arXiv:2606.08332v1 Announce Type: new Abstract: Self-supervised learning (SSL) has achieved remarkable representation learning performance, but many existing methods rely on large batch sizes, memory banks, momentum encoders, or global synchronization mechanisms that substantially increase computational cost and training complexity. In this work, we propose Semantic Mutual Information (SMI), a lightweight self-supervised objective derived from a mutual-information-inspired dependency formulation under Gaussian assumptions. Unlike conventional correlation matching objectives that operate on high-dimensional feature correlation matrices, SMI performs optimization on a sample-level dependency matrix through a nonlinear transformation of pairwise correlations. This formulation induces distinct optimization dynamics that emphasize strongly dependent semantic pairs while maintaining representation diversity. Experimental results on ImageNet using a ResNet-50 backbone demonstrate that SMI achieves competitive linear evaluation performance relative to state-of-the-art SSL approaches while substantially reducing computational complexity. Across multiple low-resource benchmarks, SMI consistently improves transfer performance over Barlow Twins, particularly on fine-grained datasets. Furthermore, analyses of optimization dynamics and representation geometry suggest improved alignment--redundancy balance, greater feature diversity, and more spatially localized semantic representations. These results indicate that nonlinear dependency optimization provides an effective and computationally efficient alternative to conventional correlation-based self-supervised learning objectives.
Science Journals
arXiv:2606.09099v1 Announce Type: new Abstract: Autonomous UAV swarms require scalable coordination mechanisms that maintain mission performance under limited communication, environmental uncertainty, and component failures. Centralized approaches provide global coordination but suffer from communication bottlenecks and single-node vulnerabilities, whereas fully decentralized methods often lack mission-level consistency. This paper presents Layered Autonomous Edge Intelligence (LAEI), a UAV-swarm framework that combines onboard learned policies with lightweight mission-level supervision. Each UAV performs local perception, obstacle avoidance, and action selection onboard, while the supervisory layer provides adaptive goal reassignment, fault-aware recovery, and context-dependent policy guidance without directly controlling low-level actions. LAEI further incorporates recovery strategies, including dynamic reassociation, backup supervisory support, and fallback local autonomy, to maintain mission continuity under representative failure scenarios. We evaluate LAEI in simulated UAV-swarm scenarios using mission completion time, collision rate, and coverage efficiency. The results show that LAEI reduces mission completion time and improves operational efficiency while maintaining collision-aware distributed UAV-level decision-making.
arXiv:2606.09100v1 Announce Type: new Abstract: Community detection is a fundamental problem in the analysis of complex networks. It has applications across social, biological, and financial domains. Traditional algorithms such as Louvain, LPA, and modularity optimization often require manual parameter tuning. They also suffer from inaccurate cluster center selection and struggle with scalability. To address these challenges, we propose Automatic Laplacian Centrality Means (ALCMeans), a novel community detection algorithm. ALCMeans combines Laplacian energy-based automatic center identification with DeepWalk embeddings for robust node representation. Unlike existing Laplacian-based and clustering methods, ALCMeans eliminates the need to predefine the number of communities, enhances cluster center selection using structural importance, and leverages representation learning for more accurate and stable assignments. Experimental results on benchmark datasets demonstrate 10 to 20 percent higher NMI and ARI scores compared to Louvain, Newman-Girvan, LPA, Fast-Greedy, and a recent GNN-based competitor (MAGI, KDD 2024). Additional evaluations with modularity and F1-scores confirm the superiority of ALCMeans. Ablation studies highlight the critical contributions of each component. Despite its reliance on DeepWalk parameters and increased runtime relative to lightweight heuristics, ALCMeans consistently outperforms state-of-the-art methods. This makes it a promising tool for real-world network analysis.
