arXiv subjects

Tao Yang

Publications and source records attributed to Tao Yang.

At least 127 records · Page 7Linked to original sources

arXiv2025

Grey Rhino Warning: IPv6 is Becoming Fertile Ground for Reflection Amplification Attacks

Distributed Denial-of-Service (DDoS) attacks represent a cost-effective and potent threat to network stability. While extensively studied in IPv4 networks, DDoS implications in IPv6 remain underexplored. The vast IPv6 address space renders brute-force scanning and amplifier testing for all active addresses impractical. Innovatively, this work investigates AS-level vulnerabilities to reflection amplification attacks in IPv6. One prerequisite for amplification presence is that it is located in a vulnerable autonomous system (AS) without inbound source address validation (ISAV) deployment. Hence, the analysis focuses on two critical aspects: global detection of ISAV deployment and identification of amplifiers within vulnerable ASes. Specifically, we develop a methodology combining ICMP Time Exceeded mechanisms for ISAV detection, employ IPv6 address scanning for amplifier identification, and utilize dual vantage points for amplification verification. Experimental results reveal that 4,460 ASes (61.36% of measured networks) lack ISAV deployment. Through scanning approximately 47M active addresses, we have identified reflection amplifiers in 3,507 ASes. The analysis demonstrates that current IPv6 networks are fertile grounds for reflection amplification attacks, alarming network security.

cs.NI↗

arXiv2025

FreePRM: Training Process Reward Models Without Ground Truth Process Labels

Recent advancements in Large Language Models (LLMs) have demonstrated that Process Reward Models (PRMs) play a crucial role in enhancing model performance. However, training PRMs typically requires step-level labels, either manually annotated or automatically generated, which can be costly and difficult to obtain at scale. To address this challenge, we introduce FreePRM, a weakly supervised framework for training PRMs without access to ground-truth step-level labels. FreePRM first generates pseudo step-level labels based on the correctness of final outcome, and then employs Buffer Probability to eliminate impact of noise inherent in pseudo labeling. Experimental results show that FreePRM achieves an average F1 score of 53.0% on ProcessBench, outperforming fully supervised PRM trained on Math-Shepherd by +24.1%. Compared to other open-source PRMs, FreePRM outperforms upon RLHFlow-PRM-Mistral-8B (28.4%) by +24.6%, EurusPRM (31.3%) by +21.7%, and Skywork-PRM-7B (42.1%) by +10.9%. This work introduces a new paradigm in PRM training, significantly reducing reliance on costly step-level annotations while maintaining strong performance.

cs.CL↗

arXiv2025

Discriminative Policy Optimization for Token-Level Reward Models

Process reward models (PRMs) provide more nuanced supervision compared to outcome reward models (ORMs) for optimizing policy models, positioning them as a promising approach to enhancing the capabilities of LLMs in complex reasoning tasks. Recent efforts have advanced PRMs from step-level to token-level granularity by integrating reward modeling into the training of generative models, with reward scores derived from token generation probabilities. However, the conflict between generative language modeling and reward modeling may introduce instability and lead to inaccurate credit assignments. To address this challenge, we revisit token-level reward assignment by decoupling reward modeling from language generation and derive a token-level reward model through the optimization of a discriminative policy, termed the Q-function Reward Model (Q-RM). We theoretically demonstrate that Q-RM explicitly learns token-level Q-functions from preference data without relying on fine-grained annotations. In our experiments, Q-RM consistently outperforms all baseline methods across various benchmarks. For example, when integrated into PPO/REINFORCE algorithms, Q-RM enhances the average Pass@1 score by 5.85/4.70 points on mathematical reasoning tasks compared to the ORM baseline, and by 4.56/5.73 points compared to the token-level PRM counterpart. Moreover, reinforcement learning with Q-RM significantly enhances training efficiency, achieving convergence 12 times faster than ORM on GSM8K and 11 times faster than step-level PRM on MATH. Code and data are available at https://github.com/homzer/Q-RM.

cs.CL↗

arXiv2025

Structured Memory Mechanisms for Stable Context Representation in Large Language Models

