arXiv subjects

Liyuan Zhang

Publications and source records attributed to Liyuan Zhang.

At least 19 recordsLinked to original sources

arXiv2026

AREAL-DTA: Dynamic Tree Attention for Efficient Reinforcement Learning of Large Language Models

Reinforcement learning (RL)-based post-training for large language models (LLMs) is computationally expensive, as it generates many rollout sequences that frequently share long token prefixes. Existing RL frameworks usually process these sequences independently during policy training, i.e., repeatedly recomputing identical prefixes in both the forward and backward passes of policy gradient computation, leading to substantial inefficiencies in computation resources and memory usage. Although prefix sharing naturally induces a tree structure over rollouts, packed tree-mask approaches scale poorly in RL settings. In this paper, we introduce AReaL-DTA, which efficiently exploits prefix sharing in RL training. AReaL-DTA employs a depth-first search (DFS)-based execution strategy that dynamically traverses the rollout prefix tree during both forward and backward computation, materializing only a single root-to-leaf path at a time. To further improve scalability, AReaL-DTA incorporates a load-balanced distributed batching mechanism that dynamically constructs and processes prefix trees across multiple GPUs. On $τ^2$-bench, AReaL-DTA improves training throughput by up to $8.31\times$ over dense training and up to $1.70\times$ over sparse training. Our code is available at https://github.com/areal-project/AReaL/tree/feat/dta.

cs.LG↗

arXiv2026

D^2SD: Accelerating Speculative Decoding with Dual Diffusion Draft Models

Speculative decoding accelerates autoregressive large language model inference by drafting multiple tokens and verifying them in a single target-model forward pass. Recent diffusion-based drafters generate an entire block of tokens in parallel but usually commit to a single draft sequence per verification: once the first mismatch occurs, all subsequent draft tokens are discarded, resulting in a limited acceptance rate. Naively batching more draft candidate sequences only introduces a marginal improvement, as redundant or poorly placed branches increase the cost of drafting and verification without proportionally increasing the number of accepted tokens. We propose D^2SD, a dual diffusion draft speculative decoding framework that organizes candidates into a confidence-guided prefix tree, where the first diffusion drafter generates a block along with per-position confidence scores that are used to identify the most likely rejection boundary and select the top-K prefix ranges for recovery; the second variable-prefix diffusion drafter re-anchors at each selected prefix and proposes alternative continuations in one batched pass; the resulting shared-prefix candidates are jointly verified via cascade attention. Empirically, D^2SD shows clear improvements over both the underlying diffusion approach and strong autoregressive speculative decoding baselines.

cs.DC↗

arXiv2026

Entropy-Guided GRVQ for Ultra-Low Bitrate Neural Speech Codec

Neural audio codec (NAC) is essential for reconstructing high-quality speech signals and generating discrete representations for downstream speech language models. However, ensuring accurate semantic modeling while maintaining high-fidelity reconstruction under ultra-low bitrate constraints remains challenging. We propose an entropy-guided group residual vector quantization (EG-GRVQ) for an ultra-low bitrate neural speech codec, which retains a semantic branch for linguistic information and incorporates an entropy-guided grouping strategy in the acoustic branch. Assuming that channel activations follow approximately Gaussian statistics, the variance of each channel can serve as a principled proxy for its information content. Based on this assumption, we partition the encoder output such that each group carries an equal share of the total information. This balanced allocation improves codebook efficiency and reduces redundancy. Trained on LibriTTS and VCTK, our model shows improvements in perceptual quality and intelligibility metrics under ultra-low bitrate conditions, with a focus on codec-level fidelity for communication-oriented scenarios.

eess.AS↗

arXiv2025

Generative Psycho-Lexical Approach for Constructing Value Systems in Large Language Models

