arXiv subjects

Tong Yang

Publications and source records attributed to Tong Yang.

At least 109 records · Page 6Linked to original sources

arXiv2025

Ionic KdV structure in weakly collisional plasmas

We consider the one-dimensional ions dynamics in weakly collisional plasmas governed by the Vlasov-Poisson-Landau system under the Boltzmann relation with the small collision frequency $ν>0$. It is observed in physical experiments that the interplay of nonlinearities and dispersion may lead to the formation of ion acoustic solitons that are described by the Korteweg-de Vries equation. In this paper, to capture the ionic KdV structure in the weak-collision regime, we study the combined cold-ions limit and longwave limit of the rescaled VPL system depending on a small scaling parameter $ε>0$. The main goal is to justify the uniform convergence of the VPL solutions to the KdV solutions over any finite time interval as $ε\to 0$ under restriction that $ε^{3/2}\lesssim ν\lesssim ε^{1/2}$. The proof is based on the energy method near local Maxwellians for making use of the Euler-Poisson dynamics under the longwave scaling. The KdV profiles, in particular including both velocity field and electric potential, may have large amplitude, which induces the cubic velocity growth. To overcome the $ε$-singularity in such multi-parameter limit problem, we design delicate velocity weighted energy functional and dissipation rate functional in the framework of macro-micro decomposition that is further incorporated with the Caflisch's decomposition. As an application of our approach, the global-in-time existence of solutions near global Maxwellians when the KdV profile is degenerate to a constant equilibrium is also established under the same scaling with $ε^{3}\lesssim ν\lesssim ε^{5/2}$. For the proof, the velocity weight is modified to depend on the solution itself, providing an extra quartic dissipation so as to obtain the global dynamics for most singular Coulomb potentials.

math.AP↗

arXiv2025

Exploration from a Primal-Dual Lens: Value-Incentivized Actor-Critic Methods for Sample-Efficient Online RL

Online reinforcement learning (RL) with complex function approximations such as transformers and deep neural networks plays a significant role in the modern practice of artificial intelligence. Despite its popularity and importance, balancing the fundamental trade-off between exploration and exploitation remains a long-standing challenge; in particular, we are still in lack of efficient and practical schemes that are backed by theoretical performance guarantees. Motivated by recent developments in exploration via optimistic regularization, this paper provides an interpretation of the principle of optimism through the lens of primal-dual optimization. From this fresh perspective, we set forth a new value-incentivized actor-critic (VAC) method, which optimizes a single easy-to-optimize objective integrating exploration and exploitation -- it promotes state-action and policy estimates that are both consistent with collected data transitions and result in higher value functions. Theoretically, the proposed VAC method has near-optimal regret guarantees under linear Markov decision processes (MDPs) in both finite-horizon and infinite-horizon settings, which can be extended to the general function approximation setting under appropriate assumptions.

cs.LG↗

arXiv2025

FAF: A Feature-Adaptive Framework for Few-Shot Time Series Forecasting

Multi-task and few-shot time series forecasting tasks are commonly encountered in scenarios such as the launch of new products in different cities. However, traditional time series forecasting methods suffer from insufficient historical data, which stems from a disregard for the generalized and specific features among different tasks. For the aforementioned challenges, we propose the Feature-Adaptive Time Series Forecasting Framework (FAF), which consists of three key components: the Generalized Knowledge Module (GKM), the Task-Specific Module (TSM), and the Rank Module (RM). During training phase, the GKM is updated through a meta-learning mechanism that enables the model to extract generalized features across related tasks. Meanwhile, the TSM is trained to capture diverse local dynamics through multiple functional regions, each of which learns specific features from individual tasks. During testing phase, the RM dynamically selects the most relevant functional region from the TSM based on input sequence features, which is then combined with the generalized knowledge learned by the GKM to generate accurate forecasts. This design enables FAF to achieve robust and personalized forecasting even with sparse historical observations We evaluate FAF on five diverse real-world datasets under few-shot time series forecasting settings. Experimental results demonstrate that FAF consistently outperforms baselines that include three categories of time series forecasting methods. In particular, FAF achieves a 41.81\% improvement over the best baseline, iTransformer, on the CO$_2$ emissions dataset.

