augments each atomic memory note with a biometric embedding and uses biometric matching to form the retrieval candidate pool before semantic ranking. We evaluate Bio-Memory on LoCoMo in a 10-user shared-agent setting over 7 face benchmarks and 10 palmprint protocols. Across datasets, Bio-Memory consistently separates owner and non-owner queries. Under face-based personalization, the largest average gap reaches 27.29% / 21.15% in F1 / BLEU-1 on CALFW; under palmprint-based personalization, the corresponding gap is 25.75% / 19.22% on MS_Blue. These results support biometrics as a practical control signal for personalized memory retrieval in shared environments.

With the rapid development of aerospace technology and the large-scale deployment of low Earth orbit (LEO) constellations, the risk of orbital collisions has increased, creating a growing demand for reliable observations of satellite maneuvers. However, public datasets containing real maneuver records remain scarce. We present MAD-LEO, a Maneuver-Annotated orbital Dataset for LEO satellites. The mission-reported subset contains 1,134 maneuver events from eleven geodetic and altimetry satellites spanning 1992 to 2026, with labels taken directly from mission-published maneuver histories. Each event is checked against two-line element (TLE) data, precise orbit products, and satellite laser ranging (SLR) observations, with evidence tiers assigned according to data availability. The operational subset pairs operator-published ephemerides for 6,785 Starlink satellites with cataloged TLE records over a continuous 107-hour period. Technical validation across seven machine-readable experiment suites confirms the cross-source consistency of the labels and the evidence products.

In data-driven training, multivariate time-series forecasting is usually optimized with a scalar loss averaged over samples, variables, and horizons. This averaging is convenient, but the optimizer sees only the aggregated gradient, which does not reveal whether the variable-wise contributions align or oppose one another. To quantify how often this disagreement arises, we measure the variable-wise gradients directly and find that 30.6% of their pairwise cosine similarities are negative on average across seven datasets. However, conflict and harm are not the same thing. Under shared training 35 of the 64 variables do worse than a full-input single-target oracle, and the harmed fraction is not reliably predicted by how often gradients conflict. We propose Per-Variable Surgery (PV-Surgery), an optimizer-side training strategy for backbones with cache-compatible layers. One backward pass builds variable-wise gradient proxies from output-side signals and keeps the pointwise forecasting loss. Reliability-aware selection targets layers whose proxy sums closely approximate their shared-gradient slices. Conditional pooling forms anchor and conflict pools without dropping variables. Common-direction surgery aligns variable or pooled gradients with their normalized mean and restores input norms to avoid reweighting. In experiments across five backbones, seven datasets, and four horizons, PV-Surgery lowers MSE by 3.61% and MAE by 2.93% on average. For multivariate forecasting, this indicates that the variable-wise structure hidden by mean-loss training is a usable optimization signal.

Long-term memory is a core capability for personalized LLM agents. To support it, existing memory systems organize information using various criteria such as topic segments or summary hierarchies. However, we identify two major limitations in these designs. First, they lack scene awareness: memories from unrelated life scenes share the same retrieval space, which inflates the search space and introduces cross-scene interference. Second, they encode each memory from a single perspective, making it difficult to retrieve complementary views of the same event. In this paper, we propose the CreaMem architecture, which enables scene-aware memory organization by partitioning memory into several Life Scene Memories to reduce cross-scene interference at retrieval. To go beyond the single perspective and achieve cross-memory synergy, entries are dual-coded from both episodic and trait-based perspectives within each memory. We further devise a permemory balanced sampling strategy at retrieval time. Extensive experiments on two long-term memory benchmarks show that CreaMem improves QA accuracy across all evaluation metrics, with particularly large gains on multi-hop reasoning performance, validating scene-aware partitioning and cross-memory synergy. To enhance reproducibility, we release our code in a public GitHub repository.