arXiv:2606.08345v1 Announce Type: new Abstract: Large language models (LLMs) have shown strong performance in financial analysis and surface-level factual error detection, yet their ability to identify fraudulent financial misinformation in audited corporate reporting remains underexplored. Existing financial and audit benchmarks mainly focus on factual verification, numerical reasoning, rule compliance, or audit workflows, but rarely evaluate misleading disclosure narratives or management explanations that obscure the true drivers of reported performance. We introduce AuditFraudBench, an enforcement-grounded benchmark constructed from authentic company filings and regulatory materials, including original and restated 10-K and 10-Q filings, structured financial statements, MD&A disclosures, and SEC Accounting and Auditing Enforcement Releases (AAERs). AuditFraudBench contains three tasks: Profit Source Attribution, Misleading Narrative Detection, and Fraud Pattern Classification, which evaluate whether models can identify the true source of reported performance, detect misleading disclosure framing, and classify misconduct mechanisms into known manipulation patterns. We evaluate GPT, DeepSeek, and Qwen series LLMs on the benchmark. Results show that both proprietary and open models still struggle to jointly reason over financial figures, disclosure framing, restatement evidence, and enforcement-grounded fraud mechanisms. AuditFraudBench provides a challenging testbed for audit-relevant, evidence-grounded evaluation of LLMs in financial reporting.
arXiv:2606.08346v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has become a dominant paradigm for improving the reasoning capabilities of large language models (LLMs). Recent tree-based methods such as TreeRPO extend flat trajectory sampling with tree-structured rollouts to obtain dense, step-level reward signals without a separate process reward model. However, not all trees are equally informative: trees where all leaves succeed, all leaves fail, or the policy already predicts the reward distribution contribute little to gradient updates, wasting compute. We introduce CATPO (Critique-Augmented Tree Policy Optimization), which diagnoses and addresses this waste at the tree level. CATPO first scores each tree via a tree informativeness score, F(T), combining leaf-outcome diversity with policy-reward decorrelation at zero extra compute. For dead-wrong trees where all branches fail, CATPO applies critique-guided healing: it locates the shallowest failure point, generates a natural-language critique, and grafts refined continuations to recover training signal. Finally, an informativeness-weighted loss scales each tree's gradient contribution by its normalized score, concentrating parameter updates on the most informative trees while preserving overall gradient magnitude. Experiments on Qwen2.5-Math-1.5B trained with the MATH dataset show that CATPO achieves 37.5% macro accuracy across four benchmarks (AIME24, MATH-500, OlympiadBench, and MinervaMath), improving over TreeRPO by 1.9% and GRPO by 4.8%.
arXiv:2606.09104v1 Announce Type: new Abstract: Deep reinforcement learning (DRL) frameworks for portfolio optimization have shown promise for their ability to learn allocation rules dynamically from market data. However, these models fail to account for fat-tailed returns, which characterize actual market behavior with more frequent extreme events. Furthermore, historical data is treated homogeneously, without accounting for temporal importance, leading models to fail during regime changes. We propose a new BAVAR-BLED algorithm that combines methods derived from Bayesian-Averaging Vector Autoregressive (BAVAR) and the Black-Litterman model using Elliptical Distributions (BLED) within a TD3 architecture. BAVAR captures a set of vector autoregressive representations that consider multi-scale temporal features, enabling adaptive allocation decisions based on regime-aware estimates of return expectations and dispersion matrices. These estimates serve as prior inputs to BLED, a model that uses Student's t-distributions, allowing for more realistic fat tail return estimates. The BAVAR-BLED algorithm uses transformer networks for view construction and CNNs for risk-aversion estimates, which modify dynamic allocation decisions based on market conditions. An evaluation of 29 Dow Jones Industrial Average constituents over a decade-long market period shows that BAVAR-BLED significantly outperforms state-of-the-art methods, achieving Sharpe and Sortino ratios of 1.72 and 2.70, respectively, and total returns of 57.26%.