cs.LG↗

arXiv2025

ScholarSearch: Benchmarking Scholar Searching Ability of LLMs

Large Language Models (LLMs)' search capabilities have garnered significant attention. Existing benchmarks, such as OpenAI's BrowseComp, primarily focus on general search scenarios and fail to adequately address the specific demands of academic search. These demands include deeper literature tracing and organization, professional support for academic databases, the ability to navigate long-tail academic knowledge, and ensuring academic rigor. Here, we proposed ScholarSearch, the first dataset specifically designed to evaluate the complex information retrieval capabilities of Large Language Models (LLMs) in academic research. ScholarSearch possesses the following key characteristics: Academic Practicality, where question content closely mirrors real academic learning and research environments, avoiding deliberately misleading models; High Difficulty, with answers that are challenging for single models (e.g., Grok DeepSearch or Gemini Deep Research) to provide directly, often requiring at least three deep searches to derive; Concise Evaluation, where limiting conditions ensure answers are as unique as possible, accompanied by clear sources and brief solution explanations, greatly facilitating subsequent audit and verification, surpassing the current lack of analyzed search datasets both domestically and internationally; and Broad Coverage, as the dataset spans at least 15 different academic disciplines. Through ScholarSearch, we expect to more precisely measure and promote the performance improvement of LLMs in complex academic information retrieval tasks. The data is available at: https://huggingface.co/datasets/PKU-DS-LAB/ScholarSearch

cs.IR↗

arXiv2025

SciDA: Scientific Dynamic Assessor of LLMs

Advancement in Large Language Models (LLMs) reasoning capabilities enables them to solve scientific problems with enhanced efficacy. Thereby, a high-quality benchmark for comprehensive and appropriate assessment holds significance, while existing ones either confront the risk of data contamination or lack involved disciplines. To be specific, due to the data source overlap of LLMs training and static benchmark, the keys or number pattern of answers inadvertently memorized (i.e. data contamination), leading to systematic overestimation of their reasoning capabilities, especially numerical reasoning. We propose SciDA, a multidisciplinary benchmark that consists exclusively of over 1k Olympic-level numerical computation problems, allowing randomized numerical initializations for each inference round to avoid reliance on fixed numerical patterns. We conduct a series of experiments with both closed-source and open-source top-performing LLMs, and it is observed that the performance of LLMs drop significantly under random numerical initialization. Thus, we provide truthful and unbiased assessments of the numerical reasoning capabilities of LLMs. The data is available at https://huggingface.co/datasets/m-a-p/SciDA

cs.CL↗

arXiv2025

Continuous Semi-Implicit Models

Semi-implicit distributions have shown great promise in variational inference and generative modeling. Hierarchical semi-implicit models, which stack multiple semi-implicit layers, enhance the expressiveness of semi-implicit distributions and can be used to accelerate diffusion models given pretrained score networks. However, their sequential training often suffers from slow convergence. In this paper, we introduce CoSIM, a continuous semi-implicit model that extends hierarchical semi-implicit models into a continuous framework. By incorporating a continuous transition kernel, CoSIM enables efficient, simulation-free training. Furthermore, we show that CoSIM achieves consistency with a carefully designed transition kernel, offering a novel approach for multistep distillation of generative models at the distributional level. Extensive experiments on image generation demonstrate that CoSIM performs on par or better than existing diffusion model acceleration methods, achieving superior performance on FD-DINOv2.

stat.ML↗

arXiv2025

HATA: Trainable and Hardware-Efficient Hash-Aware Top-k Attention for Scalable Large Model Inference

Large Language Models (LLMs) have emerged as a pivotal research area, yet the attention module remains a critical bottleneck in LLM inference, even with techniques like KVCache to mitigate redundant computations. While various top-$k$ attention mechanisms have been proposed to accelerate LLM inference by exploiting the inherent sparsity of attention, they often struggled to strike a balance between efficiency and accuracy. In this paper, we introduce HATA (Hash-Aware Top-$k$ Attention), a novel approach that systematically integrates low-overhead learning-to-hash techniques into the Top-$k$ attention process. Different from the existing top-k attention methods which are devoted to seeking an absolute estimation of qk score, typically with a great cost, HATA maps queries and keys into binary hash codes, and acquires the relative qk score order with a quite low cost, which is sufficient for realizing top-k attention. Extensive experiments demonstrate that HATA achieves up to 7.2$\times$ speedup compared to vanilla full attention while maintaining model accuracy. In addition, HATA outperforms the state-of-the-art top-$k$ attention methods in both accuracy and efficiency across multiple mainstream LLM models and diverse tasks. HATA is open source at https://github.com/gpzlx1/HATA.