Values are core drivers of individual and collective perception, cognition, and behavior. Value systems, such as Schwartz's Theory of Basic Human Values, delineate the hierarchy and interplay among these values, enabling cross-disciplinary investigations into decision-making and societal dynamics. Recently, the rise of Large Language Models (LLMs) has raised concerns regarding their elusive intrinsic values. Despite growing efforts in evaluating, understanding, and aligning LLM values, a psychologically grounded LLM value system remains underexplored. This study addresses the gap by introducing the Generative Psycho-Lexical Approach (GPLA), a scalable, adaptable, and theoretically informed method for constructing value systems. Leveraging GPLA, we propose a psychologically grounded five-factor value system tailored for LLMs. For systematic validation, we present three benchmarking tasks that integrate psychological principles with cutting-edge AI priorities. Our results reveal that the proposed value system meets standard psychological criteria, better captures LLM values, improves LLM safety prediction, and enhances LLM alignment, when compared to the canonical Schwartz's values.

cs.CL↗

arXiv2025

Accelerated Convolutive Transfer Function-Based Multichannel NMF Using Iterative Source Steering

Among numerous blind source separation (BSS) methods, convolutive transfer function-based multichannel non-negative matrix factorization (CTF-MNMF) has demonstrated strong performance in highly reverberant environments by modeling multi-frame correlations of delayed source signals. However, its practical deployment is hindered by the high computational cost associated with the iterative projection (IP) update rule, which requires matrix inversion for each source. To address this issue, we propose an efficient variant of CTF-MNMF that integrates iterative source steering (ISS), a matrix inversion-free update rule for separation filters. Experimental results show that the proposed method achieves comparable or superior separation performance to the original CTF-MNMF, while significantly reducing the computational complexity.

cs.SD↗

arXiv2025

Boosting Pathology Foundation Models via Few-shot Prompt-tuning for Rare Cancer Subtyping

Rare cancers comprise 20-25% of all malignancies but face major diagnostic challenges due to limited expert availability-especially in pediatric oncology, where they represent over 70% of cases. While pathology vision-language (VL) foundation models show promising zero-shot capabilities for common cancer subtyping, their clinical performance for rare cancers remains limited. Existing multi-instance learning (MIL) methods rely only on visual features, overlooking cross-modal knowledge and compromising interpretability critical for rare cancer diagnosis. To address this limitation, we propose PathPT, a novel framework that fully exploits the potential of vision-language pathology foundation models through spatially-aware visual aggregation and task-specific prompt tuning. Unlike conventional MIL, PathPT converts WSI-level supervision into fine-grained tile-level guidance by leveraging the zero-shot capabilities of VL models, thereby preserving localization on cancerous regions and enabling cross-modal reasoning through prompts aligned with histopathological semantics. We benchmark PathPT on eight rare cancer datasets(four adult and four pediatric) spanning 56 subtypes and 2,910 WSIs, as well as three common cancer datasets, evaluating four state-of-the-art VL models and four MIL frameworks under three few-shot settings. Results show that PathPT consistently delivers superior performance, achieving substantial gains in subtyping accuracy and cancerous region grounding ability. This work advances AI-assisted diagnosis for rare cancers, offering a scalable solution for improving subtyping accuracy in settings with limited access to specialized expertise.

cs.CV↗

arXiv2025

T$^\text{3}$SVFND: Towards an Evolving Fake News Detector for Emergencies with Test-time Training on Short Video Platforms

The existing methods for fake news videos detection may not be generalized, because there is a distribution shift between short video news of different events, and the performance of such techniques greatly drops if news records are coming from emergencies. We propose a new fake news videos detection framework (T$^3$SVFND) using Test-Time Training (TTT) to alleviate this limitation, enhancing the robustness of fake news videos detection. Specifically, we design a self-supervised auxiliary task based on Mask Language Modeling (MLM) that masks a certain percentage of words in text and predicts these masked words by combining contextual information from different modalities (audio and video). In the test-time training phase, the model adapts to the distribution of test data through auxiliary tasks. Extensive experiments on the public benchmark demonstrate the effectiveness of the proposed model, especially for the detection of emergency news.

cs.CV↗

arXiv2025

Enhanced anomalous Hall effect in the topological Kagome metal Cs(V$_{1-x}$Mn$_x$)$_3$Sb$_5$