This paper addresses the limitations of large language models in understanding long-term context. It proposes a model architecture equipped with a long-term memory mechanism to improve the retention and retrieval of semantic information across paragraphs and dialogue turns. The model integrates explicit memory units, gated writing mechanisms, and attention-based reading modules. A forgetting function is introduced to enable dynamic updates of memory content, enhancing the model's ability to manage historical information. To further improve the effectiveness of memory operations, the study designs a joint training objective. This combines the main task loss with constraints on memory writing and forgetting. It guides the model to learn better memory strategies during task execution. Systematic evaluation across multiple subtasks shows that the model achieves clear advantages in text generation consistency, stability in multi-turn question answering, and accuracy in cross-context reasoning. In particular, the model demonstrates strong semantic retention and contextual coherence in long-text tasks and complex question answering scenarios. It effectively mitigates the context loss and semantic drift problems commonly faced by traditional language models when handling long-term dependencies. The experiments also include analysis of different memory structures, capacity sizes, and control strategies. These results further confirm the critical role of memory mechanisms in language understanding. They demonstrate the feasibility and effectiveness of the proposed approach in both architectural design and performance outcomes.

cs.CL↗

arXiv2025

DASH: Input-Aware Dynamic Layer Skipping for Efficient LLM Inference with Markov Decision Policies

Large language models (LLMs) have achieved remarkable performance across a wide range of NLP tasks. However, their substantial inference cost poses a major barrier to real-world deployment, especially in latency-sensitive scenarios. To address this challenge, we propose \textbf{DASH}, an adaptive layer-skipping framework that dynamically selects computation paths conditioned on input characteristics. We model the skipping process as a Markov Decision Process (MDP), enabling fine-grained token-level decisions based on intermediate representations. To mitigate potential performance degradation caused by skipping, we introduce a lightweight compensation mechanism that injects differential rewards into the decision process. Furthermore, we design an asynchronous execution strategy that overlaps layer computation with policy evaluation to minimize runtime overhead. Experiments on multiple LLM architectures and NLP benchmarks show that our method achieves significant inference acceleration while maintaining competitive task performance, outperforming existing methods.

cs.CL↗

arXiv2025

LCD: Advancing Extreme Low-Bit Clustering for Large Language Models via Knowledge Distillation

Large language models (LLMs) have achieved significant progress in natural language processing but face challenges in deployment due to high memory and computational requirements. Weight quantization is a common approach to address these issues, yet achieving effective low-bit compression remains challenging. This paper presents LCD, which unifies the learning of clustering-based quantization within a knowledge distillation framework. Using carefully designed optimization techniques, LCD preserves LLM performance even at ultra-low bit widths of 2-3 bits. Additionally, LCD compresses activations through smoothing and accelerates inference with a LUT-based design. Experimental results show that LCD outperforms existing methods and delivers up to a 6.2x speedup in inference. Notably, LCD is shown to be more cost-effective, making it a practical solution for real-world applications.

cs.LG↗

arXiv2025

Tianyi: A Traditional Chinese Medicine all-rounder language model and its Real-World Clinical Practice

Natural medicines, particularly Traditional Chinese Medicine (TCM), are gaining global recognition for their therapeutic potential in addressing human symptoms and diseases. TCM, with its systematic theories and extensive practical experience, provides abundant resources for healthcare. However, the effective application of TCM requires precise syndrome diagnosis, determination of treatment principles, and prescription formulation, which demand decades of clinical expertise. Despite advancements in TCM-based decision systems, machine learning, and deep learning research, limitations in data and single-objective constraints hinder their practical application. In recent years, large language models (LLMs) have demonstrated potential in complex tasks, but lack specialization in TCM and face significant challenges, such as too big model scale to deploy and issues with hallucination. To address these challenges, we introduce Tianyi with 7.6-billion-parameter LLM, a model scale proper and specifically designed for TCM, pre-trained and fine-tuned on diverse TCM corpora, including classical texts, expert treatises, clinical records, and knowledge graphs. Tianyi is designed to assimilate interconnected and systematic TCM knowledge through a progressive learning manner. Additionally, we establish TCMEval, a comprehensive evaluation benchmark, to assess LLMs in TCM examinations, clinical tasks, domain-specific question-answering, and real-world trials. The extensive evaluations demonstrate the significant potential of Tianyi as an AI assistant in TCM clinical practice and research, bridging the gap between TCM knowledge and practical application.

