Science & Mathematics

Coarse-to-real (C2R) video generation converts coarse animation, simulation, or game-rendered videos into photorealistic videos while preserving scene layout, camera motion, and dynamics. Training such models requires aligned (coarse, realistic) video pairs, which are expensive to create and confine prior systems to narrow synthetic domains. We exploit a simple asymmetry: paired videos are hard to obtain, but paired real-to-coarse images are easy to construct with existing editing and 3D reconstruction tools. Using these image pairs, we train the DINO-R2C model to transfer realistic images to coarse images in DINO feature space, and apply it frame-wise to ordinary real videos to synthesize pseudo-coarse videos. The C2R# model then learns to generate each real video from its pseudo-coarse counterpart, free from needing (coarse, realistic) video pairs. At inference, the DINO-R2C model is discarded and actual coarse videos drive the C2R# model directly. Thus, our pair-free training scales with any real-video corpora rather than with paired synthetic videos. We will release the model and project webpage at https://gonzalognogales.github.io/coarse2real-sharp/.

Safe robotic task planning requires planners that can reliably achieve task goals while satisfying safety constraints. Classical planners search over explicit symbolic task representations and can provide formal guarantees with respect to encoded safety constraints, but require rigid input formats. Their computational cost also increases rapidly with the search space. In contrast, large language models (LLMs) offer flexible task interfaces and strong generalization capabilities, but they are large, power-hungry, and pre-trained with little robot-specific safety data and knowledge. Thus, they can be difficult to deploy locally on robotic systems with limited resources. To address these limitations, we present SLoSa, a neuro-symbolic compression and fine-tuning framework for training foundation-model-based robotic task planners with Small size, Low power demand, and Safety awareness. The framework first constructs a safety-aware dataset using classical planners, an advanced teacher LLM, and formal verifiers, followed by two training stages: supervised fine-tuning and reinforcement learning (RL) fine-tuning guided by a verification-based reward machine to enforce safety alignment. We evaluate SLoSa across four Planning Domain Definition Language (PDDL) domains, two household simulation environments, local edge platforms, and a desktop robotic arm. Across the PDDL tasks, verifier-guided post-training raises the success rate of compact 4B-14B planners to 70.5-76.0%, while INT4 deployment on a Jetson Orin requires 17.85-26.41 W of average power. Results in household simulators and on the physical robot further demonstrate the applicability of SLoSa across diverse robotic task settings.

Scanning every passage in a long document seems thorough, but does it preserve more useful evidence or produce better answers? We compare 11 retrieval configurations spanning lexical and dense search, hybrid reranking, and full-document scanning with Jev, OpenAI Decisions, and Cloudflare Clef. Our evaluation uses 396 QASPER questions and 80 questions from eight complete textbooks, with common passages, a five-passage limit, and a fixed answer generator. On books, Hybrid + Jev achieves the highest observed evidence coverage (94.1%), compared with 88.7% for Jev full scan and 81.3\% for OpenAI dense retrieval. Full scans examine every passage initially, yet later reduction steps can discard relevant evidence. Higher evidence coverage also does not consistently produce better answers: Hybrid + Decisions achieves the highest observed book Answer Accuracy among retrieval methods (0.892), despite recovering less reference evidence. Full scanning incurs substantially higher latency on books under the tested deployments, while reusable dense retrieval has lower projected API cost than paid decision-based methods as queries accumulate. Most QASPER articles already fit within the hybrid candidate pool, limiting the benefit of scanning. These results highlight evidence retention, answer quality, and reuse cost as complementary criteria for choosing a long-document retrieval strategy.

Transformers are deployed as a special model class. Neural processing units (NPU) add dedicated operators for attention and normalization, and quantization relies on Transformer-specific algorithms. We show the computation does not force this. Every Transformer block, with attention, LayerNorm and a GELU MLP, compiles exactly into a generic deep neural network built from fixed affine maps, routing and the square, sigmoid and inverse square root, plus GELU or its sigmoid approximation. Without activation-activation products, every transform-wise compilation adds at least $N^2$ scalar nonlinear evaluations per head for dense scores. Keeping that one operation type removes this cost; its elementwise form, present on every surveyed NPU, suffices with the sigmoid and inverse square root. For top-$k$ attention with exchangeable tail scores, we prove that exact or conditional-mean tail mass gives an unbiased, minimum-mean-square-error estimate of the attention distribution, while tail drop is conditionally biased. When queries or agents cluster, centroids choose the keys first, reducing the number of computed scores from $N^2$ to $Nk$ per head. On DeiT-Tiny, MambaIRv2-Light and Pythia-70M the compiled models match their references with all keys, standard post-training quantization applies, and the exact tail-mass estimate beats tail dropping with half as many keys on DeiT-Tiny and, for $k\le8$, on Pythia-70M.