As a fundamental physical phenomenon, achieving and controlling a large anomalous Hall effect (AHE) is crucial for advancing the understanding of topological physics and for developing applied technologies in spintronics. The recently discovered topological Kagome metal $A$V$_3$Sb$_5$ ($A =$ K, Rb, Cs)exhibits a significant AHE along with charge density wave (CDW) and superconductivity, providing an ideal platform to study the interactions between nontrivial band topology, CDW, and superconductivity. In this study, we systematically investigated the evolution of CDW, superconductivity, and AHE in electron (Mn)-doped Cs(V$_{1-x}$Mn$_x$)$_3$Sb$_5$ single crystals. The experimental results show that electron doping rapidly suppresses superconductivity, while the CDW order remains relatively robust. Meanwhile, a significantly enhanced AHE, with a maximum anomalous Hall conductivity (AHC) of ~25331 \Ohm ^{-1}\cm^{-1} and an anomalous Hall angle of 6.66% occurs at a relatively low doping level of $x = 0.03$. Based on the Tian-Ye-Jin (TYJ) scaling model, such a significant enhancement AHC is mainly dominated by the skew scattering. We speculated enhanced skew scattering between electrons and Mn originating from the strengthened spin-orbital coupling. Our finding provides important guidance for the design and development of transverse transport properties in topological Kagome materials.

cond-mat.str-el↗

arXiv2025

Geometry-Aware 3D Salient Object Detection Network

Point cloud salient object detection has attracted the attention of researchers in recent years. Since existing works do not fully utilize the geometry context of 3D objects, blurry boundaries are generated when segmenting objects with complex backgrounds. In this paper, we propose a geometry-aware 3D salient object detection network that explicitly clusters points into superpoints to enhance the geometric boundaries of objects, thereby segmenting complete objects with clear boundaries. Specifically, we first propose a simple yet effective superpoint partition module to cluster points into superpoints. In order to improve the quality of superpoints, we present a point cloud class-agnostic loss to learn discriminative point features for clustering superpoints from the object. After obtaining superpoints, we then propose a geometry enhancement module that utilizes superpoint-point attention to aggregate geometric information into point features for predicting the salient map of the object with clear boundaries. Extensive experiments show that our method achieves new state-of-the-art performance on the PCSOD dataset.

cs.CV↗

arXiv2025

SmartAgent: Chain-of-User-Thought for Embodied Personalized Agent in Cyber World

Recent advances in embodied agents with multimodal perception and reasoning capabilities based on large vision-language models (LVLMs), excel in autonomously interacting either real or cyber worlds, helping people make intelligent decisions in complex environments. However, the current works are normally optimized by golden action trajectories or ideal task-oriented solutions toward a definitive goal. This paradigm considers limited user-oriented factors, which could be the reason for their performance reduction in a wide range of personal assistant applications. To address this, we propose Chain-of-User-Thought (COUT), a novel embodied reasoning paradigm that takes a chain of thought from basic action thinking to explicit and implicit personalized preference thought to incorporate personalized factors into autonomous agent learning. To target COUT, we introduce SmartAgent, an agent framework perceiving cyber environments and reasoning personalized requirements as 1) interacting with GUI to access an item pool, 2) generating users' explicit requirements implied by previous actions, and 3) recommending items to fulfill users' implicit requirements. To demonstrate SmartAgent's capabilities, we also create a brand-new dataset SmartSpot that offers a full-stage personalized action-involved environment. To our best knowledge, our work is the first to formulate the COUT process, serving as a preliminary attempt towards embodied personalized agent learning. Our extensive experiments on SmartSpot illuminate SmartAgent's functionality among a series of embodied and personalized sub-tasks. We will release code and data upon paper notification at https://github.com/tsinghua-fib-lab/SmartAgent.

cs.AI↗

arXiv2025

Gradient Based Method for the Fusion of Lattice Quantizers

In practical applications, lattice quantizers leverage discrete lattice points to approximate arbitrary points in the lattice. An effective lattice quantizer significantly enhances both the accuracy and efficiency of these approximations. In the context of high-dimensional lattice quantization, previous work proposed utilizing low-dimensional optimal lattice quantizers and addressed the challenge of determining the optimal length ratio in orthogonal splicing. Notably, it was demonstrated that fixed length ratios and orthogonality yield suboptimal results when combining low-dimensional lattices. Building on this foundation, another approach employed gradient descent to identify optimal lattices, which inspired us to explore the use of neural networks to discover matrices that outperform those obtained from orthogonal splicing methods. We propose two novel approaches to tackle this problem: the Household Algorithm and the Matrix Exp Algorithm. Our results indicate that both the Household Algorithm and the Matrix Exp Algorithm achieve improvements in lattice quantizers across dimensions 13, 15, 17 to 19, 21, and 22. Moreover, the Matrix Exp Algorithm demonstrates superior efficacy in high-dimensional settings.