arXiv:2606.09111v1 Announce Type: new Abstract: Unmanned Aerial Vehicle (UAV) multispectral point clouds (MPC) provide high-dimensional spatial-spectral data for sub-canopy target detection; however, their efficacy is significantly compromised by severe illumination heterogeneity caused by vegetation shadows. To address this, we propose a prior-free anomaly detection framework capable of robustly handling lighting variations. First, we formulate solar angle estimation as an inverse optimization problem. By coupling spectral indices with a ray-tracing model, this strategy achieves Prior-Free Shadow Extraction without relying on flight metadata, effectively distinguishing dark objects from true shadows. Second, to mitigate spectral distortions, we introduce an Illumination-Consistent Sparse Representation mechanism. Unlike standard reconstruction methods, we construct a background dictionary strictly from neighbors sharing the same illumination state. This constraint effectively disentangles spectral reflectance from lighting variations, ensuring that targets are represented solely by physically consistent background points. Experimental results indicate that the proposed method significantly improves the separability between anomalies and background in complex forest environments, demonstrating superior performance over state-of-the-art baselines. This framework is particularly suited for identifying camouflaged military targets, mapping fallen tree trunks, and uncovering archaeological ruins hidden beneath dense foliage.
arXiv:2605.28912v2 Announce Type: replace Abstract: The rapid growth of AI-driven data centers and large-scale energy storage systems is increasing the reliance of power system operation on real-time measurement data and automated decision-making. However, many existing detection methods rely on statistical or data-driven analysis of measurements and can fail when attackers exploit the same data structure to craft stealthy perturbations. To illustrate this limitation, we demonstrate a blind False Data Injection Attack (FDIA) in which an Autoencoder learns the measurement manifold and generates perturbations aligned with the Jacobian null space, thereby allowing the attack to evade both residual-based baddata detectors and time-series anomaly detectors. To mitigate data-driven FDIAs which exploit the null space, we propose a topology-informed Cycle-Space Detector (CSD) that leverages the Cycle-Space of the network to impose structural constraints that enhance null space estimation. In addition, we prove that by using the Minimum Cycle Basis (MCB), the proposed CSD achieves the optimal generalization error for attack detection. By exploiting topology-derived cycle constraints rather than relying solely on numerical null space estimation, the proposed method does not require precise line parameters and improves the separation between normal and attacked measurements. Simulation results on IEEE 14-, 30-, 57-, and 118-bus systems demonstrate that the proposed method effectively detects data-driven FDIAs under realistic measurement noise.
arXiv:2605.29475v2 Announce Type: replace Abstract: Large language models (LLMs) show remarkable potential in scientific hypothesis discovery. However, existing approaches face two critical limitations: they treat divergent exploratory search and convergent fine-grained refinement as isolated tasks, and they operate autonomously with little to no human guidance. We present MOOSE-Copilot, the first unified framework to bridge this abstraction gap through a formalized human-AI interaction (HAII) protocol. Our system empowers scientists to steer the generative process via three explicit signals: initial blueprints, inter-stage routing, and intra-stage feedback. Using an oracle-simulated evaluation in which an LLM provides idealized expert signals, we show that injecting these structured signals significantly outperforms purely autonomous baselines, characterizing the gains achievable under high-quality guidance. Furthermore, we build a web-based interface that turns the framework into a no-code workflow: researchers pose a question, watch the hypothesis search unfold as an interactive tree, and steer it by selecting hypotheses, routing between stages, and injecting feedback-no command-line agents required. This makes end-to-end hypothesis discovery directly accessible to interdisciplinary researchers.
arXiv:2605.29823v2 Announce Type: replace Abstract: Deep networks often exhibit a preference for "simple" solutions, and such a simplicity bias is widely believed to play a key role in generalization. Yet a broadly applicable, quantitative measure of simplicity remains elusive. We introduce polynomial representations as a distribution-aware, low-dimensional surrogate for neural functions: we approximate a network's predictive behavior along data-dependent interpolation paths using orthogonal polynomial bases, yielding a compact functional representation. We show that the effective degree of this representation serves as a practical simplicity metric that is predictive of generalization across tasks and architectures, and consistently outperforms existing generalization proxies such as sharpness. Finally, polynomial representations naturally yield a differentiable simplicity regularizer, which consistently improves generalization in image and text classification, fine-tuning contrastive vision-language models, and reinforcement learning.
