arXiv:2606.06984v1 Announce Type: new
Abstract: This paper presents a general acceleration mechanism for multi-objective Bayesian optimisation (MOBO) that leverages Gaussian process predictive gradients as auxiliary signals. Rather than replacing existing Pareto-compliant acquisition functions, the proposed approach augments them with local stationarity information derived from surrogate-derived gradients, enabling faster convergence toward the global Pareto set under limited evaluation budgets. Two catalyst instantiations are investigated: an adaptive Multiple-Gradient Descent Algorithm-Based Catalyst (MGDA) and a predefined-weight variant that enables focused exploration when budgets are tight. Experiments on the DTLZ benchmark suite (using 2 objectives and 10 decision variables) show that predictive gradient catalysis can deliver significant acceleration compared to other acquisition functions (EHVI, AugTch, tMPoI, SAF) when surrogates are accurate, particularly for stationary problems.
Science Journals
arXiv:2602.14209v2 Announce Type: replace
Abstract: Block diffusion LLMs are an emerging paradigm for parallel language generation, but their KV caching makes memory access the dominant bottleneck in long-context inference. Sparse attention, which attends only to a small KV subset per query, can reduce this latency with minimal accuracy loss. In block diffusion, however, the B tokens of each block must share a single KV subset, and we show this per-block constraint degrades existing sparse KV estimators by up to 25% in recall. We address this challenge by exploiting a property that emerges from the block-diffusion training objective: it aligns the block-average query across denoising steps, so the All-[MASK] block at the first step already reveals the per-block KV subset for the entire trajectory. We exploit this in MAGE ([MASK]-Guided Sparse Attention), a training-free method that runs one exact attention pass at the first step and reuses its top-k index sets for all remaining steps within the block. Across three block-diffusion families on LongBench, MAGE matches Exact Attention at k=512 with near-lossless accuracy, achieves up to 6.82x end-to-end speedup at 128K context, and runs up to 3.35x and 2.28x faster than Quest and SparseD, designed for AR LLMs and fully bidirectional diffusion LLMs, respectively.
arXiv:2602.16073v2 Announce Type: replace
Abstract: Developing autonomous driving systems for complex traffic environments requires balancing multiple objectives, such as avoiding collisions, obeying traffic rules, and making efficient progress. In many situations, these objectives cannot be satisfied simultaneously, and explicit priority relations naturally arise. Also, driving rules require context, so it is important to formally model the environment scenarios within which such rules apply. Existing benchmarks for evaluating autonomous vehicles lack such combinations of multi-objective prioritized rules and formal environment models. In this work, we introduce ScenicRules, a benchmark for evaluating autonomous driving systems in stochastic environments under prioritized multi-objective specifications. We first formalize a diverse set of objectives to serve as quantitative evaluation metrics. Next, we design a Hierarchical Rulebook framework that encodes multiple objectives and their priority relations in an interpretable and adaptable manner. We then construct a compact yet representative collection of scenarios spanning diverse driving contexts and near-accident situations, formally modeled in the Scenic language. Experimental results show that our formalized objectives and Hierarchical Rulebooks align well with human driving judgments and that our benchmark effectively exposes agent failures with respect to the prioritized objectives. Our benchmark can be accessed at https://github.com/BerkeleyLearnVerify/ScenicRules/.
arXiv:2606.06985v1 Announce Type: new
Abstract: Code-switching (CS), the alternation between multiple languages within a single utterance, remains challenging for Automatic Speech Recognition (ASR). To address this issue, we propose a Point-of-Interest (POI)-aware contrastive training framework that improves recognition at CS-critical regions. We first identify CS spans by adopting POI detection method from literature, then construct acoustically plausible near-miss hypotheses by perturbing POIs in ASR N-best outputs and expanding candidates with a large language model. Hard but plausible negatives are retained through filtering with acoustic, phonemic, and textual constraints. Finally, we fine-tune Whisper-small with LoRA using a POI-weighted cross-entropy anchor objective together with a multi-negative contrastive ranking loss. Experiments on CS-FLEURS (cmn-eng) and ViMedCSS (vie-eng) show consistent reductions of over 2% in both general and CS-aware error rates compared to standard LoRA fine-tuning.