Reconfigurable intelligent surfaces (RISs) make the wireless propagation environment programmable, while simultaneously transmitting and reflecting RISs (STAR-RISs) extend this capability to users located on both sides of a surface. However, conventional STAR-RIS architectures retain a fixed physical geometry after deployment. Movable-element STAR-RIS (ME-STAR-RIS) introduces an additional spatial degree of freedom by allowing the surface elements to reposition within prescribed regions while maintaining electronic control of their transmission and reflection responses. This combination of electromagnetic and geometric reconfiguration can alter propagation distances, multipath combinations, spatial correlation, interference, near-field focusing, and sensing geometry. This article presents a system-level perspective on ME-STAR-RIS through the concept of programmable geometry. We discuss its operating principles, movement architectures, and promising applications in communications, security, near-field systems, sensing, and high-mobility networks. A representative case study comparing optimized fixed and movable STAR-RIS architectures illustrates measurable spectral-efficiency gains from limited local displacement and the resulting saturation behavior. Finally, key hardware, channel-acquisition, electromagnetic, energy, reliability, and control challenges are discussed toward practical ME-STAR-RIS deployment.

Despite recent advances in surgical vision-language models (VLMs), temporal reasoning remains limited because existing supervision is largely frame-centric. Frame-level scene graphs (SGs) have proven effective in providing structured representations of surgical environments but do not explicitly model the dynamics of surgical workflows. To explicitly model how surgical states evolve across time, we introduce a multi-level structured temporal supervision methodology that augments frame-level surgical SGs with object-level continuity, event-level interaction continuity, and procedure-level connectivity. We then execute temporal queries over the resulting spatio-temporal scene graphs (STSGs) to generate evidence-grounded question-answer pairs, which together form the STSG-VQA benchmark. Each question is linked to the temporal interval and STSG evidence used to derive its reference answer, enabling traceable verification. The benchmark contains 18,458 question-answer pairs across seven temporal categories. Fine-tuning Qwen3-VL-4B and Hulu-Med-4B with STSG-derived supervision improves question-level micro accuracy by 24.39 and 19.56 percentage points over their zero-shot baselines and by 16.50 and 14.25 points over static scene-graph supervision, respectively. These gains span all temporal categories, indicating that STSG-derived supervision helps surgical VLMs reason over temporally grounded interactions rather than isolated frames. The code and dataset will be made publicly available upon acceptance.

We study the time-uniform convergence of the raw iterate of standard stochastic gradient descent (SGD) for unconstrained smooth convex objectives. We prove that, under standard noise assumptions, the time-uniform convergence rate gets arbitrarily close to $\sqrt{\log n / n}$ but never reaches it. More specifically, we prove that for every positive, eventually nondecreasing sequence $h$ satisfying $h(n) = o(\sqrt{n})$, a bound of order $h(n)/\sqrt{n}$, holding simultaneously for all $n$ with probability at least $1-\alpha$ and uniformly over the problem class, is achievable if and only if \[ \sum_{j = 1}^{\infty} \frac{1}{h(2^j)^2} < \infty. \] The constructive sufficiency result follows from a dyadic horizon-free schedule together with an additive conditional-restart inequality. The necessity counterpart applies to every deterministic nonnegative schedule and holds even for a one-dimensional analytic smooth convex objective with Gaussian noise.

Every data-driven equation discovery method assumes (implicitly and without verification) that the target system obeys a single governing law ($K{=}1$). Here we show that this assumption is the primary bottleneck limiting scientific discovery in multi-mechanism systems, and introduce Physical Law Ecology, a framework that makes $K^*$ (the number of coexisting independent mechanisms) itself the first quantity to be determined from data. The framework automatically mines a pool of topologically distinct candidate equations, constructs a continuous dominance weight field across parameter space, and discovers analytic evolution laws governing mechanism succession—with optional monotonicity constraints encoding irreversible physics. Across four unrelated systems (elastomer mechanics, pool boiling, galactic dynamics, and droplet evaporation), BIC consistently identifies $K^*{=}3$ independent governing topologies. Applied to 163 SPARC galaxies (3,269 spatially resolved measurements), the framework autonomously recovers three gravitational laws whose coexistence provides evidence against the single-universal-acceleration hypothesis of MOND ($p<10^{-34}$). In engineering applications, multi-law weighted prediction reduces error by 67-72% over single-equation baselines while retaining full interpretability. By establishing the determination of $K^*$ as the zeroth step of scientific discovery-prior to and independent of equation search—this work opens a direction orthogonal to existing symbolic regression: not finding better equations, but mapping the ecology of mechanisms that govern complex systems.

