Papers

Vision-Language-Action (VLA) policies fuse multimodal sensory inputs, but training on limited and homogeneous robot demonstrations encourages spurious inter-sensor correlations rather than task-relevant signal, a failure we term modality entanglement. Under real-world occlusions and distractors, this manifests as nuisance sensitivity to corruption of uninformative sensors and single-modality insufficiency when only one informative sensor remains intact. We propose Evidence-Gated Regularization (EGR), a modality-agnostic training objective that introduces zero inference-time overhead. EGR derives a per-frame and per-sensor task-relevance signal to gate two state-conditional consistency objectives: invariance on low-evidence sensors, and single-sensor sufficiency on high-evidence ones. We introduce a benchmark based on BEHAVIOR-1K, comprising a fast inference-only diagnostic suite and 47 rollout-based skills targeting modality entanglement. We validate EGR on this benchmark and on two real-robot setups with fundamentally different embodiments: a bi-manual setup with two Kinova arms and three RGB cameras, and a single-arm MELFA ASSISTA setup combining vision and GelSight tactile sensors. EGR improves simulation success rates (SR) from 12.5% to 16.4% under full modalities (+31%), from 9.4% to 16.5% under uninformative-sensor corruption (+75%), and from 2.8% to 6.1% under single-sensor fallback (+120%). Under physical-object distractors, EGR boosts SR from 30% to 85% on the bi-manual setup (+183%) and from 55% to 70% on the tactile setup (+27%).

The bitter lesson poses an existential question for the data systems community, whereby large language models (LLMs) trained end-to-end are rapidly internalizing new capabilities that previously required carefully engineered data agents. Guided by empirical insights, we argue that as models continue to improve, many proposed system layers designed to compensate for model limitations on a given task will increasingly be subsumed by the model itself. We instead identify enduring research opportunities, which lie in supporting data agents across many queries with curated contextual information about the data environment, which we call persistent semantic context. We find that these context layers demonstrate strong promise for improving data agent performance, but they also raise significant system challenges. Thus, a key requirement for future data systems will lie in natively serving persistent semantic contexts as a first-class abstraction in order to enable capable data agents working over huge, complex knowledge corpora. Towards this vision, we outline exciting new research opportunities, including designing efficient context data structures, storage methods, compression techniques, and semantic consistency protocols, to ensure integrity and correctness of the stored contextual knowledge.

Deep neural networks (DNNs) are increasingly deployed in real-world vision systems, yet their predictions can be covertly manipulated by backdoor attacks, in which malicious triggers cause targeted misclassification while preserving high clean accuracy. Existing defenses often rely on model internals, training data, or clean validation samples, making them difficult to deploy when only black-box access to a trained model is available. We propose TRIM (Trigger Removal by Identifying Manipulated Regions), a deployment-oriented black-box defense that detects and selectively removes backdoor triggers at inference time without requiring model internals, training data, or clean samples. The key insight behind TRIM is to identify image regions that are responsible for anomalous model behavior and purify only those regions while preserving benign content. TRIM innovates via three key components: (i) region-based segmentation with deep feature representations, (ii) adaptive trigger discovery through inpainting and diffusion-based reconstruction to isolate regions responsible for misclassification—without assumptions about trigger type, shape, or location, and (iii) selective region purification that cleans poisoned regions while retaining benign content. To support practical deployment, TRIM further caches feature embeddings of previously identified triggers, enabling efficient recognition and avoiding redundant detection and purification. Extensive experiments across diverse datasets and backdoor types, including blended, sparse, varying-size, and multiple triggers, show that TRIM consistently outperforms existing black-box defenses, reducing attack success rates (ASR) to as low as 1.16% while preserving clean accuracy of up to 87.87%. These results demonstrate that effective backdoor mitigation is possible at inference time even when the defender has no access to any auxiliary data.