arXiv:2606.02084v2 Announce Type: replace-cross
Abstract: In magnetic confinement fusion reactors, hydrogen plasma irradiation causes material saturation and recycling, where hydrogen released from the tungsten wall significantly impacts the peripheral plasma. Kinetic Monte Carlo (kMC) simulations are essential for investigating the dynamic balance between incident and emitted fluxes at the atomic scale. However, standard kMC frameworks are inadequate for handling realistic material complexities, such as polycrystalline structures and dynamic evolution under irradiation, being computationally bottlenecked by continuous transition parameter updates. Conventionally, evaluating migration barriers in disordered systems (e.g., grain boundaries) relies on computationally prohibitive on-the-fly atomistic calculations like the Nudged Elastic Band (NEB) method.
Here, we present a deep learning-accelerated Dynamic kMC framework that eliminates this reliance. Our approach integrates a three-stage deep learning pipeline: a pix2pix model for predicting local 3D potential energy distributions, a U-Net for extracting hydrogen trapping sites, and a 3D-CNN for directly evaluating migration barriers. To achieve macroscopic timescales, we implemented a hierarchical spatial index combined with a differential local-update algorithm operating in O(1) complexity. This architecture restricts recalculations to the immediate vicinity of moving atoms, accelerating updates. Demonstrated on a large-scale realistic polycrystalline tungsten model, the framework successfully reproduces preferential hydrogen trapping along grain boundaries, bridging the gap between atomic-scale accuracy and macroscopic timescales for full-scale plasma-wall interaction simulations.
arXiv:2602.16864v2 Announce Type: replace
Abstract: Time series (TS) modeling has come a long way from early statistical, mainly linear, approaches to the current trend in TS foundation models. With a lot of hype and industrial demand in this field, it is not always clear how much progress there really is. To advance TS forecasting and analysis to the next level, here we argue that the field needs a dynamical systems (DS) perspective. TS of observations from natural or engineered systems almost always originate from some underlying DS, and arguably access to its governing equations would yield theoretically optimal forecasts. This is the promise of DS reconstruction (DSR), a class of ML/AI approaches that aim to infer surrogate models of the underlying DS from data. But models based on DS principles offer other profound advantages: Beyond short-term forecasts, they enable to predict the long-term statistics of an observed system, which in many practical scenarios may be the more relevant quantities. DS theory furthermore provides domain-independent theoretical insight into mechanisms underlying TS generation, and thereby will inform us, e.g., about upper bounds on performance of any TS model, generalization into unseen regimes as in tipping points, or potential control strategies. After reviewing some of the central concepts, methods, measures, and models in DS theory and DSR, we will discuss how insights from this field can advance TS modeling in crucial ways, enabling better forecasting with much lower computational and memory footprints. We conclude with a number of specific suggestions for translating insights from DSR into TS modeling.
arXiv:2602.18838v2 Announce Type: replace
Abstract: Light springs are space-time beams that have a helical wavepacket. Due to this special property, light springs result into a rotating pulse when intercepting a plane lying orthogonal to their propagation direction. Associated to this, we introduce here the orbital group velocity, an additional tunable property of light springs. The orbital group velocity quantifies the speed of the light spring intensity rotation, distinctly from the conventional longitudinal group velocity, which describes the motion of the wavepacket envelope along its propagation axis. We demonstrate experimentally by tunable Fourier synthesis that the orbital group velocity can assume sub- and superluminal values, thus becoming a new platform for synthetic motion studies and control of laser-matter interactions. Particularly, in the superluminal regime, when interacting with a thin overdense plasma, we reveal by particle-in-cell simulations that the light spring unlocks superradiant radiation, due to the coherent excitation of the electrons in the plasma acting as a quasiparticle. This superradiant source inherits the ultrafast temporal dynamics of the light springs while emitting in the terahertz region, thus creating a new source of terahertz radiation controlled by the properties of spatiotemporal coupling of the laser. Therefore, spatiotemporal tuning of light springs is at the frontier of controlling laser-matter interaction and generating new tunable sources of radiation.