cs.LG↗

arXiv2025

Beyond Invisibility: Learning Robust Visible Watermarks for Stronger Copyright Protection

As AI advances, copyrighted content faces growing risk of unauthorized use, whether through model training or direct misuse. Building upon invisible adversarial perturbation, recent works developed copyright protections against specific AI techniques such as unauthorized personalization through DreamBooth that are misused. However, these methods offer only short-term security, as they require retraining whenever the underlying model architectures change. To establish long-term protection aiming at better robustness, we go beyond invisible perturbation, and propose a universal approach that embeds \textit{visible} watermarks that are \textit{hard-to-remove} into images. Grounded in a new probabilistic and inverse problem-based formulation, our framework maximizes the discrepancy between the \textit{optimal} reconstruction and the original content. We develop an effective and efficient approximation algorithm to circumvent a intractable bi-level optimization. Experimental results demonstrate superiority of our approach across diverse scenarios.

cs.LG↗

arXiv2025

Detecting Flow Gaps in Data Streams

Data stream monitoring is a crucial task which has a wide range of applications. The majority of existing research in this area can be broadly classified into two types, monitoring value sum and monitoring value cardinality. In this paper, we define a third type, monitoring value variation, which can help us detect flow gaps in data streams. To realize this function, we propose GapFilter, leveraging the idea of Sketch for achieving speed and accuracy. To the best of our knowledge, this is the first work to detect flow gaps in data streams. Two key ideas of our work are the similarity absorption technique and the civilian-suspect mechanism. The similarity absorption technique helps in reducing memory usage and enhancing speed, while the civilian-suspect mechanism further boosts accuracy by organically integrating broad monitoring of overall flows with meticulous monitoring of suspicious flows.We have developed two versions of GapFilter. Speed-Oriented GapFilter (GapFilter-SO) emphasizes speed while maintaining satisfactory accuracy. Accuracy-Oriented GapFilter (GapFilter-AO) prioritizes accuracy while ensuring considerable speed. We provide a theoretical proof demonstrating that GapFilter secures high accuracy with minimal memory usage. Further, extensive experiments were conducted to assess the accuracy and speed of our algorithms. The results reveal that GapFilter-AO requires, on average, 1/32 of the memory to match the accuracy of the Straw-man solution. GapFilter-SO operates at a speed 3 times faster than the Straw-man solution. All associated source code has been open-sourced and is available on GitHub.

cs.DB↗

arXiv2025

Hidden Sketch: A Space-Efficient Reversible Sketch for Tracking Frequent Items in Data Streams

Modern data stream applications demand memory-efficient solutions for accurately tracking frequent items, such as heavy hitters and heavy changers, under strict resource constraints. Traditional sketches face inherent accuracy-memory trade-offs: they either lose precision to reduce memory usage or inflate memory costs to enable high recording capacity. This paper introduces Hidden Sketch, a space-efficient reversible data structure for key and frequency encoding. Our design uniquely combines a Reversible Bloom Filter (RBF) and a Count-Min (CM) Sketch for invertible key and frequency storage, enabling precise reconstruction for both keys and their frequencies with minimal memory. Theoretical analysis establishes Hidden Sketch's space complexity and guaranteed reversibility, while extensive experiments demonstrate its substantial improvements in accuracy and space efficiency in frequent item tracking tasks. By eliminating the trade-off between reversibility and space efficiency, Hidden Sketch provides a scalable foundation for real-time stream analytics in resource-constrained environments.

cs.DB↗

arXiv2025

ResidualSketch: Enhancing Layer Efficiency and Error Reduction in Hierarchical Heavy Hitter Detection with ResNet Innovations

In network management, swiftly and accurately identifying traffic anomalies, including Distributed Denial-of-Service (DDoS) attacks and unexpected network disruptions, is essential for network stability and security. Key to this process is the detection of Hierarchical Heavy Hitters (HHH), which significantly aids in the management of high-speed IP traffic. This study introduces ResidualSketch, a novel algorithm for HHH detection in hierarchical traffic analysis. ResidualSketch distinguishes itself by incorporating Residual Blocks and Residual Connections at crucial layers within the IP hierarchy, thus mitigating the Gradual Error Diffusion (GED) phenomenon in previous methods and reducing memory overhead while maintaining low update latency. Through comprehensive experiments on various datasets, we demonstrate that ResidualSketch outperforms existing state-of-the-art solutions in terms of accuracy and update speed across multiple layers of the network hierarchy. All related codes of ResidualSketch are open-source at GitHub.