Computation of adaptive beamforming weights in Minimum Variance Distortionless Response (MVDR) processing is a latency-critical operation that poses significant challenges for real-time hardware implementation. This paper presents an FPGA implementation of a systolic Givens-rotation QR decomposition pipeline for MVDR beamforming on a simulated 32-element ultrasound transducer array, using single-precision floating-point arithmetic. The design is deployed on a Zynq UltraScale+ FPGA at 100 MHz with three parallel kernel instances operating concurrently, achieving a measured throughput of 31,123 weight vectors/s at 90.9% parallel efficiency relative to the measured single-instance rate. At an estimated 2.451 W of programmable-logic power, and 5.286 W including the processing system, this corresponds to 12,698 and 5,888 weight vectors/s/W, respectively. Under a matched three-way dispatch, a 24-core Intel Xeon Gold 5220R at 2.20 GHz achieves 465,699 weight vectors/s at a measured 83.08 W package power, corresponding to 5,606 weight vectors/s/W. The FPGA therefore attains 2.3x the power-normalised throughput of the processor on a programmable-logic basis and 1.05x on a total on-chip basis. In contrast, the processor retains a raw throughput advantage of approximately 15x at this operating point. Numerical precision is validated against MATLAB float32 reference outputs from a Field II cyst phantom simulation, achieving a 100% pass rate with a root mean square error of 5.10x10^-7, confirming near-theoretical finite-precision behaviour without systematic bias.

Modern feedback design for distributed parameter systems presupposes that the closed-loop dynamics define a well-posed evolution problem. This presupposition becomes nontrivial for temporally degenerate parabolic systems, where temporal degeneracy affects not only the analytical properties of the evolution equation but also the mathematical formulation of the feedback interconnection itself. It is shown that admissible feedback interconnections for temporally degenerate parabolic systems are completely characterized by an operator compatibility condition linking the singular reaction operator with the actuator and observation operators. This characterization removes the singular component of the closed-loop dynamics and reduces the degenerate evolution equation to a regular evolution equation. Building upon this regularized formulation, a critical–residual decomposition yields a uniform exponential stability certificate, which is subsequently extended to the original infinite-dimensional evolution through a finite-to-infinite lifting theorem. A constructive static output feedback synthesis is finally obtained as a consequence of these results. Numerical experiments illustrate the regularization mechanism, validate the stability certificate, and confirm the finite-to-infinite lifting.

The Train Real-time Data Protocol (TRDP) is widely used on rolling stock but it provides limited native support for cryptographic protection. Furthermore, the multicast traffic profile used in TRDP Process Data to exchange critical information between onboard subsystems makes the introduction of cryptographic protection a challenge. This paper presents a lightweight security layer for secure TRDP communication that implements a number of security profiles built around modern cryptographic algorithms. This additional layer relies on an On-board Key Management System (OKMS) for both security profile negotiation, dynamic key distribution and key lifecycle management. The security profiles allow for cryptographic agility and flexibility, ranging from simple authentication to Authenticated Encryption with Associated Data (AEAD) algorithms. The security profile negotiation procedure guarantees all TRDP End Devices (ED) on a common Communication ID (ComID) share the same security profile and can therefore process each other's messages. A prototype implementation based on mbedTLS and Arm Platform Security Architecture (PSA) was developed and evaluated. Experimental results demonstrate manageable overhead, suitable for the real-time and time-sensitive communication found on rolling stock.

Several classical machine-learning methods, such as KRRs and SVRs, are both computationally and analytically tractable since their estimators either admit closed-form expressions or are obtained by minimizing convex training objectives; neither feature is generally available for deep neural networks. We address this by introducing a simple closed-form “two-stage” compositional formula $\hat{f}$ for reconstructing an unknown Lipschitz function $f:\mathcal{X}\to \mathbb{R}$ on a metric space $(\mathcal X,\rho)$ from $N$ i.i.d. noisy observations. Our main result is a high-probability uniform ($L^{\infty}$) recovery guarantee that jointly controls approximation and statistical errors while enjoying an optimization error of zero; in particular, we do not assume oracle access to an approximate ERM. Our secondary main results establish the optimality of our formula in three complementary senses. 1) Function space: On Ahlfors-regular metric spaces, the hypothesis class parameterized by our formula attains the optimal fat-shattering dimension. 2) Parameter space: Its dependence on the parameters is maximally numerically stable, in the sense that a smaller approximation error cannot be achieved with a smaller Lipschitz dependence on the model parameters. 3) Forward pass: Its dependence on the input is maximally regular, matching the Lipschitz constant of the target function $f$. When $\mathcal X=[0,1]^d$ is equipped with the $\ell^\infty$ norm, $\hat{f}$ admits algorithmic ReLU-MLP and exact ReLU-multi-head transformer realizations of depth $\mathcal{O}(\log(N))$ with $\mathcal{O}(N)$ nonzero parameters.

