arXiv subjects

Xin Liu

Publications and source records attributed to Xin Liu.

At least 271 records · Page 15Linked to original sources

arXiv2025

WeaveRec: An LLM-Based Cross-Domain Sequential Recommendation Framework with Model Merging

Cross-Domain Sequential Recommendation (CDSR) seeks to improve user preference modeling by transferring knowledge from multiple domains. Despite the progress made in CDSR, most existing methods rely on overlapping users or items to establish cross-domain correlations-a requirement that rarely holds in real-world settings. The advent of large language models (LLM) and model-merging techniques appears to overcome this limitation by unifying multi-domain data without explicit overlaps. Yet, our empirical study shows that naively training an LLM on combined domains-or simply merging several domain-specific LLMs-often degrades performance relative to a model trained solely on the target domain. To address these challenges, we first experimentally investigate the cause of suboptimal performance in LLM-based cross-domain recommendation and model merging. Building on these insights, we introduce WeaveRec, which cross-trains multiple LoRA modules with source and target domain data in a weaving fashion, and fuses them via model merging. WeaveRec can be extended to multi-source domain scenarios and notably does not introduce additional inference-time cost in terms of latency or memory. Furthermore, we provide a theoretical guarantee that WeaveRec can reduce the upper bound of the expected error in the target domain. Extensive experiments on single-source, multi-source, and cross-platform cross-domain recommendation scenarios validate that WeaveRec effectively mitigates performance degradation and consistently outperforms baseline approaches in real-world recommendation tasks.

cs.IR↗

arXiv2025

Audio-Visual Speech Enhancement In Complex Scenarios With Separation And Dereverberation Joint Modeling

Audio-visual speech enhancement (AVSE) is a task that uses visual auxiliary information to extract a target speaker's speech from mixed audio. In real-world scenarios, there often exist complex acoustic environments, accompanied by various interfering sounds and reverberation. Most previous methods struggle to cope with such complex conditions, resulting in poor perceptual quality of the extracted speech. In this paper, we propose an effective AVSE system that performs well in complex acoustic environments. Specifically, we design a "separation before dereverberation" pipeline that can be extended to other AVSE networks. The 4th COGMHEAR Audio-Visual Speech Enhancement Challenge (AVSEC) aims to explore new approaches to speech processing in multimodal complex environments. We validated the performance of our system in AVSEC-4: we achieved excellent results in the three objective metrics on the competition leaderboard, and ultimately secured first place in the human subjective listening test.

cs.SD↗

arXiv2025

Emerging axion detection in artificial magnetoelectric materials

Axions are considered a key component of dark matter, characterized by very weak couplings to fermions and Chern-Simons couplings to gauge fields. We propose a novel detection mechanism based on symmetry-breaking magnetoelectric materials with a linear axionic coupling between magnetization and ferroelectric polarization. The focus is on a strain gradient Sr2IrO4 film, where the breaking of space-inversion symmetry results in an emergent polar phase and an out-of-plane magnetic moment, exhibiting a flexomagnetoelectric effect. In this material, the linear P||M enables a direct coupling between the external axion field and the intrinsic axion-like field, which amplifies the weak electromagnetic signals induced by axions, paving the way for pioneering axion detection. In contrast to conventional detection techniques, this mechanism is expected to enhance the sensitivity of the axion-electron and axion-photon coupling, providing a novel platform for axion detection and advancing the study of dark matter through the magnetoelectric effect.

hep-ph↗

arXiv2025

Partial Envelope for Optimization Problem with Nonconvex Constraints

In this paper, we consider the nonlinear constrained optimization problem (NCP) with constraint set $\{x \in \mathcal{X}: c(x) = 0\}$, where $\mathcal{X}$ is a closed convex subset of $\mathbb{R}^n$. Building upon the forward-backward envelope framework for optimization over $\mathcal{X}$, we propose a forward-backward semi-envelope (FBSE) approach for solving (NCP). In the proposed semi-envelope approach, we eliminate the constraint $x \in \mathcal{X}$ through a specifically designed envelope scheme while preserving the constraint $x \in \mathcal{M} := \{x \in \mathbb{R}^n: c(x) = 0\}$. We establish that the forward-backward semi-envelope for (NCP) is well-defined and locally Lipschitz smooth over a neighborhood of $\mathcal{M}$. Furthermore, we prove that (NCP) and its corresponding forward-backward semi-envelope have the same first-order stationary points within a neighborhood of $\mathcal{X} \cap \mathcal{M}$. Consequently, our proposed forward-backward semi-envelope approach enables direct application of optimization methods over $\mathcal{M}$ while inheriting their convergence properties for (NCP). Additionally, we develop an inexact projected gradient descent method for minimizing the forward-backward semi-envelope over $\mathcal{M}$ and establish its global convergence. Preliminary numerical experiments demonstrate the practical efficiency and potential of our proposed approach.