arXiv:2605.30123v2 Announce Type: replace Abstract: Homomorphic encryption (HE) enables privacy-preserving aggregation in federated learning (FL) by allowing the server to operate on encrypted data without decryption. Existing HE-over-the-air (OTA) methods mainly rely on single-key HE schemes and require channel estimation or pre-equalization to compensate for wireless fading. However, single-key HE remains vulnerable to honest-but-curious (HBC) clients holding the shared secret key, while multi-key HE provides stronger client-level security by assigning each device its own secret key. We propose a four-phase protocol that enables the aggregation of xMK-CKKS over a shared wireless channel without channel estimation. The protocol retransmits partial public keys and ciphertexts through the same channel realization, so that the dominant large-modulus encryption terms cancel algebraically during decryption. We integrate this protocol with zero-order FL over slowly varying LoS-dominant channels, where each device transmits a single encrypted scalar per round and the communication/encryption overhead is independent of the model dimension. We show that the residual noise induced by encryption and wireless aggregation preserves the standard convergence rate \(O(1/\sqrt{K})\) up to a negligible noise floor, where $K$ is the number of communication rounds. The protocol assumes an non-trusted server and is secure against HBC clients, preventing any client from recovering the local updates of other participants. Numerical results on MNIST validate the theoretical analysis.
arXiv:2605.30226v2 Announce Type: replace Abstract: Vision-Language-Action (VLA) models have emerged as a promising paradigm for grounding visual-language understanding into real-world robotic manipulation. However, dexterous manipulation remains challenging for VLA policies due to high-dimensional hand control and compounding execution errors, which makes real-world RL post-training essential for bridging the gap between visually grounded action generation and physically reliable dexterous execution. However, high-dimensional dexterous exploration often triggers temporal inconsistency, sample inefficiency and hardware risks in the real world. To address these challenges, we propose BORA, an offline-to-online RL post-training framework designed for real-world dexterous VLA models. In the offline phase, BORA constructs a critic that takes both the VLM's cognition tokens and action chunks as inputs. This design enables action-conditioned value guidance, allowing the critic to evaluate dexterous hand motions beyond visual context alone. During the subsequent online phase, BORA freezes the VLA base and introduces a lightweight, Human-in-the-Loop (HiL) chunk-wise residual adaptation mechanism to mitigate real-world execution errors and further correct the offline-learned intents within the actual physical environment. By inheriting the offline critic and employing intervention-driven rewards, BORA effectively corrects execution discrepancies and adapts to real-world physical variances while preserving the pretrained policy as a stable prior. Extensive evaluations across five complex real-world dexterous tasks demonstrate that BORA significantly outperforms pure imitation learning and traditional decoupled RL baselines, achieving a 33% absolute increase in average success rate under standard settings and up to a 43% improvement in unseen object generalization.
arXiv:2605.31014v2 Announce Type: replace Abstract: Multi-omics data provide complementary molecular characterizations of disease phenotypes and play an important role in disease diagnosis and subtype classification in precision medicine. However, acquiring complete multi-omics profiles is expensive and time-consuming, while most existing deep learning methods assume full modality availability during inference, resulting in substantial redundancy and limited practicality in clinical settings. To address this issue, we propose SDM-Q, a reinforcement learning framework for adaptive and cost-aware multi-omics classification. Specifically, multi-omics diagnosis is reformulated as a finite-horizon sequential decision problem, where the currently acquired omics modalities define the diagnostic state at each stage. An action--value function determines whether to acquire an additional modality or terminate the decision process and output the final prediction. To balance diagnostic utility and acquisition cost, the reward is defined only at the terminal stage and jointly determined by classification correctness and cumulative modality acquisition cost. A backward stage-wise optimization strategy is introduced to improve policy consistency and training stability. Experiments on four public multi-omics datasets, including ROSMAP, LGG, BRCA, and KIPAN, demonstrate that SDM-Q effectively reduces redundant modality acquisition while maintaining competitive classification performance compared with methods using complete multi-omics inputs. In the BRCA and KIPAN datasets, more than 99\% and 95\% of subjects, respectively, achieve accurate classification using only a single omics modality, while the average number of acquired modalities remains below two for ROSMAP and LGG. These results suggest that cost-aware sequential decision-making provides an effective paradigm for improving the efficiency of precision medicine workflows.