This study examines whether individual text corpora (ICs) from search histories can be used to simulate individual knowledge. We collected ICs from 316 adults, who answered 36 multiple-choice knowledge items, and compared several large language models (LLMs) on this task, of which only Qwen3-1.7B proved viable. After task-specific fine-tuning via Low-Rank Adaptation (LoRA), Qwen3-1.7B outperformed both participants and a representative German norm sample on publicly available items. On non-public questions, however, the LLM performed worse than our participants, suggesting possible training data contamination for the public questions. When integrating ICs into retrieval-augmented generation to predict individual responses, LLM-participant Match accuracies significantly exceeded chance, which demonstrates a detectable individual knowledge signal. The probabilities assigned to the participants' answers were, however, low and far below the probability of correct answers, indicating poor calibration toward individual response patterns. Knowledge-gap prediction was sub-optimal, though it improved for corpora exceeding five million tokens. We discuss our entropy based evaluation benchmarks as calibration indices for individualized knowledge simulation.

The phase retrieval problem involves reconstructing a function or signal solely from the magnitude of linear measurements. Most theoretical analyses of phase retrieval algorithms rely on i.i.d. Gaussian random measurements or sub-Gaussian random measurements. In this paper, our focus is on the fusion frame phase retrieval problem, where the sampling matrices are i.i.d. rank-$r$ orthogonal projections drawn from the Haar measure. We present concentration inequalities for functions on the set of rank-$r$ orthogonal projection matrices. These inequalities are crucial for the theoretical analysis of the fusion frame phase retrieval problem. Based on these inequalities, we demonstrate that gradient descent, combined with a two-stage initialization, achieves linear convergence to the target signal up to a global phase with a measurement complexity of $O(d\log^2 d)$ when the rank $r = O(1)$. We verify this convergence through numerical results.

We introduce a method for constructing predictive models of non cyclic physical processes directly from time-series data, without assuming an underlying differential equation. The observations define a discrete evolution rule whose recurrent behaviour captures the essential dynamics of the process. Analysing this behaviour across multiple geometric scales leads to probabilistic models in the form of Markov chains. Hyperbolicity criteria identify when these models provide a consistent statistical description of the data. The method is inspired by, and illustrated through, the analysis of a biological imaging data set referred to as the Cell Process.

We propose a new approach for synthesizing large sets of winning strategies in stochastic parity games (2.5-player games) with quantitative objectives. Instead of computing a single, fully specified winning strategy, we introduce Actionable Strategy Certificates (ASCerts) as a local and permissive representation of a large class of system player winning strategies. To this end, we extend known certificates for stochastic invariants to the setting of games. Our certificates prove that synthesized strategies remain within a safe region of the game with probability at least $\lambda \in [0,1]$. As such, the certificates enhance the trustworthiness of synthesized strategies. The crux of our approach is to reinterpret and leverage the certificates as concise, local, and permissive representation of (possibly infinitely many) strategies. By carefully combining our certificates for stochastic invariants with strategy templates for almost-sure winning, we obtain a novel local representation of quantitatively winning strategies in stochastic parity games. This enables efficient synthesis, adaptation, and runtime strategy extraction, making ASCerts well suited for logical control in uncertain and adversarial environments. We provide a proof-of-concept implementation and demonstrate the potential of applying ASCerts in runtime adaptation in a case study.