arXiv:2602.19213v2 Announce Type: replace
Abstract: Medical image segmentation is vital for clinical diagnosis and quantitative analysis, yet remains challenging due to the heterogeneity of imaging modalities and the high cost of pixel-level annotations. Although general interactive segmentation models like SAM have achieved remarkable progress, their transfer to medical imaging still faces two key bottlenecks: (i) the lack of adaptive mechanisms for modality- and anatomy-specific tasks, which limits generalization in out-of-distribution medical scenarios; and (ii) current medical adaptation methods fine-tune on large, heterogeneous datasets without selection, leading to noisy supervision, higher cost, and negative transfer. To address these issues, we propose SegMoTE, an efficient and adaptive framework for medical image segmentation. SegMoTE preserves SAM's original prompt interface, efficient inference, and zero-shot generalization while introducing only a small number of learnable parameters to dynamically adapt across modalities and tasks. In addition, we design a progressive prompt tokenization mechanism that enables fully automatic segmentation, significantly reducing annotation dependence. Trained on MedSeg-HQ, a curated dataset less than 1% of existing large-scale datasets, SegMoTE achieves SOTA performance across diverse imaging modalities and anatomical tasks. It represents the first efficient, robust, and scalable adaptation of general segmentation models to the medical domain under extremely low annotation cost, advancing the practical deployment of foundation vision models in clinical applications.
arXiv:2603.02220v2 Announce Type: replace
Abstract: Time series forecasting remains a challenging problem due to the intricate entanglement of intra-period fluctuations and inter-period trends. While recent advances have attempted to reshape 1D sequences into 2D period-phase representations, they suffer from two principal limitations. Firstly, treating reshaped tensors as static images results in a topological mismatch, as standard spatial operators sever chronological continuity at grid boundaries. Secondly, relying on uniform fixed-size representations allocates modeling capacity inefficiently and fails to provide the adaptive resolution required for compressible, non-stationary temporal patterns. To address these limitations, we introduce TimeGS, a novel framework that fundamentally shifts the forecasting paradigm from regression to 2D generative rendering. By reconceptualizing the future sequence as a latent 2D temporal surface, TimeGS utilizes the inherent anisotropy of Gaussian kernels to adaptively model complex variations with flexible geometric alignment. To realize this, we introduce a Multi-Basis Gaussian Kernel Generation (MB-GKG) block that synthesizes kernels from a fixed dictionary to stabilize optimization, and a Multi-Period Chronologically Continuous Rasterization (MP-CCR) block that enforces strict temporal continuity across periodic boundaries. Comprehensive experiments on standard benchmark datasets demonstrate that TimeGS attains state-of-the-art or competitive performance. The code is at https://github.com/yixinwang1/TimeGS.
