Searcharxiv⌕ Search

arXiv subjects

Hong Wang

Publications and source records attributed to Hong Wang.

At least 55 records · Page 3Linked to original sources

ArborKV: Structure-Aware KV Cache Management for Scaling Tree-based LLM Reasoning

Recent progress in LLM reasoning has increasingly shifted from single-pass generation to explicit search over intermediate reasoning states. Tree-of-Thoughts (ToT) organizes inference to tree-structured search with branching and backtracking, but it substantially amplifies the Key--Value (KV) cache: retaining KV states for a frontier of partial trajectories quickly becomes a memory bottleneck that limits throughput and constrains search depth and width under fixed hardware budgets. We address this challenge by observing that KV reuse in ToT-style inference is governed by search dynamics: near-term decoding depends primarily on the active branch and its ancestors, whereas inactive subtrees have low short-term reuse probability yet must remain recoverable for backtracking. Motivated by this, we propose ArborKV, a structure-aware eviction framework that couples a lightweight value estimator with a tree-aware allocation policy, and performs purely token-extractive eviction with lazy rehydration to support revisits. Experiments on ToT-style reasoning benchmarks show that ArborKV achieves up to ~4x peak KV-memory reduction while preserving near-full-retention accuracy, enabling larger search configurations under fixed device budgets that would otherwise run out of memory.

cs.AI↗

AOT-POT: Adaptive Operator Transformation for Large-Scale PDE Pre-training

Pre-training neural operators on diverse partial differential equation (PDE) datasets has emerged as a promising direction for building general-purpose surrogate models in scientific machine learning. However, the inherent complexity and structural diversity of PDE solution operators make multi-PDE pre-training fundamentally challenging. Existing methods mainly address this by increasing model capacity, while leaving the target solution operators unchanged. Inspired by classical numerical analysis, we instead propose to transform complex and diverse solution operators into simpler, better-aligned forms that are easier to model jointly. Since the optimal transformation varies across PDE types, it must be adaptive and input-dependent, allowing a single neural operator to approximate an entire family of operators. We instantiate this idea as AOT-POT (adaptive operator-transformation for pre-training operator transformer), which expands hidden representations into multiple parallel streams, adaptively aggregates and redistributes them before and after each sub-layer, and mixes streams through Sinkhorn-projected doubly stochastic matrices for stable training. These mechanisms together reshape diverse solution operators into a unified form that can be effectively modeled by a single architecture. Empirically, AOT-POT achieves state-of-the-art performance on 12 PDE benchmarks with only 3\% additional parameters, reducing relative L2 error by up to 77.6\% (40.9\% on average). Fine-tuning AOT-POT further reduces L2 error by up to 92\% on in-domain PDEs and 89\% on out-of-domain PDEs (unseen types during pre-training), demonstrating that adaptive operator transformation is an effective and complementary direction for advancing PDE foundation models beyond simply scaling model capacity.

cs.LG↗

Energy-Regularized Sequential Model Editing on Hyperspheres

Large language models (LLMs) require constant updates to remain aligned with evolving real-world knowledge. Model editing offers a lightweight alternative to retraining, but sequential editing often destabilizes representations and induces catastrophic forgetting. In this work, we seek to better understand and mitigate performance degradation caused by sequential editing. We hypothesize that hyperspherical uniformity, a property that maintains uniform distribution of neuron weights on a hypersphere, helps the model remain stable, retain prior knowledge, while still accommodate new updates. We use Hyperspherical Energy (HE) to quantify neuron uniformity during editing, and examine its correlation with editing performance. Empirical studies across widely used editing methods reveals a strong correlation between HE dynamics and editing performance, with editing failures consistently coinciding with high HE fluctuations. We further theoretically prove that HE dynamics impose a lower bound on the degradation of pretrained knowledge, highlighting why HE stability is crucial for knowledge retention. Motivated by these insights, we propose SPHERE (Sparse Projection for Hyperspherical Energy-Regularized Editing), an HE-driven regularization strategy that stabilizes neuron weight distributions, ultimately preserving prior knowledge while enabling reliable sequential updates. Specifically, SPHERE identifies a sparse space complementary to the principal hyperspherical directions of the pretrained weight matrices and projects new knowledge onto it, attenuating perturbations on the principal directions. Extensive experiments on LLaMA3 (8B) and Qwen2.5 (7B) show that SPHERE outperforms the best baseline in editing capability by an average of 16.41%, while most faithfully preserving general model performance, thereby offering a principled path toward reliable large-scale knowledge editing.