cs.CL↗

arXiv2025

MPS-Prover: Advancing Stepwise Theorem Proving by Multi-Perspective Search and Data Curation

Automated Theorem Proving (ATP) in formal languages remains a formidable challenge in AI, demanding rigorous logical deduction and navigating vast search spaces. While large language models (LLMs) have shown promising performance, existing stepwise provers often suffer from biased search guidance, leading to inefficiencies and suboptimal proof strategies. This paper introduces the Multi-Perspective Search Prover (MPS-Prover), a novel stepwise ATP system designed to overcome these limitations. MPS-Prover incorporates two key innovations: a highly effective post-training data curation strategy that prunes approximately 40% of redundant training data without sacrificing performance, and a multi-perspective tree search mechanism. This search integrates a learned critic model with strategically designed heuristic rules to diversify tactic selection, prevent getting trapped in unproductive states, and enhance search robustness. Extensive evaluations demonstrate that MPS-Prover achieves state-of-the-art performance on multiple challenging benchmarks, including miniF2F and ProofNet, outperforming prior 7B parameter models. Furthermore, our analyses reveal that MPS-Prover generates significantly shorter and more diverse proofs compared to existing stepwise and whole-proof methods, highlighting its efficiency and efficacy. Our work advances the capabilities of LLM-based formal reasoning and offers a robust framework and a comprehensive analysis for developing more powerful theorem provers.

cs.AI↗

arXiv2025

One-Point Sampling for Distributed Bandit Convex Optimization with Time-Varying Constraints

This paper considers the distributed bandit convex optimization problem with time-varying constraints. In this problem, the global loss function is the average of all the local convex loss functions, which are unknown beforehand. Each agent iteratively makes its own decision subject to time-varying inequality constraints which can be violated but are fulfilled in the long run. For a uniformly jointly strongly connected time-varying directed graph, a distributed bandit online primal-dual projection algorithm with one-point sampling is proposed. We show that sublinear dynamic network regret and network cumulative constraint violation are achieved if the path-length of the benchmark also increases in a sublinear manner. In addition, an $\mathcal{O}({T^{3/4 + g}})$ static network regret bound and an $\mathcal{O}( {{T^{1 - {g}/2}}} )$ network cumulative constraint violation bound are established, where $T$ is the total number of iterations and $g \in ( {0,1/4} )$ is a trade-off parameter. Moreover, a reduced static network regret bound $\mathcal{O}( {T^{2/3 + 4g /3}} )$ is established for strongly convex local loss functions. Finally, a numerical example is presented to validate the theoretical results.

eess.SY↗

arXiv2025

Dynamic Superblock Pruning for Fast Learned Sparse Retrieval

This paper proposes superblock pruning (SP) during top-k online document retrieval for learned sparse representations. SP structures the sparse index as a set of superblocks on a sequence of document blocks and conducts a superblock-level selection to decide if some superblocks can be pruned before visiting their child blocks. SP generalizes the previous flat block or cluster-based pruning, allowing the early detection of groups of documents that cannot or are less likely to appear in the final top-k list. SP can accelerate sparse retrieval in a rank-safe or approximate manner under a high-relevance competitiveness constraint. Our experiments show that the proposed scheme significantly outperforms state-of-the-art baselines on MS MARCO passages on a single-threaded CPU.

cs.IR↗

arXiv2025

A Deep Learning Framework for Sequence Mining with Bidirectional LSTM and Multi-Scale Attention

This paper addresses the challenges of mining latent patterns and modeling contextual dependencies in complex sequence data. A sequence pattern mining algorithm is proposed by integrating Bidirectional Long Short-Term Memory (BiLSTM) with a multi-scale attention mechanism. The BiLSTM captures both forward and backward dependencies in sequences, enhancing the model's ability to perceive global contextual structures. At the same time, the multi-scale attention module assigns adaptive weights to key feature regions under different window sizes. This improves the model's responsiveness to both local and global important information. Extensive experiments are conducted on a publicly available multivariate time series dataset. The proposed model is compared with several mainstream sequence modeling methods. Results show that it outperforms existing models in terms of accuracy, precision, and recall. This confirms the effectiveness and robustness of the proposed architecture in complex pattern recognition tasks. Further ablation studies and sensitivity analyses are carried out to investigate the effects of attention scale and input sequence length on model performance. These results provide empirical support for structural optimization of the model.

