arXiv:2606.25822v1 Announce Type: cross Abstract: Motivated by sequence reconstruction and reconstruction codes, we study extremal intersections of deletion balls over a fixed $q$-ary alphabet. Let $\Sigma_q^n$ be the set of sequences of length $n$ over $\Sigma_q$, and let $D_t(x)$ denote the set of all sequences obtained from $x\in\Sigma_q^n$ by deleting exactly $t$ symbols. Our first result gives a finite upper bound under a lower-order deletion-correction constraint. We prove that if $x,y\in\Sigma_q^n$ satisfy $D_{s-1}(x)\cap D_{s-1}(y)=\varnothing$, then \[ |D_t(x)\cap D_t(y)| \le \binom{2s}{s}\binom{n-s}{t-s}. \] For binary alphabets, this strengthens a recent asymptotic upper bound of Pham, Goyal, and Kiah (2025, JCTA). We then investigate deletion-ball intersections under simultaneous constraints on run counts and lower-order deletion-ball intersections. For fixed $0<\gamma\le1$, integers $1\le s\le t$, and $m\ge1$, we show that if $x,y\in\Sigma_q^n$ have at most $\gamma n$ runs and satisfy $|D_s(x)\cap D_s(y)|\le m$, then \[ |D_t(x)\cap D_t(y)|\le \frac{m\gamma^{t-s}}{(t-s)!}n^{t-s}+O_{s,t,m}(n^{t-s-1}). \] Moreover, the leading term can be attainable whenever $m$ is realized by a fixed finite-length seed pair. As a consequence, we obtain a direct lifting theorem for deletion reconstruction codes, transferring reconstruction properties from radius $s$ to larger radii $t$. Finally, we establish a parallel insertion theory and derive corresponding results for insertion-ball intersections and insertion reconstruction codes.
Science Journals
arXiv:2606.25930v1 Announce Type: cross Abstract: Titanium dioxyde (TiO2) is a promising dielectric material for the realization of metamaterials operating in the terahertz (THz) range. Indeed, these necessitate a high permittivity and low loss material. In this paper, we compare the processes of fabrication and the results of characterisation of bulk TiO2 pellets. From the results of this characterization, we have numerically designed 2D all dielectric metamaterials (ADM) showing that they may exhibit negative or near-zero effective index. Our previous simulations show that the relative permittivity epsilon has to be around 100, while the loss tangent has to be lower than 0.02. We have thus compared conventional sintering (CS) vs spark plasma sintering (SPS), and investigated the effect of post-sintering annealing on the loss to satisfy these two criteria. The samples were characterized by THz Time Domain Spectroscopy (THz-TDS). One of the samples exhibits a loss tangent as low as 0.006, with a permittivity epsilon = 103. These results highlight the importance of the fabrication process on the EM properties of bulk TiO2, and demonstrate that it is a promising material for the development of metamaterial in the THz band.
arXiv:2606.25687v1 Announce Type: cross Abstract: Straight dark solitons in two-dimensional (2D) quantum fluids usually decay by transverse modulational instability, with no intrinsic suppression in contact-interacting Bose--Einstein condensates (BECs). We theoretically show that anisotropic long-range interactions in a quasi-2D dipolar BEC stabilize an embedded straight soliton, with spontaneous stripe order providing stronger pinning. The excitation spectra show that the lowest transverse solitonic branch remains gapped, while stripe-supersolid density modulation further hardens this branch and increases the soliton bending stiffness, penalizing transverse deformation. Accessible in current $^{166}$Er and $^{164}$Dy platforms, these results establish interaction-driven protection for straight dark solitons in structured quantum fluids.
arXiv:2606.25745v1 Announce Type: cross Abstract: Mean Field Variational Inference (MFVI) is widely understood to underestimate posterior variance. By analysing conjugate Bayesian Linear Regression (BLR), we show that this characterization is incomplete: while MFVI underestimates the variance in parameter space, it can overestimate the predictive variance compared to the exact posterior. We show that if the MFVI posterior underestimates predictive variances in some directions, it necessarily overestimates them in others. Crucially, this overestimation occurs in directions where the training data concentrates. This leads to the surprising result that, for a test point drawn from the training distribution, MFVI's expected predictive variance exceeds that of the exact posterior. We demonstrate a pathological case of this effect, where the MFVI posterior fails to reduce predictive variance compared to the prior on in distribution data. We connect these results to the Cold Posterior Effect, arguing that varying the temperature can correct this overestimation, yielding predictions closer to those of the exact posterior. We validate our theory on synthetic and real-world regression tasks.