cs.CL↗

Rongzai agent: A Large Language Model-Based Autonomous Assistant for Rietveld Refinement of Neutron Diffraction Data

Neutron diffraction (ND) is an indispensable technique for determining atomic positions (especially light elements) and thus serves as a critical probe for revealing microscopic structures in materials science. However, traditional Rietveld refinement of ND data relies heavily on manual operation of specialized software, which is time-consuming, labor-intensive, and highly dependent on user expertise, severely hindering automated analysis. The automation of Rietveld refinement has long been a long-standing and challenging problem in crystallography. To address this challenge, this paper presents the Dr.Sai-Rongzai agent, an autonomous refinement assistant based on a large language model (LLM), a specialist knowledge base, and the GSAS-II refinement engine, achieving for the first time an intelligent refinement that integrates knowledge-driven decision-making. The agent accomplishes a fully automated workflow from natural language task parsing to autonomous decision-making, execution of refinement strategies, and report generation. Evaluation on five representative samples shows that the Rongzai agent achieves lower Rwp values than human specialists on three samples (2.88% vs. 4.42%, 5.06% vs. 5.40%, 7.60% vs. 9.00%), while on the other two samples its results are very close to those of the specialists. The agent is currently deployed at the China Spallation Neutron Source (CSNS) and is open for external user registration, providing an intelligent and user-friendly analytical tool for materials research. This work fully leverages the cutting-edge advantages of LLM, offers a new path to solve the long-standing problem of automated refinement, takes a key step toward intelligent and fully automated crystallographic analysis, and holds great potential to accelerate AI for Science discoveries in neutron-based materials characterization.

cond-mat.mtrl-sci↗

DeepSight: Long-Horizon World Modeling via Latent States Prediction for End-to-End Autonomous Driving

End-to-end autonomous driving systems are increasingly integrating Vision-Language Model (VLM) architectures, incorporating text reasoning or visual reasoning to enhance the robustness and accuracy of driving decisions. However, the reasoning mechanisms employed in most methods are direct adaptations from general domains, lacking in-depth exploration tailored to autonomous driving scenarios, particularly within visual reasoning modules. In this paper, we propose a driving world model that performs parallel prediction of latent semantic features for consecutive future frames in the bird's-eye-view (BEV) space, thereby enabling long-horizon modeling of future world states. We also introduce an efficient and adaptive text reasoning mechanism that utilizes additional social knowledge and reasoning capabilities to further improve driving performance in challenging long-tail scenarios. We present a novel, efficient, and effective approach that achieves state-of-the-art (SOTA) results on the closed-loop Bench2drive benchmark. Codes are available at: https://github.com/hotdogcheesewhite/DeepSight.