arXiv:2605.31498v3 Announce Type: replace Abstract: A long standing challenge in computational chemistry and biophysics is efficiently sampling the Boltzmann distribution of molecules. Advances in generative modeling have been proposed to address the limitations of conventional sampling techniques by eliminating the computational cost of simulation. A promising direction is iteratively finetuning diffusion models along a temperature ladder whereby training data is generated via importance sampling during inference-time annealing. Unfortunately, these methods require computing a divergence over the score field to estimate importance weights, rendering them intractable for larger systems. Here we present scalable inference-time annealing (SITA), which retrains flow-based models to generate samples at progressively lower temperatures using an energy-based model to facilitate fast surrogate likelihoods. We demonstrate state-of-the-art performance on both Alanine Dipeptide and Alanine Tripeptide while avoiding costly divergence terms. Our code is available at https://github.com/countrsignal/sita.git
arXiv:2606.00229v2 Announce Type: replace Abstract: Natural language is a powerful reasoning medium for language and vision-language models, but it is mismatched to the granularity of continuous control. Text and explicit subgoals operate at task-level granularity, whereas vision-language-action (VLA) policies must choose actions at a much finer temporal scale; a single reasoning step can therefore span many action chunks while remaining only weakly coupled to the action needed now. This suggests a different question for VLA: what should play the role of language? We argue that a useful VLA reasoning medium must be shareable across model instances, verifiable through downstream action improvement, and aligned with temporally extended control structure. Based on this view, we propose Continuous Reasoning for Vision-Language-Action. Our model first predicts continuous reasoning in the form of a structured set of continuous thoughts, then reuses them as shared context for chunk-structured action generation. Better action prediction alone does not certify good reasoning: if the same internal medium cannot be shared across model instances and independently verified through improved downstream control, the added latent may simply become a model-private shortcut that helps on seen behaviors without supporting generalizable control. We therefore instantiate continuous reasoning as a shared Gaussian latent interface and train it with a self-verification objective in which an exponential-moving-average teacher must successfully consume the student's reasoning when predicting target actions. Empirically, Continuous Reasoning improves LIBERO-PRO robustness and performs strongly on real robots, raising mean subtask success over {\pi}0.5 by 40.4% on TX-G2, an AgiBot G2-compatible variant, and 26.3% on HSR. This suggests that reasoning in VLA is less about extra tokens than about a shareable, verifiable internal language for action.
arXiv:2606.01342v3 Announce Type: replace Abstract: Learning-augmented paging has been extensively studied in recent years. A key advantage over naive ML-based approaches is \emph{bounded robustness}, which guarantees worst-case performance even when predictions are inaccurate, making these algorithms valuable for real-world systems. Prior work achieves robustness bounds of $2H_k + O(1)$ in the randomized setting, leaving a gap to the optimal competitive ratio $H_k$. In this paper, we study how to close this gap. We begin by reviewing online optimality and proving a new property of the latest $H_k$-competitive algorithm, which facilitates our analysis in the learning-augmented setting. Then, we review existing learning-augmented paging algorithms and introduce a unifying primitive, the \emph{relative prediction budget}, which captures the essence of establishing robustness and reveals that prior algorithms either overuse or underutilize predictions. Guided by the above analysis, we develop a new framework that achieves the best-possible robustness up to an additive constant for learning-augmented paging: $H_k + O(1)$. Experiments further demonstrate strong practical performance.