arXiv:2606.25400v1 Announce Type: new Abstract: Brain-Computer Interfaces (BCIs) and brain signal understanding are pivotal for clinical health and next-generation interactions. Despite this significance, its widespread adoption in real-world scenarios remains restricted, primarily because current analytical paradigms lack sufficient agentic intelligence. First, existing methodologies impose prohibitive technical barriers, requiring extensive specialized expertise. Second, they remain inherently static and task-specific, failing to execute the complex, long-horizon workflows essential for real-world deployment. To accelerate the democratization of brain signal understanding, we draw inspiration from Large Language Models (LLMs) to introduce BrainAgent, an LLM-driven multi-agent framework designed to ground abstract natural language intent into rigorous, executable, and end-to-end processing pipelines. BrainAgent employs a hierarchical architecture where a central supervisor orchestrates specialized sub-agents for adaptive task decomposition and execution. Furthermore, we establish a comprehensive, systematic benchmark for evaluating agentic systems in brain signal analysis. Empirical results demonstrate that BrainAgent effectively automates complex workflows with superior reliability, marking a paradigm shift toward democratized brain signal understanding.
arXiv:2606.25838v1 Announce Type: new Abstract: Production vision pipelines silently degrade on blurry input, wasting compute on downstream OCR, retrieval, and vision-language model (VLM) calls that cannot recover a usable output. We present MagikaDocumentFromPixel, a lightweight, CPU-friendly image quality gate that classifies a single image as sharp, blurred, or uncertain in roughly 7 ms on a single CPU core. The contributions are (i) a recipe selected from a 46-configuration, 8-sweep empirical search that isolates input resolution as the dominant lever and shows architecture capacity only pays off at >= 384 px; (ii) a confidence-aware routing formalism grounded in classical selective prediction; (iii) the Edge Prior Module (EPM), a Laplacian-magnitude auxiliary input channel that gives the network direct access to the spectral evidence that classical blur heuristics rely on and that lifts test F1 by +1.3 points in a matched-env comparison; and (iv) an observation that the gate is one instance of a recurring design pattern that appears independently in Magika content-type detection, risk-controlled OCR with VLMs, and DocVLM. The final recipe MobileNetV3-Large with the EPM trained at 384x384 on paired GoPro Large frames, evaluated with 5-scale test-time augmentation reaches F1 = 0.9803 (AUC 0.9989) with a 17 MB ONNX artifact, improving over our fixed-scale baseline on the same hardware (F1 = 0.9672) by +1.31 points. We are explicit about limitations: results are on a single motion-blur distribution, numbers are from a single seed, and calibration is qualitative rather than measured.
arXiv:2606.25844v1 Announce Type: new Abstract: A common practice converts a one-dimensional signal into an image so that a vision backbone pretrained on natural photographs can be reused for recognition, yet the encoded image is rarely examined. We ask how the visual naturalness of an encoded image relates to its transfer accuracy under a frozen backbone. We build WorldStream, a corpus of 299 heterogeneous current-value series from key-free public APIs (weather, air quality, earthquakes, gold and oil, equities, crypto, foreign exchange, web activity and space weather), with a nine-way source-recognition task over 3143 temporally split windows. Across seven encodings and six frozen backbones, the Frechet distance of an encoding to natural images (FID) predicts its accuracy: Spearman $\rho=-0.72$. Two controlled interventions show this is not causal in the spectrum. Our invertible encoder has a single adjustable part, a spectral exponent $\beta$ (power $\propto |f|^{-\beta}$); varying $\beta$ moves the image toward or away from the natural-image manifold at fixed content. FID is lowest near the natural value $\beta \approx 2$, but frozen accuracy stays flat and far below the structured baselines (19.2% vs. 73.0%), and FID and accuracy are only weakly related over the sweep (Pearson $-0.32$). A second intervention, phase scrambling, holds the power spectrum exactly fixed while removing local structure; now FID and accuracy fall together (Pearson $-0.89$). The cross-encoding correlation is thus mediated by local structure, not spectral naturalness: FID predicts accuracy because Inception reads the same structure the backbones do. Full fine-tuning does not close the gap (27% vs. 67%), so the deficit is structural. The encoder is exactly invertible, recovering the signal from the 8-bit image at 72.9 dB, so the image doubles as a lossless record of the data.