cs.CV↗

Rethinking Entropy Interventions in RLVR: An Entropy Change Perspective

Reinforcement Learning with Verifiable Rewards (RLVR) serves as a cornerstone technique for enhancing the reasoning capabilities of Large Language Models (LLMs). However, its training is often plagued by \emph{entropy collapse}, a rapid decline in policy entropy that limits exploration and undermines training effectiveness. While recent works attempt to mitigate this issue via several heuristic entropy interventions, the underlying mechanisms remain poorly understood. In this work, we conduct comprehensive theoretical and empirical analyses of entropy dynamics in RLVR, offering two main insights: (1) We derive a tight analytical approximation for token-level entropy change at each update step, revealing four governing factors and providing a unified theoretical framework to explain how existing methods influence entropy; (2) We reveal a fundamental limitation of recent approaches: they rely on heuristic adjustments to one or two of these factors, leaving other relevant factors unconsidered, thus inherently limiting their effectiveness. Motivated by these findings, we propose STEER, a principled entropy-modulation method that adaptively reweights tokens based on theoretically-estimated entropy variations. Extensive experiments across six mathematical reasoning and three coding benchmarks demonstrate that STEER effectively mitigates entropy collapse and consistently outperforms state-of-the-art baselines.

cs.LG↗

Accelerating Eigenvalue Dataset Generation via Chebyshev Subspace Filter

Eigenvalue problems are among the most important topics in many scientific disciplines. With the recent surge and development of machine learning, neural eigenvalue methods have attracted significant attention as a forward pass of inference requires only a tiny fraction of the computation time compared to traditional solvers. However, a key limitation is the requirement for large amounts of labeled data in training, including operators and their eigenvalues. To tackle this limitation, we propose a novel method, named Sorting Chebyshev Subspace Filter (SCSF), which significantly accelerates eigenvalue data generation by leveraging similarities between operators -- a factor overlooked by existing methods. Specifically, SCSF employs truncated fast Fourier transform sorting to group operators with similar eigenvalue distributions and constructs a Chebyshev subspace filter that leverages eigenpairs from previously solved problems to assist in solving subsequent ones, reducing redundant computations. To the best of our knowledge, SCSF is the first method to accelerate eigenvalue data generation. Experimental results show that SCSF achieves up to a 3.5 times speedup compared to various numerical solvers.

cs.LG↗

Scheduling Your LLM Reinforcement Learning with Reasoning Trees

Using Reinforcement Learning with Verifiable Rewards (RLVR) to optimize Large Language Models (LLMs) can be conceptualized as progressively editing a query's `Reasoning Tree'. This process involves exploring nodes (tokens) and dynamically modifying the model's policy at each node. When combined with data scheduling, this process yields further gains in data efficiency and accuracy. However, existing RLVR data scheduling methods typically rely on path-based metrics to rank queries, overlooking the reasoning tree structures of these queries. In this paper, we introduce a novel metric, namely Reasoning Score (r-score), which measures the query's learning difficulty based on the structure of its reasoning tree. Based on the r-score, we propose the Reasoning Tree Schedule (Re-Schedule), a scheduling algorithm that constructs a curriculum progressing from structurally simple (high r-score) to complex (low r-score) queries. Experiments on six math-reasoning benchmarks show that Re-Schedule significantly improves average accuracy, achieving gains of up to 3.2%. These strong results validate our approach and demonstrate that a structural understanding of the reasoning tree provides a more powerful and principled foundation for RLVR data scheduling.

cs.AI↗

Towards Lawful Autonomous Driving: Deriving Scenario-Aware Driving Requirements from Traffic Laws and Regulations

Driving in compliance with traffic laws and regulations is a basic requirement for human drivers, yet autonomous vehicles (AVs) can violate these requirements in diverse real-world scenarios. To encode law compliance into AV systems, conventional approaches use formal logic languages to explicitly specify behavioral constraints, but this process is labor-intensive, hard to scale, and costly to maintain. With recent advances in artificial intelligence, it is promising to leverage large language models (LLMs) to derive legal requirements from traffic laws and regulations. However, without explicitly grounding and reasoning in structured traffic scenarios, LLMs often retrieve irrelevant provisions or miss applicable ones, yielding imprecise requirements. To address this, we propose a novel pipeline that grounds LLM reasoning in a traffic scenario taxonomy through node-wise anchors that encode hierarchical semantics. On Chinese traffic laws and OnSite dataset (5,897 scenarios), our method improves law-scenario matching by 29.1\% and increases the accuracy of derived mandatory and prohibitive requirements by 36.9\% and 38.2\%, respectively. We further demonstrate real-world applicability by constructing a law-compliance layer for AV navigation and developing an onboard, real-time compliance monitor for in-field testing, providing a solid foundation for future AV development, deployment, and regulatory oversight.