cs.LG↗

arXiv2025

Invariant Federated Learning for Edge Intelligence: Mitigating Heterogeneity and Asynchrony via Exit Strategy and Invariant Penalty

This paper provides an invariant federated learning system for resource-constrained edge intelligence. This framework can mitigate the impact of heterogeneity and asynchrony via exit strategy and invariant penalty. We introduce parameter orthogonality into edge intelligence to measure the contribution or impact of heterogeneous and asynchronous clients. It is proved in this paper that the exit of abnormal edge clients can guarantee the effect of the model on most clients. Meanwhile, to ensure the models' performance on exited abnormal clients and those who lack training resources, we propose Federated Learning with Invariant Penalty for Generalization (FedIPG) by constructing the approximate orthogonality of the invariant parameters and the heterogeneous parameters. Theoretical proof shows that FedIPG reduces the Out-Of-Distribution prediction loss without increasing the communication burden. The performance of FedIPG combined with an exit strategy is tested empirically in multiple scales using four datasets. It shows our system can enhance In-Distribution performance and outperform the state-of-the-art algorithm in Out-Of-Distribution generalization while maintaining model convergence. Additionally, the results of the visual experiment prove that FedIPG contains preliminary causality in terms of ignoring confounding features.

cs.LG↗

arXiv2025

RORem: Training a Robust Object Remover with Human-in-the-Loop

Despite the significant advancements, existing object removal methods struggle with incomplete removal, incorrect content synthesis and blurry synthesized regions, resulting in low success rates. Such issues are mainly caused by the lack of high-quality paired training data, as well as the self-supervised training paradigm adopted in these methods, which forces the model to in-paint the masked regions, leading to ambiguity between synthesizing the masked objects and restoring the background. To address these issues, we propose a semi-supervised learning strategy with human-in-the-loop to create high-quality paired training data, aiming to train a Robust Object Remover (RORem). We first collect 60K training pairs from open-source datasets to train an initial object removal model for generating removal samples, and then utilize human feedback to select a set of high-quality object removal pairs, with which we train a discriminator to automate the following training data generation process. By iterating this process for several rounds, we finally obtain a substantial object removal dataset with over 200K pairs. Fine-tuning the pre-trained stable diffusion model with this dataset, we obtain our RORem, which demonstrates state-of-the-art object removal performance in terms of both reliability and image quality. Particularly, RORem improves the object removal success rate over previous methods by more than 18\%. The dataset, source code and trained model are available at https://github.com/leeruibin/RORem.

cs.CV↗

arXiv2025

Adaptive and Efficient Log Parsing as a Cloud Service

Logs are a critical data source for cloud systems, enabling advanced features like monitoring, alerting, and root cause analysis. However, the massive scale and diverse formats of unstructured logs pose challenges for adaptable, efficient, and accurate parsing methods. This paper introduces ByteBrain-LogParser, an innovative log parsing framework designed specifically for cloud environments. ByteBrain-LogParser employs a hierarchical clustering algorithm to allow real-time precision adjustments, coupled with optimizations such as positional similarity distance, deduplication, and hash encoding to enhance performance. Experiments on large-scale datasets show that it processes 229,000 logs per second on average, achieving an 840% speedup over the fastest baseline while maintaining accuracy comparable to state-of-the-art methods. Real-world evaluations further validate its efficiency and adaptability, demonstrating its potential as a robust cloud-based log parsing solution.