arXiv:2510.00190v2 Announce Type: replace Abstract: We present an automated approach for identifying magnetospheric regions using supervised machine learning techniques applied to Magnetospheric MultiScale mission data. Our method utilizes ion energy spectra, total magnetic field, total ion temperature, ion velocity component, ion density and spacecraft position data to classify five distinct plasma environments: solar wind, magnetosheath, inner magnetosphere, plasma sheet, and lobe regions. The approach combines a convolutional neural network (CNN) for analyzing ion energy spectrogram data with a Random Forest classifier for scalar plasma parameters. The CNN method employs 2D convolution to identify spatial and temporal patterns in the ion energy spectrogram treated as image-like data, while the Random Forest model processes averaged magnetic field, temperature, velocity, density and position parameters. Our hybrid model achieves 99% accuracy on test dataset with an F1 score of 0.99, providing reliable automated region identification at 3-minute temporal resolution. This lightweight approach requires minimal manual data labeling and can be readily applied to other magnetospheric missions with similar data products.
arXiv:2606.25863v1 Announce Type: new Abstract: We study how security patches in highly configurable C/C++ systems map onto the space of compile-time variants. We formalize the Vulnerability Impact Condition (VIC) - a Boolean predicate over configuration options that denotes all variants that contained the original flaw - and introduce PatchLens, a purely static technique that recovers VICs by aligning AST-level patch hunks with source-level presence conditions and resolving file inclusion via lightweight build system analysis. Evaluating PatchLens on 1,192 Linux kernel, 289 FFmpeg, and 100 PHP patches, we compute precise, human-readable VICs without the need to compile any system variant. The resulting predicates are compact (avg. 1.84 variables for Linux, 3.23 for FFmpeg, 1.04 for PHP) and show that only a small fraction of vulnerabilities are system-wide, which carry higher CVSS scores; meanwhile, CVE texts almost never encode the required options ($\approx$ 1% average recall), motivating automated enrichment of CVE descriptions with VICs. PatchLens and the accompanying dataset enable immediate applications in CI (variant-aware triage and test selection), targeted sampling and fuzzing, and feature risk scoring, offering a scalable, explainable path to vulnerability assessment in highly configurable software.
arXiv:2606.25867v1 Announce Type: new Abstract: Stakeholder interviews are an important source of information for requirements elicitation, yet many relevant requirements remain implicit in such conversations. Stakeholders frequently describe workflows, challenges, and operational practices without explicitly articulating the software capabilities that could address them. Recent work has considered the use of LLMs to analyze conversational data and extract requirements from stakeholder interviews. Existing approaches, however, primarily focus on identifying explicitly stated requirements, leaving implicit opportunities largely unexplored. In this paper, we present LENS (LLM-Enabled Needs Discovery from Stakeholder Interviews), an approach that analyzes stakeholder interview transcripts to both extract explicit requirements and infer additional latent requirements. LENS performs this inference by reasoning over stakeholder statements together with contextual information about organizational tools and infrastructure. Both extracted and inferred requirements are represented as user stories and linked to transcript excerpts to ensure traceability. We conduct a preliminary evaluation of LENS using twelve stakeholder interview transcripts collected in an industrial setting involving cybersecurity operations. We show that LENS achieves an average F1-score of 84.4% for extracting explicit requirements, while, on average, 75% of the latent requirements identified by LENS were perceived as providing useful automation or time-saving potential by domain experts.