Let $R$ be a commutative ring and $f(a_1, \ldots, a_n) = \sum_{i=1}^k r_{i1}^{a_1} \cdots r_{in}^{a_n} m_i$ be a linear-exponential map over an $R$-module $M$. Dong and Shafrir (2026) showed that, when $\ell M = 0$ for some $\ell \in \mathbb{N}_{>0}$, the zero set of $f$ is the intersection of effectively computable $p$-normal sets, where $p$ ranges over the prime divisors of $\ell$. This generalizes an earlier theorem of Derksen and Masser (2012) on the solution set of $S$-unit equations over fields of positive characteristic. The purpose of this paper is twofold. First, we give a shorter proof of Dong and Shafrir's result, using the theorem of Derksen-Masser as a blackbox. Our proof also yields a decomposition of the zero set as a positive Boolean combination of affine transformations of zero sets of linear-exponential equations over fields. Second, we prove a multi-dimensional generalization of the Skolem-Mahler-Lech theorem over rings of finite characteristic. Specifically, we show that the zero set of every $n$-dimensional linear recurrence sequence over an $R$-module $M$ satisfying $\ell M = 0$ is the intersection of effectively computable $p$-normal sets (in $\mathbb{N}^n$), where $p$ ranges over the prime divisors of $\ell$. For example, this gives a decision procedure for whether two classical linear recurrence sequences have a common value over a ring of characteristic $p^a$ or $p^a q^b$, where $p$ and $q$ are primes.

This work introduces a time-encoded analog photonic interposer that enables long-distance, high-fidelity transport of analog signals between spatially separated chiplets. Unlike prior silicon-photonic links limited to digital data, the interposer preserves analog information by converting amplitudes into timing intervals using an analog-to-time converter (ATC), transmitting them over a wavelength-division-multiplexed photonic link, and reconstructing values at the receiver without an explicit high-precision ADC/DAC data-converter pipeline. The link instead embeds an implicit 6-bit time-domain quantization and uses a single wavelength per processing element independent of bit precision. Evaluated in a fully analog vision pipeline with an in-pixel computing (IPC) sensor, it achieves a 2.04x energy–delay product (EDP) improvement over an 8-bit digital electrical baseline on the 560x560 Visual Wake Words dataset, with the advantage widening with link length even against a precision-matched 6-bit baseline. The pipeline holds 89.87% and 86.15% accuracy on ResNet18 and MobileNetV2 for VWW and generalizes across CIFAR-10 and ModelNet40 within 2% of the digital baseline.

This paper is the first in a series on Turn-Based Combat Arena, a configurable framework for turn-based strategy games designed to support the efficient training and evaluation of machine learning agents. The proposed framework enables flexible modification of game rules and parameters, allowing rapid experimentation across diverse scenarios. Its architecture is optimized for high-throughput simulation, supporting tens of thousands of games per second and enabling the storage and processing of billions of gameplay records on a single machine. The problem of balancing the game, and particularly the parameters of game units, is investigated in detail. We evaluate several optimization approaches and show that multiple methods converge to comparable solutions, suggesting robustness in identifying balanced game configurations. These results indicate that the framework can serve as a practical platform for both game design analysis and agent training.

Classical asymptotic stability guarantees convergence but does not quantify the rate at which convergence occurs. This paper introduces a gradation of asymptotic stability where degree zero corresponds to exponential stability and degree $m>0$ corresponds to algebraic decay of order $t^{-1/m}$. We provide direct and converse Lyapunov characterizations for admissible degrees and conditions for certifying the exact stability degree. Hopf, Bautin, fractional-degree, and time-varying examples demonstrate how the degree identifies the leading stabilizing mechanism.