arXiv:2603.02536v2 Announce Type: replace
Abstract: Satellite-terrestrial communications are severely constrained by high path loss, limited spectrum resources, and time-varying channel conditions, rendering conventional bit-level transmission schemes inefficient and fragile, particularly in low signal-to-noise ratio (SNR) regimes. Semantic communication has emerged as a promising paradigm to address these challenges by prioritizing task-relevant information over exact bit recovery. In this paper, we propose a semantic forwarding-based semantic communication (SFSC) framework optimized for satellite-terrestrial networks. Specifically, we develop a vector-quantized joint semantic coding and modulation scheme, in which the semantic encoder and semantic codebook are jointly optimized to shape the constellation symbol distribution, improving channel adaptability and semantic compression efficiency. To mitigate noise accumulation and reduce on-board computational burden, we introduce a satellite semantic forwarding mechanism, enabling relay satellites to forward signals directly at the semantic level without full decoding and re-encoding. Furthermore, we design a channel-aware semantic reconstruction scheme based on feature-wise linear modulation (FiLM) to fuse the received SNR with semantic features, enhancing robustness under dynamic channel conditions. To support multi-user access, we further propose a codebook split-enhanced model division multiple access (CS-MDMA) method to improve spectral efficiency. Simulation results show that the proposed SFSC framework achieves a peak signal-to-noise ratio (PSNR) gain of approximately 7.9 dB over existing benchmarks in the low-SNR regime, demonstrating its effectiveness for robust and spectrum-efficient semantic transmission in satellite-terrestrial networks.
arXiv:2605.31051v2 Announce Type: replace
Abstract: The Linear Ordering Problem (LOP) is a fundamental combinatorial optimization problem with important applications in areas such as economics, social choice, and machine learning. Its most prominent use is the triangulation of economic input-output tables, which helps identify critical industries in an economy. Most existing algorithms have been evaluated on benchmarks derived from outdated macroeconomic data, which no longer reflect the structure of contemporary economies. Furthermore, LOP instances often exhibit many distinct global optima that can differ substantially from one another, creating challenges for applications that rely on a single solution. To address these limitations, we introduce a novel benchmark suite derived from up-to-date real-world economic data and an algorithmic scheme that leverages state-of-the-art LOP metaheuristics to generate diverse sets of high-quality solutions, together with metrics for assessing both quality and diversity. Experiments were conducted to report results on the proposed benchmark suite under both the traditional single-solution setting and the newly introduced multi-solution scenario
arXiv:2605.31071v2 Announce Type: replace
Abstract: TREE CONTAINMENT is a central decision problem in mathematical phylogenetics, asking whether a given rooted phylogenetic tree is embeddable in ("displayed by") a given rooted phylogenetic network. While the problem is NP-complete for general networks, many algorithmic advances have relied on structural parameters that capture how "tree-like" a network is. In this paper we investigate TREE CONTAINMENT under the structural parameter scanwidth, a directed width measure generalizing popular parameters measuring tree-likeness of phylogenetic networks. We first present a parameterized algorithm that solves the problem in $O(4^{k + k\log{k}} n + nm^2)$ time, where $n$ and $m$ are the numbers of nodes and arcs in the network and $k$ is the width of a given tree-extension. Complementing this upper bound, we prove a matching lower bound under the Exponential-Time Hypothesis (ETH), showing that there is no algorithm for TREE CONTAINMENT that runs in $2^{o(c\log{c})} n^{O(1)}$ time, even on binary inputs, where $c$ is the directed cutwidth of the input network, which upper-bounds the scanwidth $k$.
arXiv:2606.07004v1 Announce Type: new
Abstract: Estimating the logarithm of the determinant of large sparse positive definite symmetric matrices is an important task in numerical linear algebra, machine learning, Gaussian processes, and uncertainty quantification. In this work, we introduce two scalable and efficient methods for large-scale log-determinant termed the Optimal Stochastic Arnoldi with Incomplete Orthogonalization Procedure (OSA-IOP) and the Optimal Stochastic Lanczos Quadrature (OSLQ). The OSA-IOP approach extends the Incomplete Orthogonalization Procedure (IOP), originally developed for matrix exponential functions for exponential time stepping integrators, to compute the action of the matrix algorithm on a vector. We observe that combining IOP with a randomized Hutch++ algorithm, the OSA-IOP significantly reduces computational cost while maintaining high accuracy. The OSLQ method estimates log-determinants by coupling Lanczos quadrature with Hutch++ and controlled orthogonalization, leveraging Krylov subspaces as efficient quadrature mechanisms to approximate quadratic forms involving the matrix logarithm. We derive error bounds for both methods. Extensive numerical experiments on large-scale sparse matrices from real-world applications demonstrate the accuracy, robustness, and scalability of the proposed approaches.