math.OC↗

arXiv2025

DEEMO: De-identity Multimodal Emotion Recognition and Reasoning

Emotion understanding is a critical yet challenging task. Most existing approaches rely heavily on identity-sensitive information, such as facial expressions and speech, which raises concerns about personal privacy. To address this, we introduce the De-identity Multimodal Emotion Recognition and Reasoning (DEEMO), a novel task designed to enable emotion understanding using de-identified video and audio inputs. The DEEMO dataset consists of two subsets: DEEMO-NFBL, which includes rich annotations of Non-Facial Body Language (NFBL), and DEEMO-MER, an instruction dataset for Multimodal Emotion Recognition and Reasoning using identity-free cues. This design supports emotion understanding without compromising identity privacy. In addition, we propose DEEMO-LLaMA, a Multimodal Large Language Model (MLLM) that integrates de-identified audio, video, and textual information to enhance both emotion recognition and reasoning. Extensive experiments show that DEEMO-LLaMA achieves state-of-the-art performance on both tasks, outperforming existing MLLMs by a significant margin, achieving 74.49% accuracy and 74.45% F1-score in de-identity emotion recognition, and 6.20 clue overlap and 7.66 label overlap in de-identity emotion reasoning. Our work contributes to ethical AI by advancing privacy-preserving emotion understanding and promoting responsible affective computing.

cs.CV↗

arXiv2025

Robust LLM Training Infrastructure at ByteDance

The training scale of large language models (LLMs) has reached tens of thousands of GPUs and is still continuously expanding, enabling faster learning of larger models. Accompanying the expansion of the resource scale is the prevalence of failures (CUDA error, NaN values, job hang, etc.), which poses significant challenges to training stability. Any large-scale LLM training infrastructure should strive for minimal training interruption, efficient fault diagnosis, and effective failure tolerance to enable highly efficient continuous training. This paper presents ByteRobust, a large-scale GPU infrastructure management system tailored for robust and stable training of LLMs. It exploits the uniqueness of LLM training process and gives top priorities to detecting and recovering failures in a routine manner. Leveraging parallelisms and characteristics of LLM training, ByteRobust enables high-capacity fault tolerance, prompt fault demarcation, and localization with an effective data-driven approach, comprehensively ensuring continuous and efficient training of LLM tasks. ByteRobust is deployed on a production GPU platform and achieves 97% ETTR for a three-month training job on 9,600 GPUs.

cs.LG↗

arXiv2025

MegaScale-MoE: Large-Scale Communication-Efficient Training of Mixture-of-Experts Models in Production

We present MegaScale-MoE, a production system tailored for the efficient training of large-scale mixture-of-experts (MoE) models. MoE emerges as a promising architecture to scale large language models (LLMs) to unprecedented sizes, thereby enhancing model performance. However, existing MoE training systems experience a degradation in training efficiency, exacerbated by the escalating scale of MoE models and the continuous evolution of hardware. Recognizing the pivotal role of efficient communication in enhancing MoE training, MegaScale-MoE customizes communication-efficient parallelism strategies for attention and FFNs in each MoE layer and adopts a holistic approach to overlap communication with computation at both inter- and intra-operator levels. Additionally, MegaScale-MoE applies communication compression with adjusted communication patterns to lower precision, further improving training efficiency. When training a 352B MoE model on 1,440 NVIDIA Hopper GPUs, MegaScale-MoE achieves a training throughput of 1.41M tokens/s, improving the efficiency by 1.88$\times$ compared to Megatron-LM. We share our operational experience in accelerating MoE training and hope that by offering our insights in system design, this work will motivate future research in MoE systems.

cs.LG↗

arXiv2025

MSF-Mamba: Motion-aware State Fusion Mamba for Efficient Micro-Gesture Recognition

Micro-gesture recognition (MGR) targets the identification of subtle and fine-grained human motions and requires accurate modeling of both long-range and local spatiotemporal dependencies. While CNNs are effective at capturing local patterns, they struggle with long-range dependencies due to their limited receptive fields. Transformer-based models address this limitation through self-attention mechanisms but suffer from high computational costs. Recently, Mamba has shown promise as an efficient model, leveraging state space models (SSMs) to enable linear-time processing However, directly applying the vanilla Mamba to MGR may not be optimal. This is because Mamba processes inputs as 1D sequences, with state updates relying solely on the previous state, and thus lacks the ability to model local spatiotemporal dependencies. In addition, previous methods lack a design of motion-awareness, which is crucial in MGR. To overcome these limitations, we propose motion-aware state fusion mamba (MSF-Mamba), which enhances Mamba with local spatiotemporal modeling by fusing local contextual neighboring states. Our design introduces a motion-aware state fusion module based on central frame difference (CFD). Furthermore, a multiscale version named MSF-Mamba+ has been proposed. Specifically, MSF-Mamba supports multiscale motion-aware state fusion, as well as an adaptive scale weighting module that dynamically weighs the fused states across different scales. These enhancements explicitly address the limitations of vanilla Mamba by enabling motion-aware local spatiotemporal modeling, allowing MSF-Mamba and MSF-Mamba to effectively capture subtle motion cues for MGR. Experiments on two public MGR datasets demonstrate that even the lightweight version, namely, MSF-Mamba, achieves SoTA performance, outperforming existing CNN-, Transformer-, and SSM-based models while maintaining high efficiency.