The evolution of the electric power grid towards a “smart grid” is an example of an emerging large-scale cyber-physical system. To study the potential of cyberattacks or other sources of extreme events, high-fidelity co-simulation is one of the few realistic strategies. In this work, we demonstrate a practical and high-performance co-simulation strategy for smart grids, which provides lessons about co-simulation integration that can be carried over to other cyber-physical systems. Specifically, we present a co-simulation platform based on the Mosaik framework that integrates several federates including OpenDSS for power flow, a refined NS-3 for communication networks, and custom Python-based simulators for on-load tap changer control, distributed state estimation, and data collection. We compare synchronization strategies – exhaustive lock-step time advancement versus an optimized event-driven approach that exploits lower-bound time stamps (LBTS) and next-event prediction – and introduce targeted refinements to NS-3 event handling and relevance filtering of internal events. Performance is evaluated on two standard IEEE benchmark systems: the 13-node test feeder with tap-changer voltage regulation and a large-scale 33-bus system augmented with 32 European low-voltage feeders (1,793 nodes total) performing distributed state estimation with thousands of phasors and smart meters. Experimental results demonstrate that the refined event-based synchronization that uses LBTS-based NS-3 event filtering reduces execution time by up to 50% over non-filtered LBTS, and more than 4$\times$ in the tap control scenario compared to naïve lock-step methods, while strictly preserving temporal correctness and simulation accuracy.

Neural fields (NFs) map continuous coordinates to signals such as color or density, but fast high-quality reconstruction from sparse observations remains difficult. Classical Neural Tangent Kernel (NTK) regression gives closed-form fits, yet it is fundamentally linear and cannot accumulate reusable task priors. We develop three algorithms that address these gaps. NTK-KIP learns a distilled support set of coordinates (and optional labels) so that a finite NTK can inpaint large missing regions from little observed data, yielding a compact non-linear representation instead of a raw kernel solve. MetaQuill meta-learns a shared initialization for an INR so that new scenes can be adapted by updating only a small task-specific weight offset, which provides true feature learning and a reusable prior. Finally, MetaQuill-KIP fuses both ideas: it seeds the task with a KIP-style non-linear warm start, then refines only that small offset around the meta-learned initialization. MetaQuill-KIP achieves high-PSNR reconstructions and semantically plausible inpainting under very sparse observations, while requiring only lightweight per-instance adaptation, whereas diffusion-style baselines typically depend on large pretrained generative priors and costly per-image tuning. This shows that NTK-driven neural fields can be made both non-linear and meta-learnable, narrowing the gap between analytic kernels and practical few-shot reconstruction.

This paper describes our experience extending the polyhedral compiler Pluto with a lightweight, coordinate-wise hill-climbing tuner that adjusts numeric transformation parameters, such as tile sizes and thread-block dimensions, after Pluto selects the kernel's loop structure. To ensure fast convergence and escape local minima, hill climbing is augmented with two techniques: expanded neighborhood exploration and a shortest-hop refinement phase. On x86 and ARM CPUs, tuned kernels outperform Pluto's default configuration (1.06-1.28x geometric mean speedup across 11 benchmarks) and static optimizers (Clang -O3, Polly, IOOpt), reaching performance competitive with the AutoTVM autotuner at substantially lower search cost. Applying the same technique to GPU thread-block allocation on an NVIDIA A100 yields 5.5-8.5% improvement over default configurations. These results position post-optimization parameter tuning as a practical middle ground between fixed-cost-model polyhedral compilation and full autotuning.

Current AI agents compellingly describe slides. However, AI-assisted slide editing requires more than understanding: the output must retain layout, style, component structure, and native editability. Towards, AI-assisted slide editing, existing agents operate on screenshots or weak document representations and often fragment coherent visual units, rasterize editable content, or break layout. In contrast, for controllable slide editing, we introduce an agentic framework, SLIDEFORGE, which builds a Deck State Graph, an executable slide state that links visual decomposition, native pptx object structure, and perceptual organization. By recovering human-referable components while retaining fine-grained editable structure, SLIDEFORGE supports theme-preserving reconstruction through slide-native operations and rendered-state verification. We further introduce an evaluation paradigm for controllable slide transformation that jointly measures component recovery, preservation, restyling consistency, visual quality, and native editability. Experiments show that SLIDEFORGE outperforms direct prompting, screenshot-based agents, and generic code-agent baselines across these dimensions. Code is available at https://github.com/UIUC-MONET/SLIDEFORGE.