arXiv:2606.07044v1 Announce Type: new
Abstract: Accurate and coherent passenger demand forecasting is essential for Urban Rail Transit (URT) operations. Passenger demand has a hierarchical structure in which origin-destination (OD) flows aggregate to station-level inflows and outflows through conservation constraints. In practice, station-level and OD-level forecasts are often generated independently, producing incoherent predictions that violate these constraints and introduce inconsistencies into operational decision-making. Such issues become more severe during disruptions, when forecasting reliability is most critical. This paper presents the first hierarchical forecast reconciliation framework for joint station-level and OD-level URT demand prediction. A neural Fully Connected Reconciler (FCR) learns a non-linear mapping from incoherent base forecasts to coherent hierarchical predictions while guaranteeing exact structural consistency by construction. The method is benchmarked against OLS, WLS, and Minimum Trace (MinT) variants using Rejsekort smart-card data from the Copenhagen S-train network under one-step, multi-step, and disruption forecasting scenarios. Results show that reconciliation consistently improves OD forecasting accuracy while ensuring hierarchical coherence. Under normal conditions, FCR performs competitively with MinT-based methods. An oracle analysis indicates that perfect station-level forecasts could reduce OD prediction error by up to 34 percent, highlighting the value of improved base forecasts. Under severe disruptions, FCR outperforms classical methods, reducing OD forecasting error by up to 17.45 percent in multi-step destination-side delay scenarios. These findings establish hierarchical reconciliation as an effective mechanism for improving forecast robustness, with the largest benefits occurring under the most challenging operating conditions.
arXiv:2606.07079v1 Announce Type: new
Abstract: Standard diffusion models corrupt an entire sample with a single shared noise level, forcing all spatial regions to follow the same denoising trajectory. We introduce AsyncPatch Diffusion, a joint-diffusion framework that assigns distinct noise levels to different input dimensions, such as image pixels, or latent tokens. We show how this asynchronous corruption defines a valid generative process while supporting a richer family of spatially heterogeneous denoising trajectories, and prove the first valid ELBO for this process. We show that a single pretrained model can perform spatially adaptive generation, where different regions are denoised on different schedules. A key challenge is training: naive independent noise-level sampling overemphasizes highly heterogeneous configurations and underrepresents homogeneous noise levels, that are crucial during sampling. We address this with a controlled noise-level sampler that regulates both the average corruption level and its spatial variability. AsyncPatch achieves generation quality comparable to conventional diffusion on ImageNet 256 and LSUN, while being natively suited for inpainting without task-specific fine-tuning. We further introduce input guidance, which uses clean or partially corrupted regions to guide the generation of unknown regions, improving local consistency and texture matching. Finally, we demonstrate adaptive generation strategies including uncertainty-guided acceleration and autoregressive sampling.
arXiv:2606.07080v1 Announce Type: new
Abstract: We present dots.tts, a 2B-parameter continuous autoregressive text-to-speech (TTS) foundation model that models speech in a continuous latent space. Compared with existing continuous autoregressive models, our key innovations are threefold. First, we train an AudioVAE with multiple objectives to build a semantically structured and prediction-friendly continuous speech space. Second, we use full-history conditioning in the flow-matching head to preserve long-range consistency and reduce drift during generation. Third, we apply reward-free self-corrective post-training to the flow-matching head to further improve robustness and acoustic quality. After being trained on a large-scale multilingual corpus, dots.tts achieves the best average performance on Seed-TTS-Eval, with WERs of 0.94%/1.30%/6.60% and SIM scores of 81.0/77.1/79.5 on the zh/en/zh-hard test sets, respectively. Across other benchmarks, dots.tts also consistently demonstrates open-source state-of-the-art performance, exhibiting strong generation stability, voice cloning ability, and emotional expressiveness. For efficient inference, we further apply CFG-aware MeanFlow distillation, enabling low-latency speech generation with first-packet latencies of 85/54 ms in output streaming and dual-streaming modes, respectively. To facilitate reproducible research and practical deployment, we release the training and inference code, together with the pretrained, post-trained, and MeanFlow-distilled checkpoints, under the Apache 2.0 license.