CleanCity-BinSense addresses inefficiencies in urban waste management in developing cities, where fixed-schedule collection routes lead to overflowing bins and wasted fuel. This paper presents CleanCity-BinSense, a low-cost, end-to-end IoT-enabled smart waste management system for scalable real-time waste monitoring and demand-driven collection. The system integrates a solar-powered sensor node with an ultrasonic sensor for real-time bin fill-level monitoring. A key contribution is a configurable sensing model based on two calibration parameters, FULL_DISTANCE and EMPTY_DISTANCE, enabling deployment across bins of varying sizes and geometries without firmware modification. Fill percentage is computed using a geometry-configurable linear normalization algorithm, validated through hardware experiments with a mean absolute error (MAE) of 0.38 cm, within the manufacturer-specified sensor tolerance. Sensor readings are transmitted via Wi-Fi to a centralized web platform providing role-based dashboards for administrators, operators, and drivers, along with a public real-time bin-status map. The system also incorporates a lightweight nearest-neighbor route planning algorithm using SQL Server's spatial function to generate proximity-based collection routes with low computational overhead. Experimental evaluation shows an average end-to-end system latency of 5.3 seconds, dominated by the sensing interval rather than network overhead, while route generation for typical urban collection zones completes in under 100 ms. These results demonstrate the feasibility of a low-cost, configurable, infrastructure-light smart waste management system for heterogeneous urban waste networks in resource-constrained environments such as Dhaka, Bangladesh.

Out-of-distribution (OOD) detection is crucial for safe deployment of medical AI systems, where domain shifts arise across institutions, acquisition protocols, and patient populations. VLMs enable zero-shot OOD detection by embedding images into a language-aligned latent space, where cross-modal similarity serves as a non-parametric confidence signal for identifying in-distribution samples. Yet existing methods rely almost exclusively on final-layer embeddings, implicitly assuming that the deepest representations are universally optimal. We first show that this assumption does not hold in medical imaging: intermediate layers provide complementary OOD signals, and the optimal representational depth depends on the respective image modality. While prior work selects layer combinations via entropy minimization of normalized histograms, we demonstrate that single-resolution entropy estimation is highly sensitive to binning choices, leading to performance variations of up to 19.3% AUROC. To address this instability, we propose a multi-resolution entropy estimation strategy that aggregates histogram statistics across multiple discretization scales, enabling robust and stable intermediate-layer selection. Across two medical OOD benchmarks, namely MIDOG and OASIS, covering distinct imaging modalities, diverse shift types, and different VLM backbones, our method consistently outperforms state-of-the-art approaches, offering a lightweight and stable solution for zero-shot OOD detection.

CSS-T codes are quantum error-correcting codes that play an important role in fault-tolerant quantum computation, as they help mitigate the proliferation of errors. They admit a transversal T-gate and are defined from a pair of nested classical binary linear codes satisfying a specific algebraic condition expressed in terms of their Schur square. We extend this algebraic characterization to the propagation rule known as the $(u \mid u+v)$-construction, that is, the matrix-product code constructed from two constituent codes. Moreover, for cyclic constituent codes, we provide an explicit characterization in terms of the defining cyclotomic sets, which extends existing results for cyclic codes. This framework allows the construction of new and longer CSS-T codes.

This paper takes a critical approach to visualization of colonial regimes. Drawing on critical hermeneutics and postcolonial theory, we created interactive visu- alizations revealing the constructed nature of colonial categories and exposing the hierarchies between empires and territories. Using the Varieties of Political Regimes dataset, our approach combines temporal flow visualization and geo- graphic distribution mapping. This allows us to contextualize colonial rule within broader patterns of political change. Our visualization challenges conventional, static, and isolated representations of colonial data by making the interpre- tive frameworks underlying colonial categorization visible through coordinated temporal, spatial, and relational views.

Diffusion models have become a core paradigm for multimedia generation, offering powerful concept-driven controllability for personalization, semantic editing, and selective unlearning. However, as semantic control extends beyond natural-language prompts to learned embeddings and intervention pipelines, the safety and governance of these systems become increasingly difficult to evaluate in a unified manner, especially for safety-sensitive, identity-linked, and other privacy-relevant concepts. Existing studies mainly rely on heuristic audits, adversarial probing, or task-specific erasure benchmarks, and therefore provide limited support for systematic comparison across models, conditioning channels, and deployment conditions. We present a concept-level probabilistic audit and reporting framework for diffusion models. We formalize governance-relevant concept behaviors as Bernoulli semantic events induced by stochastic generation, and define a Concept Risk Operator that maps model-channel configurations to structured risk profiles, enabling comparison across prompting interfaces, learned embedding channels, models, and recorded conditions. We apply sample-level post-hoc calibration and configuration-level risk aggregation, and show that probability error can change thresholded actions near policy boundaries. Experiments on SD1.5, SD2.1, and SDXL reveal consistent yet non-uniform operational risk patterns across concept families, channels, recorded conditions, and shifted protocols. In particular, embedding-based access and obfuscated prompts expose risks often understated by standard-prompt evaluation. A pooled multi-protocol calibrator improves held-out probability reliability, but we do not claim transfer from a standard-only calibrator. CLRC provides a common audit schema for probabilistic and decision-aware governance of multimedia generation systems.