cs.AI↗

Towards Understanding, Analyzing, and Optimizing Agentic AI Execution: A CPU-Centric Perspective

Agentic AI serving converts monolithic LLM-based inference to autonomous problem-solvers that can plan, call tools, perform reasoning, and adapt on the fly. Due to diverse task execution need, such serving heavily rely on heterogeneous CPU-GPU systems with majority of the external tools responsible for agentic capability, either run on or are orchestrated by the CPU. Towards having a deeper understanding of its role, this paper aims to characterize and analyze the system bottlenecks introduced by agentic AI workloads from a largely overlooked CPU-centric perspective. We first present a compile-time characterization of agentic AI execution and choose representative workloads to capture the algorithmic diversity. We then perform runtime characterization of the representative workloads analyzing the end-to-end latency and throughput on two different hardware systems to isolate respective architectural bottlenecks. Based on the insights on the bottlenecks, we finally present two scheduling optimizations, namely, 1. CPU-Aware Overlapped Micro-Batching (COMB) and 2. Mixed Agentic Scheduling (MAS) on homogeneous and heterogeneous agentic workloads, respectively. In specific, these methods optimize for improved CPU-GPU concurrent utilization while reducing skewed resource allocation for heterogeneous execution. Experimental evaluations on the two hardware systems demonstrate the efficacy of COMB in yielding up to 1.7x lower P50 latency in standalone homogeneous workload execution and up to 3.9x/1.8x lower service/total latency under homogeneous open-loop load. Additionally, for heterogeneous open-loop load, MAS can reduce the total latency for minority request-type by up to 2.37x/2.49x at P50/P90 percentile.

cs.AI↗

Optimization for Pinching Antennas System With Multiple Carriers and Rate Splitting Multiple Access

To meet the urgent demands for spectral efficiency and multi-user access in high-frequency application scenario for the sixth-generation wireless communication, this paper investigates a rate splitting multiple access (RSMA) system assisted by pinching antennas (PAs) with multiple waveguides and multiple carriers, aiming to maximize the overall system sum rate. To address the high sensitivity of high-frequency signals to PA movement in the overloaded scenarios, a two-stage PA position optimization method based on both path loss and phase shift error minimization is proposed under RSMA framework. Specifically, the first step is to perform coarse adjustment by minimizing large-scale path loss. Then, based on the derivation of a closed-form solution for the ideal phase shift in a single-user single-carrier case, the fine-grained positions of PAs are optimized via a one-dimensional line search to minimize the composite phase shift error across all users and carriers. In order to meet the quality of service requirements, the Lagrange dual method is employed to obtain the closed form of beamforming vectors after the PA positions are determined. Simulation results demonstrate that the proposed scheme achieves significant improvement in sum rate and confirm that RSMA exhibits stronger robustness to inaccurate PA positions caused by both discrete position channel estimation and physical hardware compared to other multiple-access techniques in PA-assisted systems. Furthermore, the results validate that fine-grained PA position adjustment is particularly crucial in high-frequency bands.

eess.SP↗

Decoding the Delta: Unifying Remote Sensing Change Detection and Understanding with Multimodal Large Language Models