cs.DS↗

arXiv2025

FairKV: Balancing Per-Head KV Cache for Fast Multi-GPU Inference

KV cache techniques in Transformer models aim to reduce redundant computations at the expense of substantially increased memory usage, making KV cache compression an important and popular research topic. Recently, state-of-the-art KV cache compression methods implement imbalanced, per-head allocation algorithms that dynamically adjust the KV cache budget for each attention head, achieving excellent performance in single-GPU scenarios. However, we observe that such imbalanced compression leads to significant load imbalance when deploying multi-GPU inference, as some GPUs become overburdened while others remain underutilized. In this paper, we propose FairKV, a method designed to ensure fair memory usage among attention heads in systems employing imbalanced KV cache compression. The core technique of FairKV is Fair-Copying, which replicates a small subset of memory-intensive attention heads across GPUs using data parallelism to mitigate load imbalance. Our experiments on popular models, including LLaMA 70b and Mistral 24b model, demonstrate that FairKV increases throughput by 1.66x compared to standard tensor parallelism inference. Our code will be released as open source upon acceptance.

cs.DC↗

arXiv2025

ReFocus: Reinforcing Mid-Frequency and Key-Frequency Modeling for Multivariate Time Series Forecasting

Recent advancements have progressively incorporated frequency-based techniques into deep learning models, leading to notable improvements in accuracy and efficiency for time series analysis tasks. However, the Mid-Frequency Spectrum Gap in the real-world time series, where the energy is concentrated at the low-frequency region while the middle-frequency band is negligible, hinders the ability of existing deep learning models to extract the crucial frequency information. Additionally, the shared Key-Frequency in multivariate time series, where different time series share indistinguishable frequency patterns, is rarely exploited by existing literature. This work introduces a novel module, Adaptive Mid-Frequency Energy Optimizer, based on convolution and residual learning, to emphasize the significance of mid-frequency bands. We also propose an Energy-based Key-Frequency Picking Block to capture shared Key-Frequency, which achieves superior inter-series modeling performance with fewer parameters. A novel Key-Frequency Enhanced Training strategy is employed to further enhance Key-Frequency modeling, where spectral information from other channels is randomly introduced into each channel. Our approach advanced multivariate time series forecasting on the challenging Traffic, ECL, and Solar benchmarks, reducing MSE by 4%, 6%, and 5% compared to the previous SOTA iTransformer. Code is available at this GitHub Repository: https://github.com/Levi-Ackman/ReFocus.

cs.LG↗

arXiv2025

LiNo: Advancing Recursive Residual Decomposition of Linear and Nonlinear Patterns for Robust Time Series Forecasting

Forecasting models are pivotal in a data-driven world with vast volumes of time series data that appear as a compound of vast Linear and Nonlinear patterns. Recent deep time series forecasting models struggle to utilize seasonal and trend decomposition to separate the entangled components. Such a strategy only explicitly extracts simple linear patterns like trends, leaving the other linear modes and vast unexplored nonlinear patterns to the residual. Their flawed linear and nonlinear feature extraction models and shallow-level decomposition limit their adaptation to the diverse patterns present in real-world scenarios. Given this, we innovate Recursive Residual Decomposition by introducing explicit extraction of both linear and nonlinear patterns. This deeper-level decomposition framework, which is named LiNo, captures linear patterns using a Li block which can be a moving average kernel, and models nonlinear patterns using a No block which can be a Transformer encoder. The extraction of these two patterns is performed alternatively and recursively. To achieve the full potential of LiNo, we develop the current simple linear pattern extractor to a general learnable autoregressive model, and design a novel No block that can handle all essential nonlinear patterns. Remarkably, the proposed LiNo achieves state-of-the-art on thirteen real-world benchmarks under univariate and multivariate forecasting scenarios. Experiments show that current forecasting models can deliver more robust and precise results through this advanced Recursive Residual Decomposition. We hope this work could offer insight into designing more effective forecasting models. Code is available at this Repository: https://github.com/Levi-Ackman/LiNo.