As Large Language Models (LLMs) are increasingly integrated into human society, aligning them with pluralistic social values has become a critical priority. However, whether LLMs exhibit consistent value preferences across languages remains underexplored, particularly for culturally grounded values, which are more abstract and difficult to evaluate and align than safety-centric principles. We investigate this issue through Chinese Social Values (CSV), a value system rooted in Chinese culture and comprising $12$ dimensions across national, societal, and personal levels. We construct C-Voices, the first comprehensive multilingual contrastive probe dataset for CSV, with 86,400 dilemma-based instances in six languages, each pairing a CSV-aligned action with a value-conflicting alternative. Building on the contrastive probes of C-Voices, we then propose a fine-tuning-free value vector steering method that derives value directions from hidden-state discrepancies and selectively intervenes on value-sensitive layers during inference. Experiments on six languages show that CSV-oriented preferences are model-dependent and language-sensitive, with the same dilemma eliciting divergent responses across languages. Our method achieves effective CSV steering, supports cross-lingual transfer of value vectors, and generalizes to existing FLAMES and ValuePrism.

Autonomous exploration on uneven terrain requires ground robots to balance exploration efficiency, coverage completeness, and terrain safety. Detailed tsrrain reasoning improves local reliability but can slow large-scale exploration, whereas coarse region guidance expands quickly in open areas but can miss narrow passages and irregular traversable boundaries. To address this challenge, this paper presents TASG-Explore, a traversability-aware sector-guided exploration framework for ground robots. The framework first performs hierarchical traversability analysis using variable-voxel ground fitting and adaptive 8-bit obstacle encoding. It then splitting cost map into sectors, incrementally updates sector clusters, extracts terrain-coupled frontier viewpoints, and maintains a dynamic topological roadmap with unknown topological hypotheses. Finally, a sector-guided planner selects region targets and inserts local viewpoints to generate efficient exploration routes. Benchmark experiments in diverse challenging environments, including caves, forests, and rugged hills, show that TASG-Explore achieves the best overall performance among six representative state-of-the-art planners. The proposed traversability analysis improves processing efficiency by 6.3 times while maintaining high accuracy, and the exploration planner improves exploration efficiency by 51% and increases coverage by up to 2.95 times in rugged hill scene. Large-scale real-world experiments further demonstrate the practical value of the proposed method.

Humanoid motion trackers can reproduce diverse whole-body motions, but their performance degrades on complex terrain where terrain-agnostic references become physically infeasible. We present PGMT, a Perceptive General Motion Tracking pipeline for humanoid robots that learns terrain adaptation from independently selected motion references and terrains. PGMT first learns a general tracking and recovery prior, then incorporates terrain perception through motion-conditioned terrain glimpses that selectively encode regions relevant to the current motion. Terrain-aware tracking relaxation allows necessary deviations from the reference while preserving its motion intent. Zero-shot deployment on a Unitree G1 demonstrates robust terrain-adaptive locomotion and whole-body motion execution over real-world terrain with obstacles up to 37 cm high, while supporting teleoperation, dynamic motion tracking, and fall recovery. PGMT extends general humanoid motion tracking beyond flat ground, providing a unified policy for terrain-adaptive locomotion, diverse whole-body behaviors, and teleoperation in complex environments. Project homepage: https://luyili.github.io/pgmt/