Recent advances in generative speech models have made it increasingly difficult to distinguish authentic from synthetic audio, enabling new forms of fraud and misinformation. Audio watermarking offers a promising defense by embedding an imperceptible signal into generated speech that can later be detected to verify its provenance. However, recent studies have shown that existing post-hoc watermarking methods fail under neural codecs and denoisers, transformations routinely applied during real-world storage, transmission, and processing, severely limiting their practical utility. Here we introduce CRAW, a codec-robust audio watermarking framework that jointly improves robustness against neural re-synthesis while maintaining high perceptual quality. CRAW combines distortion-aware training with an attention-based pooling mechanism, inference-time perceptual mask- ing, and an error-correcting code to recover the fidelity lost during robust training. Experiments demonstrate that CRAW achieves state-of-the-art robustness against neural codecs, denoisers, and vocoders while maintaining perceptual quality comparable to existing post-hoc watermarking methods. The code is available at https://github.com/DavidC1212/craw.

Scaling laws in modern deep learning describe how held-out loss improves as model capacity, training data, and compute increase, often following power-law trends. We investigate whether analogous scaling regularities arise in actuarial ratemaking, where data are tabular, heterogeneous, and noisy, and where classical models such as GLMs remain strong baselines. Using a real-world motor insurance portfolio, we train models from different families across increasing fractions of the training data and multiple random seeds, evaluating out-of-sample Poisson deviance, a likelihood-based loss for Poisson count predictions in which lower values indicate better held-out fit. We find that all model families improve with additional data, but scaling exponents differ substantially: TabM exhibits markedly stronger data scaling than purely supervised tabular Transformers and standard MLP baselines. Transformer variants show weak parameter scaling unless augmented with additional inductive biases (TabM-style adaptation or self-supervision). These results provide quantitative guidance on model selection by data regime and suggest that effective scaling on actuarial tabular tasks depends on architecture and loss function objective design, with simple increases in Transformer size providing limited gains.

Scientific datasets are most compatible with AI workflows when they are described by rich, machine-readable metadata and provenance information, following the FAIR guiding principles. The FAIR Open-Science Extensible Data Exchange Network (FOXDEN) is a set of cyberinfrastructure building blocks developed at the Cornell High Energy Synchrotron Source (CHESS) for annotating raw, reduced, and analyzed datasets with both structured and unstructured metadata and provenance records. It also allows researchers to publish these records with Digital Object Identifiers (DOIs) to create AI-ready datasets. We describe FOXDEN's architecture and its deployment at CHESS and at the National High Magnetic Field Laboratory, where it serves as a foundation for future agentic scientific workflows.

Conformal risk control is an emerging framework for the safe deployment of machine learning models with finite-sample guarantees. To accommodate a broader class of risk notions, quantile risk control extends this framework to quantile-based risk measures. However, existing methods either suffer from excessive conservatism or lack rigorous finite-sample guarantees. To address these limitations, we introduce Occupancy-based Quantile Risk Control (OQRC), a novel method that provides tight risk control bounds with finite-sample validity. Our key idea is to formulate risk control as a finite-occupancy problem by partitioning the loss space with the ordered calibration losses. Specifically, we estimate the distribution of test losses across the resulting bins and upper-bound the risk by the maximum loss attained within each bin. We then select the parameter $\lambda$ such that this upper bound does not exceed a predefined threshold $\alpha$ with high probability $1-\delta$. Theoretically, we establish a finite-sample guarantee showing that OQRC yields tight risk control bounds that converge to the optimal bounds at a provable rate of $\mathcal{O}_ p(n^{-1/2})$. Extensive experiments demonstrate the effectiveness of our method, reducing the risk gap by up to 78.64% on common benchmarks.