As quantum computing moves from the noisy intermediate-scale quantum (NISQ) era toward early fault-tolerant quantum computing (EFTQC), quantum-enhanced AI is shifting from a conceptual idea to experimental research. Quantum components are expected to extend the capabilities of existing AI systems, yet it remains unclear how such components will help within large multimodal models. In this study, we try to answer that question for vision-language models (VLMs) using classically simulated quantum layers. With LLaVA-OneVision as the backbone, we evaluate 4- and 8-qubit basic-ring and strongly entangling circuits at four insertion and four replacement locations. Insertion adds quantum layers to the classical model, while replacement substitutes them for classical components. Using the same setup, we compare performance and computational cost between classical and quantum-classical hybrid models. On the Visual Spatial Reasoning dataset, inserting a quantum layer after the visual projector reaches 80.7% development accuracy, 2.28 percentage-points above classical full fine-tuning. Replacing the final decoder feed-forward sublayer retains 79.22% accuracy with 13.07 million fewer trainable parameters. Quantum insertion can thus improve accuracy, and selective replacement can reduce parameters. These benefits depend strongly on where the quantum layer is inserted or which classical component it replaces: earlier replacements lower accuracy, and fewer parameters do not reliably shorten training or reduce computation. To assess potential security benefits, we evaluated the robustness of quantum-classical hybrid models under six attack settings. Overall, our study maps where quantum layers help and hurt within VLMs, offering a starting point for hybrid VLM design; whether these gains generalize remains open. The code is available at https://github.com/TITANS-Lab-Artifacts/QuantumVLM.

Large language model (LLM) agents are increasingly deployed in multi-agent systems (MAS) to accomplish complex tasks. Prior black-box red-teaming attacks mainly focus on single agents, but we find that these methods are far less effective in MAS, where multiple sub-agents may not directly interact with the user. Therefore, we introduce ReMAS, the first reinforcement learning-based red-teaming framework tailored to MASs, which fine-tunes attacker LLMs to generate effective adversarial prompts for system prompt extraction. The framework follows a two-step process: first, a rewrite stage refines base attack prompts to increase extraction success, and second, a template generation stage constructs attack templates that improve the likelihood of invoking specific sub-agents and thus revealing their system prompts. Extensive experiments show that our method substantially improves attack success rates compared to existing black-box attacks, with transferability across different backend LLMs and MAS structures. These results underscore the vulnerabilities of multi-agent systems and the importance of developing stronger defenses against adaptive black-box adversaries.

Silicon nanowires (Si NWs) have emerged as promising candidates for next-generation nanoelectronic devices due to their tunable electronic properties, and compatibility with aggressive scaling requirements. The integration of strain further enhances their performance by modulating carrier mobility, band structure, and transport behavior without altering material chemistry. Within the NW, the silicon-silicon dioxide (Si-SiO$_{2}$) interface serves as a critical region where strain redistribution and suboxide bonding collectively influence the mechanical and electronic response. Here, we combine reactive force field molecular dynamics and first-principles calculations to investigate how oxidation-induced interfacial structure and crystallographic facet orientation govern strain-dependent bond evolution and facet-dependent strain response of electronic properties in oxidized Si NWs. Simulated dry thermal oxidation generates chemically realistic Si-SiO$_{2}$ interfaces, enabling direct characterization of bond order redistribution and stress gradient development across the interface under tensile strain. Interfacial Si-Si bonds exhibit elongation and reduced bond order associated with partial oxidation states, leading to localized stress concentration and preferential bond rupture. Complementary density functional theory calculations on facet-specific models reveal that orientation-dependent interfacial bonding modulates facet-dependent strain response of electronic properties. Together, these results establish a direct coupling between interfacial suboxide chemistry, mechanical deformation, and electronic structure, providing atomistic insight into strain-modulated behavior at oxidized silicon interfaces in faceted NWs.