arXiv:2606.01379v2 Announce Type: replace Abstract: While prompt-based parameter-efficient continual learning mitigates catastrophic forgetting by isolating task-specific prompts, this isolation also limits later tasks from improving earlier ones, leaving backward knowledge transfer underexplored. We address this limitation by proposing Selective bAckward refinement for positive Backward knowledge transfER (SABER), a replay-free framework that enables controlled backward transfer in prompt-based continual learning. SABER determines when backward refinement is beneficial using complementary task-correlation criteria based on prompt-gradient geometry and loss-distribution similarity, and how to perform refinement safely by restricting updates to non-interfering directions in the prompt parameter space. Extensive experiments across multiple continual learning benchmarks and diverse pretrained backbones, including T5-Large, LLaMA, and Qwen, demonstrate that SABER consistently achieves positive backward transfer while maintaining strong overall average performance. Code is available at https://github.com/OptMN-Lab/SABER-ICML-2026/.
arXiv:2606.01478v2 Announce Type: replace Abstract: High-quality, large-scale synthetic data from simulations is becoming a cornerstone for pushing the capabilities of robot algorithms. While aerial robotics simulators have evolved to support specialized needs such as fidelity, differentiability, and swarms independently, a unified platform that can synthesize data across all these domains is missing. In this work, we propose Crazyflow, a simulator designed to push the limits of aerial-robotics algorithm development, from model-based to data-driven methods, gradient-based to sampling-based approaches, and single-agent to multi-agent systems. Compared to existing state-of-the-art drone simulators, it achieves speeds more than an order of magnitude faster for a single drone and can simulate thousands of swarms of 4000 drones each. Real-world experiments show Crazyflow supports both analytical-gradient-based policy learning, achieving sub-centimeter trajectory tracking accuracy without domain randomization, and sampling-based obstacle avoidance at speeds exceeding half a billion steps per second. Breaking the traditional train-then-deploy paradigm, we show that its unprecedented speed even enables in-flight reinforcement learning; we demonstrate this by throwing a physical drone into the air and training a recovery policy from scratch in 0.38 seconds, successfully stabilizing the drone. Crazyflow supports multiple levels of simulation abstraction, is directly compatible with all open-source Crazyflie models, and enables rapid reconfiguration across custom drone platforms and applications by providing a light-weight system identification pipeline. By pushing accuracy, speed, and differentiability simultaneously, Crazyflow serves as an open-source resource for synthetic data generation, with emerging capabilities for large-scale parallelization for online, in-execution learning and optimization, opening the door to novel algorithm development.
arXiv:2606.01546v2 Announce Type: replace Abstract: Sparse high-dimensional representations are conducive to uncovering nontrivial structures in unsupervised exploration of data. Such a representation can deal with the dense connectivity in graphs relevant to community detection problems. However, sparse high-dimensional representations are capable of doing more, including manifold tiling and feature learning. Conventional algorithms optimize in the space of computationally intractable completely positive matrices or relax the problem to the space of doubly nonnegative matrices that scale with sample size in a way rendering them impractical for large data sets. Some of these methods also impose a row sum constraint, such as double stochasticity. Row sum constraints have the added advantage of being shift-invariant, in the context of manifold tiling. Constraints on the row sum of output similarity matrices require nontrivial online learning rules. Addressing these needs, we propose a versatile online biologically plausible learning algorithm capable of learning sparse shift-invariant representations, useful for clustering, manifold tiling, or sparse coding, depending on the data structure.