Deformable linear objects such as wires and cables are difficult to segment because they are thin, highly deformable, and frequently self-occluded, while large-scale instance-level annotations are expensive to obtain in real scenes. Existing resources either focus on cable tracing or semantic segmentation under constrained settings, or generate visually plausible images without physically grounded wire deformation. We present WireSeg-32k, a synthetic dataset for wire instance segmentation with 32,000 RGB images, instance masks, depth maps, and a complementary real-world test set with annotations. To generate this dataset, we develop DeformX, a co-simulation pipeline that couples Cosserat-rod dynamics with photorealistic Isaac Sim rendering, enabling physically plausible, contact-consistent wire shapes, CAD-based wire assets, and diverse visually grounded scenes. As a simple baseline, LoRA fine-tuning SAM3 on WireSeg-32k alone improves real-world mAP@75 by 10.2% over the off-the-shelf model, showing that physically grounded synthetic data can transfer to real wire perception.

Geostationary atmospheric motion vectors (AMVs) provide the dense horizontal wind vectors (u,v) and heights ingested into data assimilation systems. Traditional AMVs track features using window-based cross-correlation and estimate heights via infrared brightness temperatures paired with numerical weather prediction (NWP) background states, creating a circular dependency that yields inaccurate heights, high computational cost, and sparse retrievals. Stereo winds from GEO-GEO and GEO-LEO geometrically resolve heights from parallax shifts across different poses, eliminating NWP dependence and improving accuracy, but they remain computationally heavy with limited coverage. In this work, we replace window-based tracking in stereo matching with deep optical flow for efficient, improved retrieval. Fine-tuning balances a self-supervised geometric residual loss with supervised radiosonde reconstruction. To eliminate multi-satellite overlap requirements, we distill the stereo teacher into a single-satellite student model. Chi-square and height uncertainties from the teacher are emulated by the student for quality assurance. The student generates winds across full-disk GEO imagery globally. Validation compares stereo and student models against radiosondes, operational AMVs, ERA5 reanalysis, and EarthCARE cloud profiles. Results through triple collocation show that stereo winds improve performance beyond operational AMVs for water vapor bands (6.2, 6.9, and 7.3 {\mu}m), wit degradation in the long-wave infrared (11.2 {\mu}m) band.

We model insider threat detection as a dynamic Bayesian game in which a platform coordinates a committee of strategic certifiers to sustain equilibrium among honest users and detect malicious deviations before exfiltration. Certifiers and users operate under a Bayesian Temporal Correlated Equilibrium (BTCE), where a sealed-envelope correlating device issues private recommendations over time and obedience is verified at every on-path information state. Unlike Stackelberg formulations, BTCE coordinates heterogeneous certifiers without requiring commitment power. We incorporate present bias and loss aversion to capture impulsive escalation dynamics, enabling 1.7–4.5 days earlier detection than rational baselines. We prove three guarantees: (1) calibrated intervention losses make recommended behavior a current-self best response despite behavioral biases, (2) controlled evidence accumulation guarantees intervention in bounded expected time before exfiltration, and (3) median aggregation confines implemented actions to the honest recommendation range when fewer than half of certifiers are Byzantine. On CERT r6.2 our mechanism achieves up to 28.3% pre-exfiltration detection with false positives below 1.6%, while both a transformer baseline and a streaming provenance approximation (HOLMESLite) achieve near-zero pre-exfiltration detection under comparable constraints.

Smartphone skin photographs are indispensable to teledermatology, yet assessing the diagnostic suitability of submitted cases (gradability) remains a critical bottleneck in mobile care workflows. Dermatologists routinely review multiple photographic views (regional, angled, and close-up) to identify consistent textural detail rather than relying on a single image. We present the Semantic Tri-view Pipeline, an interpretable architecture for automated teledermatology gradability screening that formalizes epidermal micro-relief as a computable biomarker of image quality. Using an expert-annotated subset of the public SCIN dataset, we train a lightweight DeepLabV3+ model to segment micro-relief fidelity. These spatial masks are then aggregated across up to three case views with a logistic regression classifier, leveraging viewpoint redundancy to support robustness under uncontrolled smartphone acquisition. This approach learns context-aware, clinically intelligible heuristics, such as penalizing high-fidelity texture in regional distance views. Evaluated at a predefined 90% sensitivity operating point, the system's apparent errors largely reflect subjective clinical variance on borderline cases where clinicians rely on non-visual metadata. On SCIN, performance improves from an AUC of 0.81 (80.6% PPV) on variance-heavy majority-consensus cases to 0.96 (97.7% PPV) on optically unambiguous unanimous cases. Overall, this work delivers an interpretable, privacy-by-design, edge-ready system that can provide real-time feedback during case submission to filter ungradable photo sets before review.