Clinical registries and wearable sensors support early health-risk assessment, yet models trained on these records may leak sensitive information when released. Differential privacy (DP) bounds this leakage by injecting noise whose total magnitude grows with the number of trainable parameters, a cost especially severe for small health cohorts. This motivates parameter efficient models including quantum machine learning (QML), whose variational circuits can represent complex decision boundaries with few trainable parameters. It raises a question: at matched input dimensionality and privacy budget, does a quantum model retain more utility than a classical one? We investigate this question on three clinical cohorts and one wearable cohort, treating each clinical record or participant as a privacy unit. We compare capacity-matched quantum, multilayer perceptron, and logistic regression models across five privacy budgets and five seeds. The quantum model leads on three clinical cohorts by up to 0.104 ROC AUC averaged across budgets; however, trails on the wearable cohort, where the small number of participants requires substantial noise. No difference survives Holm correction across 24 tests, so these results indicate a trend rather than a quantum advantage. An analytic gradient bound shows that lower clipping thresholds can improve utility by reducing noise despite clipping bias. Finally, an experiment on IBM quantum computer (ibm_fez, a 156-qubit superconducting processor) using eight fixed test cases yields mean absolute probability errors of 0.046 (private) and 0.063 (non-private) relative to simulation, supporting the feasibility of inference on noisy intermediate-scale quantum (NISQ) hardware. Code is publicly available at https://github.com/TITANS-Lab-Artifacts/DP-QML-Health.

We prove state evolution for approximate message passing (AMP) with rectangular random matrices whose entries are independent, centered, and variance-normalized. The row and column dimensions grow with a ratio that converges to a positive finite limit. The entry laws may differ across coordinates and dimensions, and we assume uniform integrability of the fourth powers of the rescaled entries. For entries sampled independently from a fixed law, a finite fourth moment suffices, with no requirement for higher moments. The nonlinearities are continuously differentiable and globally Lipschitz, with bounded continuous first state derivatives. The signal and noise have independent and identically distributed coordinates with finite moments of all orders, and are independent of each other and of the matrix. At every fixed number of iterations, empirical averages of pseudo-Lipschitz functions of order two converge in probability to the limits specified by Gaussian state evolution. These averages include empirical second moments and products of trajectory coordinates. The proof combines an aggregate bound on third derivatives with randomized Lindeberg replacement and operator-norm truncation, without matching third moments or invoking a non-Gaussian universality theorem. We also show that independence of complementary projected components for every deterministic rank-one constraint pair characterizes Gaussian entries. This identifies the limitation of extending Gaussian conditioning identities to independent non-Gaussian matrices. Finally, a variance-profile example shows that average variance normalization alone does not ensure the homogeneous state evolution limits.

Enhanced Rock Weathering (ERW) is a carbon removal solution with the potential to remove gigatons of CO$_2$ per year, but quantifying that removal today requires dense soil sampling and laboratory analysis. We evaluate five weather Foundation Models (FMs) as frozen encoders feeding lightweight decoders, under a leave-one-field-out protocol on 5,192 chronosequences from 264 fields across Brazil and the US. To our knowledge, this is the first holistic evaluation of weather FMs for ERW and the largest evaluation on operational ERW data. We find that across weather FMs, representations transfer to ERW: our best model, TerraNova-ERW, built on Aurora 1.5, positively orders measurements within fields it has never seen. Zero-shot prediction of a new field's mean removal is not yet feasible, and we attribute this to data scale: the same architecture excels on $\sim$$10^5$ natural-weathering sequences, anchoring the scale we expect ERW data must reach. Few-shot prediction, blending the model with a few samples per field, is supported today: TerraNova-ERW recovers over 60% of the highest-removal fields with two samples per field, roughly halves the samples needed to estimate a field's mean removal, and can cut sampling on our densest fields by 28-98%.

The scarcity of large annotated datasets limits deep learning (DL) segmentation in material CT analysis. DL-based pseudo-labeling can help, but it is often computationally expensive and hard to interpret. We introduce a multi-stage, model-free framework that generates pseudo-labels directly from grayscale CT scans. It combines thresholding, morphological dilation, and a custom breadth-first search region-growing algorithm to produce polygonal annotations without pre-trained weights, labeled data, or a GPU. We compare it against the zero-shot Segment Anything Model (SAM), a fully supervised 3D U-Net, and classical baselines. On a dendrite dataset with ground truth, our method reaches 63.19% IoU versus 10.74% for zero-shot SAM, with four times the instance recall at 0.5 IoU (33.3% versus 8.3%). The supervised 3D U-Net serves as a reference ceiling that requires expert labels. On a silica sand specimen, YOLO models trained on our labels performed comparably to models trained on SAM labels in a qualitative comparison. Across 1,053 extracted grains, the median equivalent diameter was 163.8 micrometers, near the specimen's known 106 to 150 micrometers sieve range, compared with 219.5 micrometers for SAM, whose labels merge adjacent particles and add spurious detections. A sensitivity analysis shows a stable operating range for the key threshold and a sharp, predictable failure point linked to phase contrast. Our method needs 15s per image on a standard CPU versus 28s for SAM on a GPU. These results support our transparent approach as a tool for annotation bootstrapping and microstructural quantification in resource-limited settings.