Knowledge graphs have become a key resource for integrating heterogeneous data and powering downstream tasks such as question answering, entity linking, and semantic search. They are built and maintained incrementally, either (i) fully automated, e.g., YAGO, (ii) semiautomatically with community oversight, e.g., DBpedia, or (iii) manually through collaborative editing, e.g., Wikidata. Understanding the evolution of knowledge graphs is essential as changes may reflect real-world updates, error corrections, or noise introduced by vandalism, all of which affect the reliability of downstream applications. Among openly available knowledge graphs, Wikidata is the most challenging case to study evolution, with over 120 million entities edited by humans and bots and an edit history spanning more than a decade. Although Wikidata exposes change data in various formats (e.g., periodic dumps and real-time event streams), none support analytical queries over the complete edit history. Therefore, we present WiDiff, a tool that extracts changes from Wikidata's complete edit history and provides a unified interface for large-scale analytical queries over it.

Reliable video generation requires more than high-quality frames to form a coherent story: a model must maintain a persistent state, transporting visual attributes such as identity, scene layout, motion, and fine details across time. Existing training-free methods mainly strengthen cross-frame attention or analyze local attention entropy, but these views do not reveal whether temporal interactions stay in a healthy transport regime. In this work, we study video generation through the perspective of Temporal State Transport. We introduce Spectral Tension, a signed diagnostic that compares local attention diffuseness with global spectral diversity, and use it to identify two opposite temporal failures: fragmented transport and over-mixing hotspots. Based on this diagnosis, we propose Spectral Transport Homeostasis, a training-free regulator that softly corrects pathological temporal states while largely preserving balanced ones. Experiments on pretrained video generation models show that the original model often occupies imbalanced temporal regimes, whereas our method selectively applies larger corrections to the worst temporal hotspots and improves temporal consistency and visual quality without finetuning. Code: https://github.com/lytang63/temporal-state-transport

Centroidal Voronoi tessellation (CVT) is a fundamental primitive for high-quality surface sampling and isotropic remeshing in computer graphics. However, computing surface CVTs with classical solvers remains expensive: each optimization step repeatedly constructs restricted Voronoi diagrams (RVDs) and integrates quantities over their surface cells. We introduce Neural CVT, a learning-based surface-CVT solver that replaces these costly geometric computations with a recurrent neural optimizer, accelerating CVT optimization by one to two orders of magnitude in our benchmarks while preserving geometric fidelity. Given an input surface, we sample a dense point cloud and extract multi-scale geometric descriptors with a graph neural encoder. A lightweight recurrent optimizer then refines seed positions over a small number of iterations, aggregating interpolated surface features and optimization history to predict per-seed displacements. The framework is trained self-supervised using CVT objectives that promote uniform distributions and, when desired, feature alignment. Across diverse organic and CAD-like shapes, Neural CVT generalizes to unseen geometry, initialization strategies, and seed densities, producing isotropic, feature-preserving remeshes comparable to state-of-the-art offline optimization methods at a fraction of the computational cost. Code and trained models will be released.

Sign language video generation demands precise hand and facial articulation, yet modern video diffusion models, trained predominantly on spoken-language video, produce artifacts that render signing unintelligible. We propose SignRefine, a sign language video generation model that produces comprehensible signing from 2D keypoint conditioning alone, generalizing across appearances and visual conditions. Our approach builds on a pretrained video diffusion transformer and introduces local adapters with spatial grounding to selectively refine hand and face regions, steering the strong base model's prior toward accurate articulation. To enable this work and support broader sign language research, we present NVSign, a large-scale dataset of video content natively produced in sign language, offering diverse signer appearances, environments, and natural conversational settings. Trained on this data, our model shows up to 30% improvement in hand pose precision metrics over the strongest baseline and is preferred by sign language users for visual quality and comprehensibility in more than 80% of comparisons.