arXiv:2606.26063v1 Announce Type: cross Abstract: Electronic excitations in solids are commonly described within a hierarchy in which the excitonic Hamiltonian is defined first and the lattice acts later through renormalization, relaxation, and dephasing. This picture assumes that the optically accessible excitonic manifold is already present at the moment of photoexcitation. Here we show that this assumption fails in a soft polar semiconductor. Using femtosecond coherent multidimensional spectroscopy on lead-halide perovskite nanocrystals, we observe quantum back-action between an electronic excitation and a collective lattice-polarization field that expands the excitonic Hilbert space in real time. The optical pulse first prepares an excitonic polarization, X1. A second configuration, X2, emerges only after the polaron field develops, while coherent X1-X2 coupling appears at later times. State formation and coherence formation are therefore resolved as distinct stages of quasiparticle formation. In contrast, CdSe quantum dots exhibit the conventional limit in which excitonic states and couplings are present at time zero and are only weakly perturbed by phonons. The observed diagonal and anti-diagonal splittings increase with nanocrystal size and correlate with radiative oscillator strength, opposite to expectations from simple quantum confinement. A dynamical polaron-field model describes the lattice polarization as an order parameter that expands the optically accessible manifold and generates time-dependent coherent coupling. These results show that strong system-bath coupling can actively create excitonic states and the coherent manifold in which they evolve.
arXiv:2006.16511v2 Announce Type: replace Abstract: A set $S$ of vertices of a graph $G$ is a \emph{geodetic set} if every vertex of $G$ lies in a shortest path between some pair of vertices of $S$. The \textsc{Minimum Geodetic Set (MGS)} problem is to find a geodetic set with minimum cardinality of a given graph. A \emph{grid embedding} of a graph is a set of points in two dimensions with integer coordinates such that each point in the set represents a vertex of the graph and, for each edge, the points corresponding to its endpoints are at Euclidean distance~$1$. A graph is a \emph{partial grid} if it has a grid embedding. In this paper, we first prove that \textsc{Minimum Geodetic Set} remains NP-hard even for subcubic partial grids of arbitrary girth. This jointly strengthens three existing hardness results: for bipartite graphs (Dourado et al., Discrete. Math, 2010), subcubic graphs (Bueno et al., Inf. Process. Lett., 2018)~\cite{bueno2018}, and planar graphs (Chakraborty et al., CALDAM, 2020). The \emph{area} of an internal face is the number of integer points lying on the boundary or interior of the face. A graph is a \emph{solid grid} if it has a grid embedding such that all interior faces have area exactly four. To complement the above hardness result, we design a linear-time algorithm for \textsc{Minimum Geodetic Set} on solid grids, improving on a $3$-approximation algorithm by Chakraborty et al. (CALDAM, 2020). Our results hold for \textsc{Edge Geodetic Set} as well. A set $S$ of vertices of a graph $G$ is a \emph{geodetic set} if every edge of $G$ lies in a shortest path between some pair of vertices of $S$. The \textsc{Minimum Edge Geodetic Set (MEGS)} problem is to find an edge geodetic set with minimum cardinality of a given graph. As corollaries, we obtain that \textsc{MEGS} remains NP-hard on partial grids and is linear-time solvable on solid grids.
arXiv:2206.04359v3 Announce Type: replace Abstract: One of the fundamental challenges in the deep learning community is to theoretically understand how well a deep neural network generalizes to unseen data. However, current approaches often yield generalization bounds that are either too loose to be informative of the true generalization error or only valid to the compressed nets. In this study, we present a simple yet non-vacuous generalization bound from the optimization perspective. We achieve this goal by leveraging that the hypothesis set accessed by stochastic gradient algorithms is essentially fractal-like and thus can derive a tighter bound over the algorithm-dependent Rademacher complexity. The main argument rests on modeling the discrete-time recursion process via a continuous-time stochastic differential equation driven by fractional Brownian motion. Numerical studies demonstrate that our approach is able to yield plausible generalization guarantees for modern neural networks such as ResNet and Vision Transformer, even when they are trained on a large-scale dataset (e.g. ImageNet-1K).