While Multimodal Large Language Models (MLLMs) excel in general vision-language tasks, their application to remote sensing change understanding is hindered by a fundamental "temporal blindness". Existing architectures lack intrinsic mechanisms for multi-temporal contrastive reasoning and struggle with precise spatial grounding. To address this, we first introduce Delta-QA, a comprehensive benchmark comprising 180k visual question-answering samples. Delta-QA unifies pixel-level segmentation and visual question answering across bi- and tri-temporal scenarios, structuring change interpretation into four progressive cognitive dimensions. Methodologically, we propose Delta-LLaVA, a novel MLLM framework explicitly tailored for multi-temporal remote sensing interpretation. It overcomes the limitations of naive feature concatenation through three core innovations: a Change-Enhanced Attention module that systematically isolates and amplifies visual differences, a Change-SEG module utilizing Change Prior Embedding to extract differentiable difference features as input for the LLM, and Local Causal Attention to prevent cross-temporal contextual leakage. Extensive experiments demonstrate that Delta-LLaVA decisively outperforms leading generalist MLLMs and specialized segmentation models in complex change deduction and high-precision boundary localization, establishing a unified framework for earth observation intelligence.

cs.CV↗

VA-FastNavi-MARL: Real-Time Robot Control with Multimedia-Driven Meta-Reinforcement Learning

Interpreting dynamic, heterogeneous multimedia commands with real-time responsiveness is critical for Human-Robot Interaction. We present VA-FastNavi-MARL, a framework that aligns asynchronous audio-visual inputs into a unified latent representation. By treating diverse instructions as a distribution of navigable goals via Meta-Reinforcement Learning, our method enables rapid adaptation to unseen directives with negligible inference overhead. Unlike approaches bottlenecked by heavy sensory processing, our modality-agnostic stream ensures seamless, low-latency control. Validation on a multi-arm workspace confirms that VA-FastNavi-MARL significantly outperforms baselines in sample efficiency and maintains robust, real-time execution even under noisy multimedia streams.

cs.RO↗

DSO: Dual-Scale Neural Operators for Stable Long-term Fluid Dynamics Forecasting

Long-term fluid dynamics forecasting is a critically important problem in science and engineering. While neural operators have emerged as a promising paradigm for modeling systems governed by partial differential equations (PDEs), they often struggle with long-term stability and precision. We identify two fundamental failure modes in existing architectures: (1) local detail blurring, where fine-scale structures such as vortex cores and sharp gradients are progressively smoothed, and (2) global trend deviation, where the overall motion trajectory drifts from the ground truth during extended rollouts. We argue that these failures arise because existing neural operators treat local and global information processing uniformly, despite their inherently different evolution characteristics in physical systems. To bridge this gap, we propose the Dual-Scale Neural Operator (DSO), which explicitly decouples information processing into two complementary modules: depthwise separable convolutions for fine-grained local feature extraction and an MLP-Mixer for long-range global aggregation. Through numerical experiments on vortex dynamics, we demonstrate that nearby perturbations primarily affect local vortex structure while distant perturbations influence global motion trends, providing empirical validation for our design choice. Extensive experiments on turbulent flow benchmarks show that DSO achieves state-of-the-art accuracy while maintaining robust long-term stability, reducing prediction error by over 88% compared to existing neural operators.

cs.LG↗

Substrate-Mediated Evaporation and Stochastic Evolution of Supported Au Nanoparticles

We use in situ transmission electron microscopy with automated tracking to study supported gold nanoparticles (NPs) during high-temperature vacuum annealing. \rev{The average mass loss per NP is governed by a flat, nearly size-independent substrate-mediated evaporation profile.} On top of \rev{this mean shrinkage}, individual NPs show significant fluctuations in apparent growth or shrinkage, and NP volume follows a \rev{random-walk-like trajectory. To rationalize both the ensemble-mean behavior and the particle-resolved variability, we develop a self-consistent theory that couples substrate-mediated evaporation to collective 2D Ostwald-type mass exchange through a shared adatom field, described in terms of a renormalized screening length and background concentration. In the experimentally relevant regime, the theory predicts an approximately size-independent mean shrinkage rate and clarifies how net mass loss suppresses classical coarsening.} \rev{Superimposed on this deterministic drift, we quantify stochastic volume trajectories and capture their fluctuation spectrum with a minimal Langevin description consistent with intermittent adatom attachment and detachment events.} In addition, we characterize the lateral diffusive motion of NPs, which is responsible for their coalescence. Altogether, our results highlight that stochasticity is intrinsic at the nanoscale \rev{and that predicting the evolution of supported NPs at early and intermediate times requires a unified framework combining substrate-mediated evaporation, collective mass exchange, and stochastic fluctuations.