arXiv:2606.07085v1 Announce Type: new
Abstract: As an emerging operating system, HarmonyOS has a significant demand for software migration from platforms such as Android and iOS, where the User Interface (UI) translation accounts for a critical link. However, the latest UI development has shifted to declarative paradigms, e.g., Kotlin Jetpack Compose (KJC) for Android, SwiftUI for iOS, and ArkUI for HarmonyOS, rendering prior translation approaches inapplicable, as they target either backend logic or legacy imperative UIs.
As such, this paper targets ArkUI and proposes an automatic translation approach, namely ArkTrans, to port UI files from Android and iOS to HarmonyOS.
ArkTrans overcomes two salient challenges during the translation: (1) Programming Language (PL) unfamiliarity, and (2) severe syntactic chaos. Towards the first challenge, ArkTrans heuristically constructs ArkUI skeletons by extracting metadata from source PL, thereby guiding LLMs' initial translation. As for the second challenge, ArkTrans executes empirically revealed post-fixing rules via pattern matching to repair most of the remaining syntactic errors. To examine the effectiveness of ArkTrans, we construct a 100-sample parallel UI page translation benchmark from KJC/SwiftUI to ArkUI at the file level. Extensive experiments demonstrate that LLMs with direct/one-shot prompting cannot translate a single compilable UI page. In contrast, at most 90.67\% ArkTrans-translated files can be successfully compiled with high visual fidelity.
Multi-Agent Reasoning with Consistency Verification Improves Uncertainty Calibration in Medical MCQA
arXiv:2603.24481v2 Announce Type: replace
Abstract: Miscalibrated confidence scores are a practical obstacle to deploying AI in clinical settings. A model that is always overconfident offers no useful signal for deferral. We present a multi-agent framework that combines domain-specific specialist agents with Two-Phase Verification (Wu et al., 2024) and S-Score Weighted Fusion to improve both calibration and discrimination in medical multiple-choice question answering. Four specialist agents (respiratory, cardiology, neurology, gastroenterology) generate independent diagnoses using Qwen2.5-7B-Instruct. Each diagnosis undergoes a two-phase self-verification process that measures internal consistency and produces a Specialist Confidence Score (S-score). The S-scores drive a weighted fusion strategy that selects the final answer and calibrates the reported confidence. We evaluate on high-disagreement subsets of MedQA-USMLE and MedMCQA (100 and 250 questions). All results are specific to this filtered regime. On MedQA-250, the full system achieves ECE = 0.091 (74.4% reduction over the single-specialist baseline) and AUROC = 0.630 (+0.056) at 59.2% accuracy. Calibration gains of 49-74% hold across all four settings. Ablation analysis reveals that Two-Phase Verification drives ECE reduction while multi-agent reasoning drives AUROC improvement, suggesting that consistency checking and ensemble aggregation address different failure modes of LLM uncertainty. Whether the resulting confidence signal is sufficient to support clinical deferral decisions in practice remains a direction for future investigation.