cs.LG↗

arXiv2025

Value-Incentivized Preference Optimization: A Unified Approach to Online and Offline RLHF

Reinforcement learning from human feedback (RLHF) has demonstrated great promise in aligning large language models (LLMs) with human preference. Depending on the availability of preference data, both online and offline RLHF are active areas of investigation. A key bottleneck is understanding how to incorporate uncertainty estimation in the reward function learned from the preference data for RLHF, regardless of how the preference data is collected. While the principles of optimism or pessimism under uncertainty are well-established in standard reinforcement learning (RL), a practically-implementable and theoretically-grounded form amenable to large language models is not yet available, as standard techniques for constructing confidence intervals become intractable under arbitrary policy parameterizations. In this paper, we introduce a unified approach to online and offline RLHF -- value-incentivized preference optimization (VPO) -- which regularizes the maximum-likelihood estimate of the reward function with the corresponding value function, modulated by a $\textit{sign}$ to indicate whether the optimism or pessimism is chosen. VPO also directly optimizes the policy with implicit reward modeling, and therefore shares a simpler RLHF pipeline similar to direct preference optimization. Theoretical guarantees of VPO are provided for both online and offline settings, matching the rates of their standard RL counterparts. Moreover, experiments on text summarization and dialog verify the practicality and effectiveness of VPO.

cs.LG↗

arXiv2025

Faster WIND: Accelerating Iterative Best-of-$N$ Distillation for LLM Alignment

Recent advances in aligning large language models with human preferences have corroborated the growing importance of best-of-N distillation (BOND). However, the iterative BOND algorithm is prohibitively expensive in practice due to the sample and computation inefficiency. This paper addresses the problem by revealing a unified game-theoretic connection between iterative BOND and self-play alignment, which unifies seemingly disparate algorithmic paradigms. Based on the connection, we establish a novel framework, WIN rate Dominance (WIND), with a series of efficient algorithms for regularized win rate dominance optimization that approximates iterative BOND in the parameter space. We provides provable sample efficiency guarantee for one of the WIND variant with the square loss objective. The experimental results confirm that our algorithm not only accelerates the computation, but also achieves superior sample efficiency compared to existing methods.

cs.LG↗

arXiv2025

Incentivize without Bonus: Provably Efficient Model-based Online Multi-agent RL for Markov Games

Multi-agent reinforcement learning (MARL) lies at the heart of a plethora of applications involving the interaction of a group of agents in a shared unknown environment. A prominent framework for studying MARL is Markov games, with the goal of finding various notions of equilibria in a sample-efficient manner, such as the Nash equilibrium (NE) and the coarse correlated equilibrium (CCE). However, existing sample-efficient approaches either require tailored uncertainty estimation under function approximation, or careful coordination of the players. In this paper, we propose a novel model-based algorithm, called VMG, that incentivizes exploration via biasing the empirical estimate of the model parameters towards those with a higher collective best-response values of all the players when fixing the other players' policies, thus encouraging the policy to deviate from its current equilibrium for more exploration. VMG is oblivious to different forms of function approximation, and permits simultaneous and uncoupled policy updates of all players. Theoretically, we also establish that VMG achieves a near-optimal regret for finding both the NEs of two-player zero-sum Markov games and CCEs of multi-player general-sum Markov games under linear function approximation in an online environment, which nearly match their counterparts with sophisticated uncertainty quantification.

cs.LG↗

arXiv2025

Structural stability of boundary layers in the entire subsonic regime

Despite the physical importance, there are limited mathematical theories for the compressible Navier-Stokes equations with strong boundary layers. This is mainly due to the absence of a stream function structure, unlike the extensively studied incompressible fluid dynamics in two dimensions. This paper aims to establish the structural stability of boundary layer profiles in the form of shear flow for the two-dimensional steady compressible Navier-Stokes equations. Our estimates are uniform across the entire subsonic regime, where the Mach number $m\in (0,1)$. As a byproduct, we provide the first result concerning the low Mach number limit in the presence of Prandtl boundary layers. The proof relies on the quasi-compressible-Stokes iteration introduced in [38], along with a subtle analysis of the interplay between density and velocity variables in different frequency regimes, and the identification of cancellations in higher-order estimates.

math.AP↗