cond-mat.mtrl-sci↗

ODIN-Based CPU-GPU Architecture with Replay-Driven Simulation and Emulation

Integration of CPU and GPU technologies is a key enabler for modern AI and graphics workloads, combining control-oriented processing with massive parallel compute capability. As systems evolve toward chiplet-based architectures, pre-silicon validation of tightly coupled CPU-GPU subsystems becomes increasingly challenging due to complex validation framework setup, large design scale, high concurrency, non-deterministic execution, and intricate protocol interactions at chiplet boundaries, often resulting in long integration cycles. This paper presents a replay-driven validation methodology developed during the integration of a CPU subsystem, multiple Xe GPU cores, and a configurable Network-on-Chip (NoC) within a foundational SoC building block targeting the ODIN integrated chiplet architecture. By leveraging deterministic waveform capture and replay across both simulation and emulation using a single design database, complex GPU workloads and protocol sequences can be reproduced reliably at the system level. This approach significantly accelerates debug, improves integration confidence, and enables end-to-end system boot and workload execution within a single quarter, demonstrating the effectiveness of replay-based validation as a scalable methodology for chiplet-based systems.

cs.DC↗

A possible connection between quantum mechanics and spacetime

Recent developments in holographic gravity suggest that spacetime structure may be deeply related to quantum mechanics. In this work, from a different perspective, we demonstrate that wave-particle duality can be interpreted as the uncertainty of spacetime for the particle. Summarizing all possible trajectories in conventional path integral quantum mechanics can be transformed into the summation of all possible spacetime metrics. Furthermore, we emphasize that in conventional quantum gravity, it is possible that the classical matter fields correspond to quantum spacetime. We argue that this is not quite reasonable and propose a new path integral quantum gravity model based on the new interpretation of wave-particle duality. In this model, the aforementioned drawback of conventional quantum gravity naturally disappears.

physics.gen-ph↗

RA-SSU: Towards Fine-Grained Audio-Visual Learning with Region-Aware Sound Source Understanding

Audio-Visual Learning (AVL) is one fundamental task of multi-modality learning and embodied intelligence, displaying the vital role in scene understanding and interaction. However, previous researchers mostly focus on exploring downstream tasks from a coarse-grained perspective (e.g., audio-visual correspondence, sound source localization, and audio-visual event localization). Considering providing more specific scene perception details, we newly define a fine-grained Audio-Visual Learning task, termed Region-Aware Sound Source Understanding (RA-SSU), which aims to achieve region-aware, frame-level, and high-quality sound source understanding. To support this goal, we innovatively construct two corresponding datasets, i.e. fine-grained Music (f-Music) and fine-grained Lifescene (f-Lifescene), each containing annotated sound source masks and frame-by-frame textual descriptions. The f-Music dataset includes 3,976 samples across 22 scene types related to specific application scenarios, focusing on music scenes with complex instrument mixing. The f-Lifescene dataset contains 6,156 samples across 61 types representing diverse sounding objects in life scenarios. Moreover, we propose SSUFormer, a Sound-Source Understanding TransFormer benchmark that facilitates both the sound source segmentation and sound region description with a multi-modal input and multi-modal output architecture. Specifically, we design two modules for this framework, Mask Collaboration Module (MCM) and Mixture of Hierarchical-prompted Experts (MoHE), to respectively enhance the accuracy and enrich the elaboration of the sound source description. Extensive experiments are conducted on our two datasets to verify the feasibility of the task, evaluate the availability of the datasets, and demonstrate the superiority of the SSUFormer, which achieves SOTA performance on the Sound Source Understanding benchmark.

cs.CV↗