arXiv:2306.01140v4 Announce Type: replace Abstract: This work is concerned with the analysis of a space-time finite element discontinuous Galerkin method on polytopal meshes (XT-PolydG) for the numerical discretization of wave propagation in coupled poroelastic-elastic media. The mathematical model consists of the low-frequency Biot's equations in the poroelastic medium and the elastodynamics equation for the elastic one. To realize the coupling, suitable transmission conditions on the interface between the two domains are (weakly) embedded in the formulation. The proposed PolydG discretization in space is then coupled with a dG time integration scheme, resulting in a full space-time dG discretization. We present the stability analysis for both the continuous and the semidiscrete formulations, and we derive error estimates for the semidiscrete formulation in a suitable energy norm. The method is applied to a wide set of numerical test cases to verify the theoretical bounds. Examples of physical interest are also presented to investigate the capability of the proposed method in relevant geophysical scenarios.
arXiv:2312.11061v3 Announce Type: replace Abstract: This technical note examines exponential stability of the null solution to a large class of compartmental systems governed by ordinary differential equations. Sufficient conditions under which these systems admit a linear Lyapunov function are provided. The coefficients of the Lyapunov functions and the exponential decay rate they yield are obtained from an eigenvalue problem. For a special case of the system class considered, we derive an equivalence between attractivity of the null solution and the existence of a linear Lyapunov function.
arXiv:2511.08659v4 Announce Type: replace Abstract: This book is an introductory textbook targeted towards computer science students who are completely new to the topic of automated negotiation. It does not require any prerequisite knowledge, except for elementary mathematics and basic programming skills. This book comes with an simple toy-world negotiation framework implemented in Python that can be used by the readers to implement their own negotiation algorithms and perform experiments with them. This framework is small and simple enough that any reader who does not like to work in Python should be able to re-implement it very quickly in any other programming language of their choice.
arXiv:2405.17423v4 Announce Type: replace Abstract: As Visual Language Models (VLMs) become increasingly embedded in everyday applications, ensuring they can recognise and appropriately handle privacy-sensitive content is thus essential to protect users. To this end, we conduct a comprehensive evaluation of twelve state-of-the-art VLMs and identify limitations in their understanding of visual privacy. However, existing privacy-related datasets often suffer from label inconsistencies, limiting their reliability. To address this, we introduce two compact, high-quality benchmarks, PrivBench and PrivBench-H, that focus on commonly recognised visual privacy categories aligned with the General Data Protection Regulation (GDPR). Additionally, we present PrivTune, an instruction-tuning dataset specifically curated to improve privacy sensitivity. We obtain multiple Privacy VLMs by fine-tuning off-the-shelf VLMs on only a few hundred samples from PrivTune, which leads to substantial gains on all benchmarks, surpassing even GPT-4, while maintaining strong performance on other tasks. Our findings show that privacy-awareness in VLMs can be substantially improved with minimal data and careful dataset design, setting the stage for safer, more privacy-aligned AI systems.
arXiv:2606.25559v1 Announce Type: new Abstract: Decentralised systems are often built to avoid central control, but their evolution almost always depends on centralised platforms, informal maintainer authority, and a surprising amount of unpaid goodwill. To address this uncomfortable mismatch, we introduce TwoStepDemocracy, a technical proof-of-concept for protocol-native software evolution. The prototype combines costly cryptographic identities, peer-to-peer dissemination, issue and solution voting, and Bitcoin-based funding campaigns. Users can express demand by proposing and voting on issues; developers can submit concrete solutions; and accepted work can be linked to voluntary, non-custodial funding. The design deliberately separates demand, approval, and payment. This way, money can support a solution, but it never buys more voting power. The prototype demonstrates that such a coordination layer can be built as a peer-to-peer implementation with local storage, signed governance objects, and Bitcoin integration. We studied performance, scalability, and costs across storage, identity management, and funding. The results show technical feasibility, but not yet social viability. A larger user study is still needed to evaluate whether real communities would, in practice, vote, fund, and coordinate through this mechanism.