cs.LG↗

arXiv2025

Noise-Resilient Point-wise Anomaly Detection in Time Series Using Weak Segment Labels

Detecting anomalies in temporal data has gained significant attention across various real-world applications, aiming to identify unusual events and mitigate potential hazards. In practice, situations often involve a mix of segment-level labels (detected abnormal events with segments of time points) and unlabeled data (undetected events), while the ideal algorithmic outcome should be point-level predictions. Therefore, the huge label information gap between training data and targets makes the task challenging. In this study, we formulate the above imperfect information as noisy labels and propose NRdetector, a noise-resilient framework that incorporates confidence-based sample selection, robust segment-level learning, and data-centric point-level detection for multivariate time series anomaly detection. Particularly, to bridge the information gap between noisy segment-level labels and missing point-level labels, we develop a novel loss function that can effectively mitigate the label noise and consider the temporal features. It encourages the smoothness of consecutive points and the separability of points from segments with different labels. Extensive experiments on real-world multivariate time series datasets with 11 different evaluation metrics demonstrate that NRdetector consistently achieves robust results across multiple real-world datasets, outperforming various baselines adapted to operate in our setting.

cs.LG↗

arXiv2025

3D Focusing-and-Matching Network for Multi-Instance Point Cloud Registration

Multi-instance point cloud registration aims to estimate the pose of all instances of a model point cloud in the whole scene. Existing methods all adopt the strategy of first obtaining the global correspondence and then clustering to obtain the pose of each instance. However, due to the cluttered and occluded objects in the scene, it is difficult to obtain an accurate correspondence between the model point cloud and all instances in the scene. To this end, we propose a simple yet powerful 3D focusing-and-matching network for multi-instance point cloud registration by learning the multiple pair-wise point cloud registration. Specifically, we first present a 3D multi-object focusing module to locate the center of each object and generate object proposals. By using self-attention and cross-attention to associate the model point cloud with structurally similar objects, we can locate potential matching instances by regressing object centers. Then, we propose a 3D dual masking instance matching module to estimate the pose between the model point cloud and each object proposal. It performs instance mask and overlap mask masks to accurately predict the pair-wise correspondence. Extensive experiments on two public benchmarks, Scan2CAD and ROBI, show that our method achieves a new state-of-the-art performance on the multi-instance point cloud registration task. Code is available at https://github.com/zlynpu/3DFMNet.

cs.CV↗

arXiv2024

A Probabilistic Fluctuation based Membership Inference Attack for Diffusion Models

Membership Inference Attack (MIA) identifies whether a record exists in a machine learning model's training set by querying the model. MIAs on the classic classification models have been well-studied, and recent works have started to explore how to transplant MIA onto generative models. Our investigation indicates that existing MIAs designed for generative models mainly depend on the overfitting in target models. However, overfitting can be avoided by employing various regularization techniques, whereas existing MIAs demonstrate poor performance in practice. Unlike overfitting, memorization is essential for deep learning models to attain optimal performance, making it a more prevalent phenomenon. Memorization in generative models leads to an increasing trend in the probability distribution of generating records around the member record. Therefore, we propose a Probabilistic Fluctuation Assessing Membership Inference Attack (PFAMI), a black-box MIA that infers memberships by detecting these trends via analyzing the overall probabilistic fluctuations around given records. We conduct extensive experiments across multiple generative models and datasets, which demonstrate PFAMI can improve the attack success rate (ASR) by about 27.9% when compared with the best baseline.

cs.LG↗

arXiv2024

Learning with Dynamics: Autonomous Regulation of UAV Based Communication Networks with Dynamic UAV Crew