arXiv:2604.03779v2 Announce Type: replace
Abstract: Diffusion models have excelled at generative tasks for both continuous and token-based domains, but their application to discrete ordinal data remains underdeveloped. We present CountsDiff, a diffusion framework designed to model distributions on the natural numbers. CountsDiff extends the Blackout diffusion framework by simplifying its formulation through a direct parameterization in terms of a survival probability schedule and an explicit loss weighting. This introduces flexibility through design parameters with direct analogues in existing diffusion modeling frameworks. Beyond this reparameterization, CountsDiff introduces features from modern diffusion models, previously absent in counts-based domains, including continuous-time training, classifier-free guidance, and churn/remasking reverse dynamics that allow non-monotone reverse trajectories. We propose an initial instantiation of CountsDiff and validate it on natural image datasets (CIFAR-10, CelebA), exploring the effects of the introduced design parameters in a complex, well-studied, and interpretable data domain. We then highlight biological count assays as a natural use case, evaluating CountsDiff on single-cell RNA-seq imputation in fetal and heart cell atlases. Remarkably, we find that even this simple instantiation matches or surpasses the performance of a state-of-the-art discrete generative model and leading scRNA-seq imputation methods, while leaving substantial headroom for further gains through optimized design choices in future work.
arXiv:2604.04226v2 Announce Type: replace
Abstract: The Agentic Web is emerging as a paradigm in which autonomous software agents interact with online resources and with each other to accomplish user goals. However, the capacity of Agentic Web is still limited by insufficient autonomous software agent population, which has become a crucial challenge for scaling Agentic Web. In order to alleviate this, we study the task of automatically converting existing code repositories into autonomous software agents via coding agents, decompose the process into critical stages, and identify key technical hurdles. To systematically evaluate this capability, we propose SoftWare Agent generation for Agentic Web Bench (SW-$A^2$-Bench), the first benchmark designed for software agent generation. SW-$A^2$-Bench evaluates not only whether software agents can be generated, but also whether generated software agents are faithful to the source repositories and interoperable with other agents in multi-agent workflows. Our experiments demonstrate that our approach effectively activates the functional capabilities of code repositories and enables interoperable multi-agent collaboration in Agentic Web. We believe that this work will provide a standardized evaluation for software agent generation and will contribute to the future of scaling the capacity of Agentic Web.
arXiv:2606.07120v1 Announce Type: new
Abstract: Autoencoders (AEs) learn low-dimensional representations by mapping data into a latent space while minimizing reconstruction error. Despite their empirical success, theoretical understanding remains limited and largely restricted to linear models or settings without a bottleneck. In this work, we study nonlinear AEs with a fixed finite-dimensional bottleneck in the mean-field (MF) regime. We derive explicit MF learning dynamics for both encoder and decoder, providing a tractable characterization of training in the nonlinear setting. We show that, over finite time horizons, the empirical risk of finite-width networks trained with stochastic gradient descent closely tracks the MF risk trajectory with high probability. At optimality, we further establish that the finite-width risk converges to the MF optimum, demonstrating that finite networks are sufficiently expressive to approximate the infinite-width solution.
arXiv:2606.07145v1 Announce Type: new
Abstract: Text-guided diffusion models have become effective tools for real-image visual editing, where the edited image must follow a target instruction while preserving editing-irrelevant structure. Most training-free editors rely on inversion: a source image is mapped to a noisy latent trajectory and the terminal latent is reused for target-prompt denoising. This reuse is useful for preservation, but it also couples source reconstruction and target editing. The resulting trajectory mismatch may either damage background/layout details or over-constrain the intended edit. This paper presents Consistent-Inversion, a training-free reverse consistency guidance framework for structure-preserving visual editing. Instead of treating the inverted source latent as a fixed initialization, Consistent-Inversion checks whether an intermediate target trajectory can be reversed toward the source inversion trajectory under the source prompt. To make this check well-defined, we construct an auxiliary target-side noise representation, perform source-guided reverse denoising, and use the resulting reverse consistency discrepancy as a correction signal for selected early target denoising steps. The method does not update model parameters, is compatible with inversion-based editors, and introduces only a small inference overhead when applied sparsely. Experiments on PIE-Bench show that Consistent-Inversion improves background and structural fidelity under a unified SD3.5 protocol while maintaining target-prompt alignment, and compatibility experiments further verify the same correction principle on classical Stable-Diffusion inversion pipelines.