arXiv:2606.25481v1 Announce Type: new Abstract: The state of the tunneling electron under the potential barrier is important in strong laser-atom interaction but is difficult to identify. Recent experiments showed that the tunneling electron may be located in a bound state with high symmetry [Phys. Rev. Lett. 134, 213201 (2025)]. However, the quantitative characteristic of the tunneling state in a tunneling event remains unclear. Here, we study tunneling ionization of atoms in strong circular laser fields. The calculated photoelectron momentum distribution (PMD) through numerical solution of time-dependent Schr\"odinger equation (TDSE) presents an isotropic ring-shaped distribution and the most probable momentum (MPM) along the ring can be easily identified. The kinetic energy related to MPM is remarkably smaller than that predicted by the strong-field approximation (SFA) that ignores Coulomb potential. Surprisingly, for different target atoms and laser parameters, the kinetic energy difference of MPM between TDSE and SFA is always close to half of the corresponding Coulomb potential at the tunnel exit. This phenomenon can be well described by a proposed model, which indicates that the tunneling electron is in an exit-position-dependent quasibound state agreeing with the virial theorem. These results quantitatively reveal the characteristics of tunneling states from a statistical perspective.
arXiv:2606.25562v1 Announce Type: new Abstract: This paper presents an energy-efficient hardware acceleration of the convolutional layers in the U-Net architecture for image segmentation, implemented on FPGA. While digit-serial arithmetic, particularly most-significant-digit-first (MSDF) techniques, offers a compact hardware footprint, it suffers from initial latency before producing the first output digit. This delay accumulates in cascaded operations like multiplication followed by addition, where each unit introduces its own startup overhead. To overcome this, we propose a merged multiply-add (MMA) architecture that fuses these operations into a unified pipeline. Instead of incurring separate delays, the MMA introduces a single streamlined latency per iteration, shorter than the combined latency of conventional cascaded units, resulting in enhanced throughput and efficiency. The MMA units are designed to process spatial input depths in parallel, achieving significantly higher performance than both standalone MSDF-based and conventional designs. We evaluate the proposed design using U-Net as a target application. Despite operating at a lower frequency than a CPU, the FPGA-based accelerator achieves up to an order of magnitude higher energy efficiency, delivering up to $15.14$ GOPS/W compared to $1.93$ GOPS/W for CPU-based inference. The design also shows approximately $9\times$ reduction in energy consumption compared to MSDF-based FPGA implementations. These results highlight the efficacy of the merged arithmetic approach for resource-constrained, latency-sensitive edge applications in medical imaging and computer vision.
arXiv:2501.02970v3 Announce Type: replace Abstract: With the growing popularity of blockchains, modern chained BFT protocols combining chaining and leader rotation to obtain better efficiency and leadership democracy have received increasing interest. Although the efficiency provisions of chained BFT protocols have been thoroughly analyzed, the leadership democracy has received little attention in prior work. In this paper, we scrutinize the leadership democracy of four representative chained BFT protocols, especially under attack. To this end, we propose a unified framework with two evaluation metrics, i.e., chain quality and censorship resilience, and quantitatively analyze chosen protocols through the Markov Decision Process (MDP). With this framework, we further examine the impact of two key components, i.e., voting pattern and leader rotation on leadership democracy. Our results indicate that leader rotation is not enough to provide the leadership democracy guarantee; an adversary could utilize the design, e.g., voting pattern, to deteriorate the leadership democracy significantly. Based on the analysis results, we propose customized countermeasures for three evaluated protocols to improve their leadership democracy with only slight protocol overhead and no change of consensus rules. We also discuss future directions toward building more democratic chained BFT protocols.