Observing objects grasped by a robot hand is challenging due to severe visual occlusions. Although in-hand manipulation can expose hidden surfaces, existing approaches often rely on predefined or open-loop reorientation strategies that do not explicitly target under-observed regions. We propose AURORA, an active 3D reconstruction framework that closes the loop between online object-centric reconstruction and in-hand reorientation. At its core, Ray-GPIS estimates direction-wise reconstruction uncertainty along candidate viewing rays and selects next-best-view targets using an uncertainty–novelty objective, which are realized through an axis-conditioned in-hand rotation policy. The resulting RGB-D observations are fused incrementally using CAD-free 6D pose tracking and lightweight geometric reconstruction. Experiments demonstrate that AURORA improves reconstruction quality and information-acquisition efficiency over non-active rotation strategies, while Ray-GPIS also outperforms active view-planning baselines in reconstruction performance, action-ranking quality, and planning efficiency. Targeted ablations further validate its robustness to hand occlusion and pose errors. The project webpage is available at https://aurorahand.github.io/

This paper presents a soft robotic drummer for accurate and efficient drum rolls. High-frequency drum rolls require the "multi-bounce technique," where a drumstick bounces multiple times with a single stroke. In robotic reproduction of this technique, the body's elasticity is key, while the fine motion during the stroke is also crucial for maximizing the potential of that elasticity. Therefore, we design two tricks: i) Tap-Pull (TP) trick to increase the number of rebounds by adding a pulling motion after impact; and ii) Micro-Pulse (MP) trick to keep the drumming volume by injecting small oscillations during the stroke. Due to the nonlinear complexity of soft embodiment, both tricks are efficiently tuned using Bayesian optimization in a data-driven manner for accomplishing the respective objectives quantified. We evaluated the optimized behaviors with soft and rigid end-effectors. As a result, the soft TP achieved the highest bounce count (12.25 per stroke) with uniform intervals. The soft MP suppressed the volume decay, yielding 6.8-times higher acoustic efficiency compared to the rigid MP. These results indicate that the proposed tricks with the combination of elasticity and optimization can make robots play excellent drum rolls.

We study strategyproof mechanisms for building a pathway between two regions of a line segment separated by an obstacle. Each of the $n$ agents has a private location within its region and may use either its original route to a facility or the new pathway, whose traversal cost is a fraction $k\in[0,1)$ of its length. We seek strategyproof (SP) and group-strategyproof (GSP) mechanisms that approximately minimize maximum cost or social cost. After characterizing optimal pathways for both objectives, we establish a tight deterministic maximum-cost approximation ratio of $\frac{2}{1+k}$ and a deterministic social-cost upper bound of $\frac{n}{1+k(n-1)}$, together with complementary lower bounds. Both upper bounds are achieved by GSP mechanisms. We then study randomized mechanisms under strategyproofness in expectation. A power-proportional mechanism achieves a social-cost approximation ratio at most $5$, independent of $n$ and $k$, with a tight guarantee of $3$ for this mechanism when $k=0$. We prove randomized lower bounds of $\frac{3+2k}{2+3k}$ for maximum cost and $\max\big\{1,\frac{285}{263+385k}\big\}$ for social cost, the latter for $n\ge7$. Finally, we improve several bounds for the real-line pathway model of [Chan and Wang, AAMAS 2023]. Our deterministic maximum-cost lower bound of $2$ matches the upper bound obtainable from [Qin, Fang, and Liu, COCOA 2024]. We strengthen the deterministic social-cost lower bound from $\frac32$ to $2$ under SP and to $\max\{2,n-1\}$ under GSP. For randomized social cost, we sharpen the guarantee of Chan and Wang's proportional mechanism from $6$ to $3$ and raise their lower bound from $1.02$ to $\frac{285}{263}\approx1.08365$ for $n\ge7$.

This paper examines the adaptation of a rhythm-based interface, originally designed for manual dexterity rehabilitation, for use in speech therapy. The interface allows users to control synthesized vocal phrases through finger tapping, leveraging the cognitive link between gesture and speech. Through interviews with four therapists and pilot tests with one speech therapist and eight children with speech impairments (autism, Down syndrome, verbal apraxia, dyslexia), we found that the interface improves motivation and therapeutic outcomes by facilitating more interactions between verbally challenged patients and the therapist. Our findings suggest that rhythmic gestures can enhance verbal communication, offering potential for broader therapeutic and educational applications.