arXiv:2606.01637v2 Announce Type: replace Abstract: Large language models are increasingly used in multi-agent systems, where they see and respond to other agents' answers. A key risk is conformity: a model may abandon its own answer simply because others agree on a different one. Prior studies show that LLMs often revise toward a majority answer, but it remains unclear whether these revisions help correct mistakes as often as they introduce new errors. In this paper, we conduct a controlled study in which an LLM first answers a question, then sees simulated peer responses before making a final decision. We manipulate two social cues: consensus structure and authority labels assigned to peers, and measure how they influence beneficial and harmful revisions. Across four open-weight LLMs and seven QA datasets, we find that peer agreement makes it much easier to mislead initially correct models than to correct initially wrong ones. Authority labels make models more likely to choose the endorsed answer, regardless of whether it is correct. More concerningly, generic reasoning interventions such as chain-of-thought and reflection do not reliably reduce harmful revision while preserving beneficial revision. These findings suggest that multi-agent LLM systems should verify peer answers rather than simply aggregate them.
arXiv:2606.08379v1 Announce Type: new Abstract: This study addresses the optimal execution of large stock sell programs by introducing TT-DAC-PS (Twin-Target Deterministic Actor-Critic with Policy Smoothing), a deterministic actor-critic architecture that combines twin exponential-moving-average critic targets with pessimistic min backup, TD3-style target policy smoothing noise, delayed actor updates, and conservative Q regularisation to curb overestimation. Exploration uses Ornstein-Uhlenbeck (OU) noise with a hybrid schedule: deterministic episode-wise decay, variance-guided adjustment based on recent reward dispersion, and a Soft Actor-Critic (SAC)-style temperature that is learned and mapped to the noise scale. The environment integrates Almgren-Chriss (AC) trade impact with Limit Order Book (LOB) prices and volumes, normalised state features, per-step volume participation caps, and a utility-based reward. The trade execution algorithm is applied to LOB data for ten U.S. stocks. Performance is assessed against reinforcement-learning baseline algorithms, including Proximal Policy Optimisation (PPO), Soft Actor-Critic (SAC), and Advantage Actor-Critic (A2C), as well as alternative trade execution algorithms, including Time-Weighted Average Price (TWAP), Volume-Weighted Average Price (VWAP), and AC. The proposed model consistently reduces mean implementation shortfall percentage with competitive variance, outperforming classical baselines and standard reinforcement-learning benchmark models.
arXiv:2606.08380v1 Announce Type: new Abstract: Domain-specific Field Programmable Gate Array (FPGA) architectures increasingly integrate specialized hardblocks, such as Tensor Slices, to accelerate artificial intelligence and machine learning workloads. Despite their efficiency benefits, these architectures remain difficult to program because designers typically rely on manual Register-Transfer Level (RTL) integration to access these hardblocks. This paper presents a compiler-agnostic methodology that enables high-level synthesis (HLS) tools to target custom FPGA hardblocks directly from C/C++ code. Architectural hardblocks are exposed as schedulable C-level operators using an RTL blackbox abstraction with explicit latency and initiation-interval contracts, allowing the HLS scheduler to optimize around specialized hardware without manual RTL orchestration. Unlike traditional uses of HLS blackboxes for external IP integration, our approach treats blackboxes as architectural abstractions, enabling scalable composition of C-level operators that target custom FPGA hardblocks without compiler modification. We evaluate the proposed flow using a Tensor Slice-based FPGA architecture with AMD Vitis HLS and the Verilog-to-Routing (VTR) toolchain. Across multiple matrix sizes, designs generated using the proposed C-Blackbox flow achieve lower area-delay product than behavioral HLS baselines while providing substantially higher productivity-adjusted efficiency than handwritten RTL implementations. These results demonstrate that domain-specific FPGA architectures can be made accessible through HLS while maintaining competitive hardware efficiency.