ACT-JEPA: Novel Joint-Embedding Predictive Architecture for Efficient Policy Representation Learning
arXiv:2501.14622v5 Announce Type: replace Abstract: Learning efficient representations for decision-making policies is a challenge in imitation learning (IL). Current IL methods require expert demonstrations, which are expensive to collect. Additionally, they are not explicitly trained to understand the environment. Consequently, they have underdeveloped world models. Self-supervised learning (SSL) offers an alternative, as it can learn a world model from diverse, unlabeled data. However, most SSL methods are inefficient because they operate in raw input space. In this work, we propose ACT-JEPA, a novel architecture that unifies IL and SSL to enhance policy representations. It is trained end-to-end to jointly predict 1) action sequences and 2) latent observation sequences. To learn in latent space, we utilize Joint-Embedding Predictive Architecture, which allows the model to filter out irrelevant details and learn a robust world model. We evaluate ACT-JEPA in different environments and across multiple tasks. Our results show that it outperforms the strongest baseline in all environments. ACT-JEPA achieves up to 40% improvement in world model understanding and up to 10% higher task success rate. Finally, we show that predicting latent observation sequences effectively generalizes to predicting action sequences. This work demonstrates how integrating IL and SSL leads to efficient policy representation learning, an improved world model, and a higher task success rate.
arXiv:2501.18916v2 Announce Type: replace Abstract: Recent work has demonstrated the potential of large language models (LLMs) for program optimization, a key challenge in programming languages. We propose a blackbox adaptation method called Retrieval Augmented Search (RAS) that performs beam search over candidate optimizations; at each step, it retrieves in-context examples from a given training dataset of slow-fast program pairs to guide the LLM. Critically, we find that performing contextual retrieval based on an LLM-generated natural language description significantly outperforms retrieval based on the source code. We also propose AEGIS, a method for improving interpretability by decomposing training examples into ''atomic edits'' that are significantly more incremental in nature. We show that RAS performs up to 2.06$\times$ better than prior state-of-the-art blackbox adaptation strategies on optimizing C++ programs, and that AEGIS performs up to 1.37$\times$ better while making significantly smaller edits. We also show that using RAS improves the mean runtime percentile of Python programs by 10.27 compared to baselines.
arXiv:2502.02205v4 Announce Type: replace Abstract: The application of deep learning for partial differential equation (PDE)-constrained control is gaining increasing attention. However, existing methods rarely consider safety requirements crucial in real-world applications. To address this limitation, we propose Safe Diffusion Models for PDE Control (SafeDiffCon), which introduce the uncertainty quantile as model uncertainty quantification to achieve optimal control under safety constraints through both post-training and inference phases. Firstly, our approach post-trains a pre-trained diffusion model to generate control sequences that better satisfy safety constraints while achieving improved control objectives via a reweighted diffusion loss, which incorporates the uncertainty quantile estimated using conformal prediction. Secondly, during inference, the diffusion model dynamically adjusts both its generation process and parameters through iterative guidance and fine-tuning, conditioned on control targets while simultaneously integrating the estimated uncertainty quantile. We evaluate SafeDiffCon on three control tasks: 1D Burgers' equation, 2D incompressible fluid, and controlled nuclear fusion problem. Results demonstrate that SafeDiffCon is the only method that satisfies all safety constraints, whereas other classical and deep learning baselines fail. Furthermore, while adhering to safety constraints, SafeDiffCon achieves the best control performance. The code can be found at https://github.com/AI4Science-WestlakeU/safediffcon.
arXiv:2502.13334v2 Announce Type: replace Abstract: Motivated by the recent popularity of machine learning training services, we introduce a contract design problem in which a provider sells a service that results in an outcome of uncertain quality for the buyer. The seller has a set of actions that lead to different distributions over outcomes. We focus on a setting in which the seller has the ability to commit to an action and the buyer is free to accept or reject the outcome after seeing its realized quality. We propose a two-stage payment scheme where the seller designs a menu of contracts, each of which specifies an action, an upfront price and a vector of outcome-dependent usage prices. Upon selecting a contract, the buyer pays the upfront price, and after observing the realized outcome, the buyer either accepts and pays the corresponding usage price, or rejects and is exempt from further payment. We show that this two-stage payment structure is necessary to maximize profit: only upfront price or only usage prices is insufficient. We then study the computational complexity of computing a profit-maximizing menu. While computing the optimal profit is NP-hard even for two buyer types, we derive a fully-polynomial time approximation scheme (FPTAS) for the optimal profit for a constant number of buyer types. Finally, we prove that in the single-parameter setting in which buyers' valuations are parametrized by a single real number, there exists a revenue-optimal menu consisting of a single contract.