cs.CV↗

arXiv2025

Laminar: A Scalable Asynchronous RL Post-Training Framework

Reinforcement learning (RL) post-training for Large Language Models (LLMs) is now scaling to large clusters and running for extended durations to enhance model reasoning performance. However, the scalability of existing RL frameworks is limited, as extreme long-tail skewness in RL trajectory generation causes severe GPU underutilization. Current asynchronous RL systems attempt to mitigate this, but they rely on global weight synchronization between the actor and all rollouts, which creates a rigid model update schedule. This global synchronization is ill-suited for the highly skewed and evolving distribution of trajectory generation latency in RL training, crippling training efficiency. Our key insight is that efficient scaling requires breaking this lockstep through trajectory-level asynchrony, which generates and consumes each trajectory independently. We propose Laminar, a scalable and robust RL post-training system built on a fully decoupled architecture. First, we replace global updates with a tier of relay workers acting as a distributed parameter service. This enables asynchronous and fine-grained weight synchronization, allowing rollouts to pull the latest weight anytime without stalling the actor's training loop. Second, a dynamic repack mechanism consolidates long-tail trajectories onto a few dedicated rollouts, maximizing generation throughput. The fully decoupled design also isolates failures, ensuring robustness for long-running jobs. Our evaluation on a 1024-GPU cluster shows that Laminar achieves up to 5.48$\times$ training throughput speedup over state-of-the-art systems, while reducing model convergence time.

cs.LG↗

arXiv2025

Ultrafast surface melting of orbital order in La0.5Sr1.5MnO4

Understanding how light modifies long-range order in quantum materials is key to improving our ability to control functionality. However, this is challenging if the response is heterogeneous. Here we address the most common form of light-induced heterogeneity, surface melting, and measure the dynamics of orbital order in the layered manganite, La0.5Sr1.5MnO4. We isolate the surface dynamics from the bulk by measuring the orbital truncation rod as well as orbital Bragg peak. After photoexcitation, the orbital Bragg peak shows an unusual narrowing, which suggests an increase in the correlation length in the probed volume. In contrast, the correlation length at the surface decreases. These differences can be reconciled if the material is heterogeneous, and light melts a less ordered surface. By isolating the surface response, we determine that the loss of long-range order is an incoherent process, which is likely accompanied by the formation of local polarons.

cond-mat.str-el↗

arXiv2025

Domain Generalization in-the-Wild: Disentangling Classification from Domain-Aware Representations

Evaluating domain generalization (DG) for foundational models like CLIP is challenging, as web-scale pretraining data potentially covers many existing benchmarks. Consequently, current DG evaluation may neither be sufficiently challenging nor adequately test genuinely unseen data scenarios. To better assess the performance of CLIP on DG in-the-wild, a scenario where CLIP encounters challenging unseen data, we consider two approaches: (1) evaluating on 33 diverse datasets with quantified out-of-distribution (OOD) scores after fine-tuning CLIP on ImageNet, and (2) using unlearning to make CLIP `forget' some domains as an approximation. We observe that CLIP's performance deteriorates significantly on more OOD datasets. To address this, we present CLIP-DCA (Disentangling Classification from enhanced domain Aware representations). Our approach is motivated by the observation that while standard domain invariance losses aim to make representations domain-invariant, this can be harmful to foundation models by forcing the discarding of domain-aware representations beneficial for generalization. We instead hypothesize that enhancing domain awareness is a prerequisite for effective domain-invariant classification in foundation models. CLIP-DCA identifies and enhances domain awareness within CLIP's encoders using a separate domain head and synthetically generated diverse domain data. Simultaneously, it encourages domain-invariant classification through disentanglement from the domain features. CLIP-DCA shows significant improvements within this challenging evaluation compared to existing methods, particularly on datasets that are more OOD.

cs.CV↗

arXiv2025

FastUMI-100K: Advancing Data-driven Robotic Manipulation with a Large-scale UMI-style Dataset