arXiv:2605.26078v3 Announce Type: replace Abstract: Wasserstein policy gradient (WPG) is a policy optimization method for reinforcement learning (RL) that exploits the optimal-transport geometry of action distributions. For the entropy-regularized RL objective, WPG evolves each state-conditional policy by transporting it along the action gradient of the soft Q-function together with a Langevin-type diffusion. Despite its appeal for continuous-control problems, its global convergence properties remain poorly understood. Standard Langevin analyses do not directly apply, because the RL objective depends on the policy through the Bellman recursion rather than through a static convex functional, and the Langevin drift is determined by the soft Q-function, whose regularity must be controlled along the policy iterates. In this paper, we develop a global convergence theory for WPG by exploiting the Bellman structure of entropy-regularized RL. We show that the role usually played by convexity can be replaced by a Bellman-based argument: the soft Bellman residual admits a statewise KL representation with respect to a Gibbs policy; Bellman contraction relates this residual to the global optimality gap; and a Bellman resolvent identity connects value improvement to relative Fisher information. Combined with a uniform log-Sobolev inequality (LSI) for the evolving Gibbs family, these ingredients yield a distributional Polyak--\L{}ojasiewicz condition. We further establish the regularity and uniform bounds needed to control the discretization error, thereby obtaining geometric contraction up to a discretization bias. Conceptually, our analysis shows that although entropy-regularized RL is not convex in the usual flat sense, the Bellman recursion induces a favorable Polyak--Lojasiewicz-type (PL) geometry that supports global convergence of WPG.
arXiv:2606.01751v2 Announce Type: replace Abstract: In long-context LLM serving, the prefill stage often dominates time-to-first-token and computational cost. Although Prefix Cache in vLLM/PagedAttention has been widely used to reuse identical prompt prefixes, repeated content in practical applications frequently appears as non-prefix, cross-request, cross-turn, and cross-agent segments, which makes conventional cache mechanisms insufficient. This paper presents SparseX, a segment-level KV Cache sharing method for common serving scenarios. SparseX uses contiguous token segments as reuse units and exploits Sparse-Q indices that naturally arise in KV Cache reuse workloads to estimate the key tokens that require correction. Based on this estimate, SparseX performs Sparse-KV Recomputation within a single forward pass, thereby restoring cross-segment contextual interactions under complex interleaved reuse patterns while avoiding additional models or separate preprocessing stages for token selection. SparseX further implements a full+sparse hybrid attention mode based on a layer-specific threshold: early layers retain full attention to obtain a more stable token-importance signal, and later layers switch to sparse recomputation to improve reuse quality on complex long-context tasks. We implement SparseX-vLLM on top of vLLM, integrating segment-level cache lookup, PagedAttention management, RoPE alignment, Sparse-Q token selection, and FlashAttention backends into a unified execution path. SparseX is model-agnostic, training-free, and compatible with Prefix Cache, and it provides unified support for common online serving scenarios including multi-round chat, retrieval-augmented generation (RAG), and agent workflows.
arXiv:2606.08388v1 Announce Type: new Abstract: Muon replaces a matrix gradient $G=U\Sigma V^\top$ by its polar factor $UV^\top$. This keeps the singular directions selected by the gradient, but makes the update spectrum flat. We study the optimization bias created by this operation. Under explicit alignment assumptions, we prove that the polar update is the one-step entropy-maximizing choice among bounded updates that use the gradient singular directions and do not adapt to the current weight spectrum. In an underdetermined regression model, we derive exact singular-value dynamics for continuous-time Muon and identify a measurement-dependent condition under which the normalized spectrum moves toward equal nonzero singular values. This geometry also rules out a common low-rank interpretation: at fixed Frobenius norm, Muon's distinguished state has a flat spectrum, whereas nuclear-norm minimization favors spectral concentration. Controlled matrix-sensing experiments separate the effect from simple gradient rescaling, show that norm-matched gradient descent does not reproduce Muon, and recover the predicted flattening trend across broad ablations. In small NanoGPT pretraining, Muon preserves stable rank, has a broad learning-rate plateau, and improves validation loss relative to AdamW; in a matched small-ViT control, the ranking reverses. The resulting picture is regime-dependent: Muon is not universally superior, but its flat-spectrum bias can help when many spectral directions need to remain active.