cs.SE↗

arXiv2025

Manipulating magnetization by orbital current from a light metal Ti

The orbital Hall effect, which does not rely on the spin-orbit coupling, has recently emerged as a promising mechanism for electrically manipulating magnetization in thin-film ferromagnets. Despite its potential, direct experimental observation of magnetization switching driven by orbital currents has been challenging, primarily because there is no direct exchange coupling between orbital angular momentum and local spin based magnetic moments. In this study, we present a compensated design to directly probe the contribution of orbital currents in the most promising light metal titanium (Ti), where symmetric layer structures allow zeroing out of the net spin current. By varying the thickness of the Ti layer in Ti(t)/Pt/Co/Pt/Co/Pt multilayers, we demonstrate the ability to control the magnetization switching polarity. We deduce the orbital charge conversion efficiency of the Ti layer to be approximately 0.17. These findings not only confirm the presence of the orbital Hall effect in Ti but also suggest that orbital currents may be promising candidates for developing energy-efficient magnetic devices with enhanced performance and scalability.

cond-mat.mtrl-sci↗

arXiv2025

Reduced Network Cumulative Constraint Violation for Distributed Bandit Convex Optimization under Slater Condition

This paper studies the distributed bandit convex optimization problem with time-varying inequality constraints, where the goal is to minimize network regret and cumulative constraint violation. To calculate network cumulative constraint violation, existing distributed bandit online algorithms solving this problem directly use the clipped constraint function to replace its original constraint function. However, the use of the clipping operation renders Slater condition (i.e, there exists a point that strictly satisfies the inequality constraints at all iterations) ineffective to achieve reduced network cumulative constraint violation. To tackle this challenge, we propose a new distributed bandit online primal-dual algorithm. If local loss functions are convex, we show that the proposed algorithm establishes sublinear network regret and cumulative constraint violation bounds. When Slater condition holds, the network cumulative constraint violation bound is reduced. In addition, if local loss functions are strongly convex, for the case where strongly convex parameters are unknown, the network regret bound is reduced. For the case where strongly convex parameters are known, the network regret and cumulative constraint violation bounds are further reduced. To the best of our knowledge, this paper is among the first to establish reduced (network) cumulative constraint violation bounds for (distributed) bandit convex optimization with time-varying constraints under Slater condition. Finally, a numerical example is provided to verify the theoretical results.

eess.SY↗

arXiv2025

MAB-Based Channel Scheduling for Asynchronous Federated Learning in Non-Stationary Environments

Federated learning enables distributed model training across clients without raw data exchange, but in wireless implementations, frequent parameter updates cause high communication overhead. Existing research often assumes known channel state information (CSI) or stationary channels, though practical wireless channels are non-stationary due to fading, user mobility, and attacks, leading to unpredictable transmission failures and exacerbating client staleness, which hampers model convergence. To tackle these challenges, we propose an asynchronous federated learning scheduling framework for non-stationary channels that aims to reduce client staleness while enhancing communication efficiency and fairness. Our framework considers two scenarios: extremely non-stationary and piecewise-stationary channels. Age of Information (AoI) quantifies client staleness under these conditions. We conduct convergence analysis to examine the impact of AoI and per-round client participation on learning performance and formulate the scheduling problem as a multi-armed bandit (MAB) problem. We derive theoretical lower bounds on AoI regret and develop scheduling strategies based on GLR-CUCB and M-exp3 algorithms, including upper bounds on AoI regret. To address imbalanced client updates, we propose an adaptive matching strategy that incorporates marginal utility and fairness considerations. Simulation results show that our algorithm achieves sub-linear AoI regret, accelerates convergence, and promotes fairer aggregation.

cs.LG↗

arXiv2025

HunyuanProver: A Scalable Data Synthesis Framework and Guided Tree Search for Automated Theorem Proving

We introduce HunyuanProver, an language model finetuned from the Hunyuan 7B for interactive automatic theorem proving with LEAN4. To alleviate the data sparsity issue, we design a scalable framework to iterative synthesize data with low cost. Besides, guided tree search algorithms are designed to enable effective ``system 2 thinking`` of the prover. HunyuanProver achieves state-of-the-art (SOTA) performances on major benchmarks. Specifically, it achieves a pass of 68.4% on the miniF2F-test compared to 65.9%, the current SOTA results. It proves 4 IMO statements (imo_1960_p2, imo_1962_p2}, imo_1964_p2 and imo_1983_p6) in miniF2F-test. To benefit the community, we will open-source a dataset of 30k synthesized instances, where each instance contains the original question in natural language, the converted statement by autoformalization, and the proof by HunyuanProver.

cs.AI↗