The Gromov-Wasserstein (GW) distance provides a principled framework for aligning metric measure (mm) spaces based solely on their intrinsic structure. Its ability to identify isomorphic representations of distributions across spaces renders it valuable for comparing data where equality up to isomorphism occurs naturally such as in graphs or, more generally, distributions on graphs. Recently, a type of dual form for the GW distance between Euclidean distributions with the squared Euclidean or inner product costs was derived, spurring the development of new statistical and algorithmic results for this setting. This work furnishes a novel duality result for GW distances with and without entropic regularization that is applicable to all finitely supported mm spaces. Leveraging this result, we derive the sample complexity of empirical GW distances between finite mm spaces, as well as limit distributions under proper centering and scaling. Furthermore, we propose new algorithms for solving the regularized GW problem which are subject to formal convergence guarantees. These statistical and algorithmic advancements give rise to a principled and efficient framework for testing whether two distributions on the set of graphs with a fixed number of nodes are isomorphic based on samples.

Continuous chain-of-thought models compress reasoning into latent tokens. Matrix-valued variants, which route each latent token through a d x d matrix bottleneck, introduce rank as a single-sample structural observable on the latent matrix Z. If matrix latents carry parallel reasoning paths via superposition, rank should track them, and truncating Z to low rank should hurt accuracy on tasks whose solutions plausibly require multiple components. Across four training regimes of a matrix-CODI model (three on ProsQA, one on GSM8K-Aug below the learning threshold), the rank-k projection ablation curve is flat to within 0.6 percentage points. A three-seed replication yields 81.0 +/- 2.0 percentage points accuracy while the final effective rank of Z spans {4, 12, 13}; the loss does not reward any particular rank. To test whether rank-blindness arises from the flatten-then-project readout alone, we trained four readouts: a bilinear reparametrization, a bilinear-plus-GELU readout nonlinear in Z, an SVD-augmented readout feeding singular values through an MLP, and a quadratic readout in Z Z^T. All four rank-k curves remain flat (Spearman p-values 0.63, 0.14, 0.82, 0.46). The flat curves persist for readouts nonlinear in Z. A linear probe on Z underperforms a raw pretrained hidden state at target prediction (AUC 0.673 vs. 0.846). A negative control on vanilla GPT-2 SFT (no matrix bottleneck, no Z, three seeds, n=500) reproduces a flat rank-k curve under the same intervention paradigm with pooled-mean range 0.20pp, and a random-h sensitivity floor lands at the same accuracy: the rank-k ablation alone conflates rank-blindness with position-irrelevance.

This article introduces Eiffel-tools, a language server protocol (LSP) implementation for the Eiffel programming language that uses Large Language Models (LLMs) to aid the development of statically verified software. The tool provides various interactive and non-interactive commands to produce code and specifications. It uses language and project specific knowledge to precisely direct the LLM and verifies the output using a static verifier. It crafts rich programmatic prompts for the input and corrects or rejects the output. Furthermore, it handles the retries until the program passes verification. The tool's bug fixing capability is evaluated on 2 public datasets using 3 models. The tool can fix 76% to 95% of bugs by combining LLMs and a formal verifier depending on the model and prompts used. The results show the trade-off between the number of fixing attempts and the success rate.