arXiv:2606.07159v1 Announce Type: new
Abstract: Compute Express Link (CXL) enables memory pooling over disaggregated memory, offering the potential to improve resource utilization in persistent memory systems. However, integrating persistence semantics into CXL-based memory pooling introduces substantial latency, which limits system scalability. This overhead arises because persist operations must traverse the entire CXL fabric, including switches, links, and protocol layers, before reaching remote persistent memory. To this end, we argue that extending CXL switches with persistence support is a promising direction for improving the scalability of persistent memory pooling. However, moving persistence support into the network breaks the traditional correctness assumptions of centralized persistence domains. In particular, enabling persistence within distributed structures, such as CXL switches, can introduce stale reads and writes if not carefully coordinated.
In this paper, we propose Distributed Persistence Domain (DPD), a new abstraction for persistent memory pooling that enables persistence support at the CXL switch level. We first formalize the concept of a distributed persistence domain and use DPD as a framework to identify the correctness hazards that arise when persistence structures are distributed across the CXL fabric. Based on this analysis, we derive the design requirements needed to guarantee correctness. Building on these insights, we present Persistent CXL Switch, a CXL switch architecture that incorporates persistence support to significantly reduce persist latency, enable read forwarding, and coalesce writes, while preserving correctness and crash consistency. We evaluated our system design using both SPLASH-4 and YCSB benchmarks. Simulation results show an average speedup of 33% over volatile CXL switches, and up to 36% speedup with read forwarding optimization across all workloads.
arXiv:2606.07171v1 Announce Type: new
Abstract: Multimodal Large Language Models (MLLMs) enable flexible instruction-driven image editing, but privacy risks arise when user images expose diverse and user-specific private content. Canonical privacy protection strategies typically substitute sensitive regions with surrogate content before cloud editing. Yet, the resulting output is often an edited surrogate rather than the desired edited source image, neglecting the local recovery in both design and evaluation scope. To this end, we introduce SPPE (Surrogate-based Privacy-Preserving Editing), the first recovery-oriented benchmark covering 36 fine-grained privacy categories and 65 editing instructions. It defines two complementary tasks: 1) editability assessment, which estimates before cloud interaction whether a surrogate can induce an edit consistent with the original image; and 2) surrogate-to-source edit recovery, which evaluates whether the edited surrogate can be transferred back to the private source with the edit effect preserved. We address each task with a dedicated method: ERMA predicts surrogate editability through instruction-aware multimodal relation modeling, while \method performs cycle-consistent recovery by using the surrogate editing pair as visual edit evidence and the source image as a source-preserving anchor. Experiments on SPPE and InstructPix2Pix show consistent improvements on both tasks. For editability assessment, ERMA improves over the best-performing baselines by 13.9% in SRCC and 12.3% in PLCC. For surrogate-to-source edit recovery, C2E-S2SER outperforms SOER across all 8 source integrity and edit consistency metrics on SPPE.
arXiv:2606.07179v1 Announce Type: new
Abstract: Streaming 3D Gaussian Splatting requires highly scalable, progressive representations. Existing progressive methods rely on \textit{discrete layering}, accumulating separate splat sets for each level of detail. This structural independence between layers inherently leads to error accumulation, severe splat redundancy, and uncontrolled quality transitions. We propose EvoGS, the first \textit{continuous-layering} representation. Organized as an Evolution Tree, EvoGS generates finer details via an explicit, wavelet-inspired parent-child refinement. This empowers child nodes to structurally correct ancestral errors, yield inherently sparse and highly compressible inter-layer signals. Extensive experiments show EvoGS eliminates splat redundancy from over 65\% to under 25\%. Compared to state-of-the-art baselines, it reduces transmission payload and GPU VRAM footprint by up to 2.4$\times$ and 5.5$\times$, respectively, and achieves smooth quality transitions optimal for real-time adaptive streaming. Project page: https://yuang-ian.github.io/evogs/