Unmanned Aerial Vehicle (UAV) based communication networks (UCNs) are a key component in future mobile networking. To handle the dynamic environments in UCNs, reinforcement learning (RL) has been a promising solution attributed to its strong capability of adaptive decision-making free of the environment models. However, most existing RL-based research focus on control strategy design assuming a fixed set of UAVs. Few works have investigated how UCNs should be adaptively regulated when the serving UAVs change dynamically. This article discusses RL-based strategy design for adaptive UCN regulation given a dynamic UAV set, addressing both reactive strategies in general UCNs and proactive strategies in solar-powered UCNs. An overview of the UCN and the RL framework is first provided. Potential research directions with key challenges and possible solutions are then elaborated. Some of our recent works are presented as case studies to inspire innovative ways to handle dynamic UAV crew with different RL algorithms.

eess.SY↗

arXiv2024

Progress of Inorganic Scintillators for Future HEP Experiments

The Caltech HEP Crystal Lab has been actively investigating novel inorganic scintillators along the following three directions. Fast and radiation hard inorganic scintillators to face the challenge of severe radiation environment expected by future HEP experiments at hadron colliders, such as the high luminosity LHC and FCC hh. Ultrafast inorganic scintillators to face the challenge of unprecedented event rate expected by future HEP experiments searching for rare decays, such as Mu2e II, and ultrafast time of flight system at hadron colliders. Cost effective inorganic scintillators for the homogeneous hadron calorimeter concept to face the challenge of both electromagnetic and jet mass resolutions required by the proposed Higgs factory. We report novel materials along all directions: LuAG:Ce ceramic fibers for the HL LHC, Lu2O3:Yb ceramic scintillators for ultrafast applications, and ABS:Ce and DSB:Ce glass scintillators for the proposed Higgs factory. The result of this investigation may also benefit nuclear physics experiments, GHz hard X ray imaging, medical imaging, and homeland security applications.

physics.ins-det↗

arXiv2024

Local excitation of kagome spin ice magnetism in HoAgGe seen by scanning tunneling microscopy

The kagome spin ice can host frustrated magnetic excitations by flipping its local spin. Under an inelastic tunneling condition, the tip in a scanning tunneling microscope can flip the local spin, and we apply this technique to kagome metal HoAgGe with a long-range ordered spin ice ground state. Away from defects, we discover a pair of pronounced dips in the local tunneling spectrum at symmetrical bias voltages with negative intensity values, serving as a striking inelastic tunneling signal. This signal disappears above the spin ice formation temperature and has a dependence on the magnetic fields, demonstrating its intimate relation with the spin ice magnetism. We provide a two-level spin-flip model to explain the tunneling dips considering the spin ice magnetism under spin-orbit coupling. Our results uncover a local emergent excitation of spin ice magnetism in a kagome metal, suggesting that local electrical field induced spin flip climbs over a barrier caused by spin-orbital locking.

cond-mat.str-el↗

arXiv2024

Physics-informed Meta-instrument for eXperiments (PiMiX) with applications to fusion energy

Data-driven methods (DDMs), such as deep neural networks, offer a generic approach to integrated data analysis (IDA), integrated diagnostic-to-control (IDC) workflows through data fusion (DF), which includes multi-instrument data fusion (MIDF), multi-experiment data fusion (MXDF), and simulation-experiment data fusion (SXDF). These features make DDMs attractive to nuclear fusion energy and power plant applications, leveraging accelerated workflows through machine learning and artificial intelligence. Here we describe Physics-informed Meta-instrument for eXperiments (PiMiX) that integrates X-ray (including high-energy photons such as $γ$-rays from nuclear fusion), neutron and others (such as proton radiography) measurements for nuclear fusion. PiMiX solves multi-domain high-dimensional optimization problems and integrates multi-modal measurements with multiphysics modeling through neural networks. Super-resolution for neutron detection and energy resolved X-ray detection have been demonstrated. Multi-modal measurements through MIDF can extract more information than individual or uni-modal measurements alone. Further optimization schemes through DF are possible towards empirical fusion scaling laws discovery and new fusion reactor designs.

physics.data-an↗