There is growing interest in adopting CLIP-style vision–language model (VLM) pretraining for mammography. However, models that directly employ the standard CLIP architecture and training objective exhibit limited zero-shot performance in clinically important tasks such as cancer, finding-type, and BI-RADS predictions. We argue that this underwhelming performance is due to neglecting two characteristics of mammography data: (1) its high-res nature, and (2) homogeneity of radiology reports, largely driven by a predominance of negative/benign findings on examinations. We propose TopKSigLIP, a VLM designed to address these two limitations through a novel architecture and learning objectives. Instead of downscaling high-res mammography images to satisfy GPU memory constraints, TopKSigLIP introduces TopK-Patch module that learns to sample a sparse set of high-res patches likely to contain lesions, sidestepping the resolution–batch size tradeoff of VLM training. The sampled patch locations additionally serve as a built-in localization tool. To address report homogeneity, we replace the contrastive loss, which falsely repels semantically similar pairs, with a Sup-sigmoid loss. Sup-sigmoid loss extends the sigmoid loss from SigLIP with soft labels derived from structured data. TopKSigLIP outperforms existing open-source mammography and general medical VLMs on both internal and external benchmarks on density assessment, BI-RADS classification, finding subtyping, and cancer prediction under zero-shot evaluation. TopKSigLIP remains competitive under linear probing despite using a significantly smaller vision encoder and smaller training batches than baselines. The TopK-Patch module additionally achieves superior lesion localization over post-hoc Grad-CAM. Code and weights are made public:https://github.com/Youngseok0001/TopKSigLIP.

Segmenting a new microscopy dataset usually means training a deep network on its annotations, and training again for every added image. Exemplar solves a segmentation readout in closed form and updates it exactly with each new mask in seconds, as accurately as networks trained for minutes. The updated readout is identical to a fit from all masks at once. Its linear readout combines frozen DINOv3 features, read at a magnification chosen from the masks, with classical filter responses at native resolution. Its regularisation and its pixel weighting are chosen exactly by leaving out one annotated image at a time, so none is set aside for validation. On ten biomedical datasets from cells to vessels and filaments, Exemplar is more accurate than every few-shot method that only applies a pretrained model, and on average than nnU-Net trained on the same masks. A fine-tuned Cellpose-SAM, the strongest baseline, is as accurate on average, but each of its fits takes a median of 25 minutes against 8.5 seconds, and each new mask means fine-tuning it again.

On-device LLM inference is attractive for privacy and responsiveness, but remains challenging on mobile and embedded devices because model weights far exceed available DRAM. Prior systems exploit activation sparsity and offload weights to SSD or flash storage, but face a fundamental systems trade-off: accurate sparse execution decisions require the latest context, whereas efficient computation-I/O overlap requires early prediction. As a result, existing designs either serialize execution or incur redundant weight fetches, extra computation, and large cache overheads. We present LeanStream, a streaming speculate-and-refine framework for efficient on-device LLM inference. LeanStream progressively refines computation, loading, and cache-retention priorities using partial GPU results, enabling fine-grained overlap between GPU execution and storage I/O. We implement LeanStream on both mobile and embedded platforms. Compared with prior on-device LLM inference systems, LeanStream reduces memory usage by 4.8$\times$ to 7.5$\times$ at the best throughput achieved by prior work, while further improving token generation throughput by 1.6$\times$ to 2.1$\times$. Source code is available at https://github.com/RenyuanLIU/LeanStream.

Threshold-based similarity search over tree-structured data using tree edit distance (TED) is computationally intensive. Given a query tree and a database of trees, the goal is to retrieve all trees within a predefined TED threshold $\tau$. Because exact TED computation is expensive, practical methods employ lower-bounds to prune dissimilar candidates before verification. Existing lower-bounds exhibit a fundamental trade-off: inexpensive statistical and structural bounds provide limited pruning power, whereas the more precise traversal-based string edit distance (SED) bound is expensive to compute using standard quadratic dynamic programming. Moreover, previous comparative studies do not cover recent structural filters or threshold-aware SED implementations, leaving their practical trade-offs unclear. In this article, we first provide a comprehensive experimental comparison of state-of-the-art TED lower-bounds in terms of pruning precision and computational cost. We then accelerate the SED lower-bound using Ukkonen's bounded string edit distance algorithm, substantially reducing its runtime without affecting its pruning power. Finally, we introduce the SED-struct threshold filter, which strengthens SED with axes-aware constraints capturing structural relationships among tree nodes. Experiments on synthetic and real-world datasets show that SED-struct consistently achieves the highest filtering precision while retaining practical filtering costs. The results suggest that SED-struct is particularly beneficial for heterogeneous tree collections in which the standard SED lower-bound achieves relatively low precision.