Muon orthogonalizes a weight matrix's momentum before each step, and on neural networks this simple preconditioner beats AdamW in a wide range of tasks. Existing analyses explain Muon within general optimization frameworks or on a few specific models, and many of them view orthogonalization as a preconditioner. Since the Hessian is the optimal preconditioner under a quadratic approximation of the loss, these works leave a basic question open: what makes orthogonalization suited to the Hessian of a neural network? Prior work shows that the layer-wise Hessian of a neural network is row-block diagonally dominant, and Muon's implicit preconditioner, the Kronecker product $(\mathbf{V}\mathbf{V}^\top)^{1/2}\otimes\mathbf{I}$ with $\mathbf{V}$ the momentum, has the same row-block structure exactly when $\mathbf{V}\mathbf{V}^\top$ is diagonally dominant. Diagonal dominance of $\mathbf{V}\mathbf{V}^\top$ has been observed in language models, but why a neural network produces it has not been explained. This work provides a quantitative asymptotic theory for neural networks at initialization, and points out the intuition that the overparameterization of the weight matrices drives the diagonal dominance. We analyze $\mathbb{E}[\mathbf{G}\mathbf{G}^\top]$, which equals $\mathbb{E}[\mathbf{V}\mathbf{V}^\top]$ at initialization, in four canonical models: symmetric matrix factorization, deep linear networks, two-layer ReLU networks, and single-layer linear attention. In each case the diagonal entries outgrow the off-diagonal ones at explicit rates as the hidden dimensions grow; for the query, key and value weights of attention the off-diagonal vanishes altogether. Simulations of the four models and experiments on language models at initialization confirm the predicted scaling.

Continuous glucose monitoring supports diabetes management but requires forecasting models that are accurate, interpretable, and efficient enough for real-time deployment. We present SSMCGM-Stream, a streaming state-space model that maintains a personalized hidden state updated with each incoming glucose reading to provide constant-time forecasting, physiologically structured counterfactual reasoning, and interpretable forecasts through hidden-state attribution. Built on this foundation, SSMCGM-Agent delivers risk-aware, natural-language decision support by translating forecasts, explanations, and counterfactuals into grounded recommendations through a guarded language-model interface. Evaluated on the T1DEXI cohort, SSMCGM-Stream achieves forecasting performance comparable to strong neural baselines while substantially reducing computational cost, enabling continuous on-device diabetes decision support.

In-context reinforcement learning (ICRL) exploits the impressive in-context learning ability of large language models to zero-shot generalize to new, unseen tasks at inference time without any parameter updates. Among many approaches, the Decision-Pretrained Transformer (DPT) is particularly appealing for its theoretical grounding: under idealized assumptions such as a sufficiently expressive model and infinite pretraining data, DPT can perform Bayesian posterior sampling. Whether this characterization survives in practice remains open. Therefore, we ask: \textit{is ICRL, specifically DPT and our proposed DPT variants, a Bayesian posterior sampler in practice?} To approach this question, we establish two necessary conditions that a Bayesian posterior sampler must satisfy: \textit{sufficient-statistic invariance} and the \textit{martingale property}. We then construct statistical hypothesis tests to statistically assess if practically-trained ICRL models satisfy these conditions. Applying these tests in Gaussian, Bernoulli, and linear bandit settings, we find that both conditions are consistently violated, suggesting that ICRL may not behave as Bayesian posterior samplers in practice. Beyond this negative result, our tests provide a principled toolkit for testing the Bayesian behavior of ICRL.

Many problems in signal processing, machine learning, and neuroimaging provide several jointly observed measurements of one process, each mixing a component common to all of them with a component private to itself, and require the common component to be recovered without supervision. This problem has been studied in several communities, largely independently. Statisticians proposed canonical correlation analysis (CCA), introduced by Hotelling in 1936. Information theorists proposed two notions of common information, due to Gács and Körner and to Wyner. Engineers developed subspace methods for sensor arrays and multiuser communications to design multiview receivers. This tutorial presents the main results and tools each community developed for this problem in a single exposition, and explains the connections that emerge where they meet, distinguishing exact equivalences from special cases and analogies. We develop the information-theoretic foundation from first principles. We then introduce CCA and its extension to several views, followed by the engineering thread from blind multiuser detection to CCA as a blind receiver. Next, we turn to nonlinear CCA, covering kernel, deep, and probabilistic formulations, the conditions under which the common factor is identifiable, and connections to self-supervised learning. We conclude with a discussion of further extensions and open directions. The exposition is self-contained, so that readers from each of these communities can follow the parts developed by the others.