Data-driven robotic manipulation learning depends on large-scale, high-quality expert demonstration datasets. However, existing datasets, which primarily rely on human teleoperated robot collection, are limited in terms of scalability, trajectory smoothness, and applicability across different robotic embodiments in real-world environments. In this paper, we present FastUMI-100K, a large-scale UMI-style multimodal demonstration dataset, designed to overcome these limitations and meet the growing complexity of real-world manipulation tasks. Collected by FastUMI, a novel robotic system featuring a modular, hardware-decoupled mechanical design and an integrated lightweight tracking system, FastUMI-100K offers a more scalable, flexible, and adaptable solution to fulfill the diverse requirements of real-world robot demonstration data. Specifically, FastUMI-100K contains over 100K+ demonstration trajectories collected across representative household environments, covering 54 tasks and hundreds of object types. Our dataset integrates multimodal streams, including end-effector states, multi-view wrist-mounted fisheye images and textual annotations. Each trajectory has a length ranging from 120 to 500 frames. Experimental results demonstrate that FastUMI-100K enables high policy success rates across various baseline algorithms, confirming its robustness, adaptability, and real-world applicability for solving complex, dynamic manipulation challenges. The source code and dataset will be released in this link https://github.com/MrKeee/FastUMI-100K.

cs.RO↗

arXiv2025

ZSMerge: Zero-Shot KV Cache Compression for Memory-Efficient Long-Context LLMs

The linear growth of key-value (KV) cache memory and quadratic computational in attention mechanisms complexity pose significant bottlenecks for large language models (LLMs) in long-context processing. While existing KV cache optimization methods address these challenges through token pruning or feature merging, they often incur irreversible information loss or require costly parameter retraining. To this end, we propose ZSMerge, a dynamic KV cache compression framework designed for efficient cache management, featuring three key operations: (1) fine-grained memory allocation guided by multi-dimensional token importance metrics at head-level granularity, (2) a residual merging mechanism that preserves critical context through compensated attention scoring, and (3) a zero-shot adaptation mechanism compatible with diverse LLM architectures without requiring retraining. ZSMerge significantly enhances memory efficiency and inference speed with negligible performance degradation across LLMs. When applied to LLaMA2-7B, it demonstrates a 20:1 compression ratio for key-value cache retention (reducing memory footprint to 5\% of baseline) while sustaining comparable generation quality, coupled with triple throughput gains at extreme 54k-token contexts that eliminate out-of-memory failures. The code is available at https://github.com/SusCom-Lab/ZSMerge.

cs.CL↗

arXiv2025

MedAgentGym: A Scalable Agentic Training Environment for Code-Centric Reasoning in Biomedical Data Science

We introduce MedAgentGym, a scalable and interactive training environment designed to enhance coding-based biomedical reasoning capabilities in large language model (LLM) agents. MedAgentGym comprises 72,413 task instances across 129 categories derived from 12 authentic real-world biomedical scenarios. Tasks are encapsulated within executable sandbox environments, each featuring detailed task specifications, interactive feedback mechanisms, verifiable ground truth annotations, and scalable training trajectory generation. Extensive benchmarking of 29 LLMs reveals substantial performance disparities in biomedical data science between commercial and open-source LLMs. Leveraging efficient multi-threaded and multi-turn trajectory sampling in MedAgentGym, Med-Copilot achieves performance gains of +43.02% and +45.28% from offline and online reinforcement learning, respectively, demonstrating MedAgentGym as an effective training ground while establishing itself as a cost-effective, privacy-preserving alternative competitive with proprietary LLMs (gpt-4o). By offering a unified execution environment with a comprehensive benchmark and accessible, extensible training resources, MedAgentGym delivers an integrated platform to develop LLM-based coding assistants for advanced biomedical data science.

cs.CL↗

arXiv2025

Large superconducting diode effect driven by edge states in twisted iron-chalcogenide Josephson junctions

The superconducting diode effect (SDE)-the unidirectional, dissipationless flow of supercurrent-is a critical element for future superconducting electronics. Achieving high efficiency under zero magnetic field is a key requirement. The Josephson junction constitutes a versatile SDE platform for exploiting quantum materials that exhibit ferromagnetism, topology, or unconventional superconductivity. However, a single two-dimensional material system that inherently offers these properties and allows for precise interface engineering, such as twisting, remains elusive. Here we report a record-high, field-free diode efficiency of ~30% in twist van der Waals Josephson heterostructures of the sign-change iron-chalcogenide superconductor FeTe0.55Se0.45 and the conventional transition-metal dichalcogenide superconductor 2H-NbSe2. The diode response shows a striking twist-angle dependence: the efficiency peaks at crystallographic alignment and collapses with a small misorientation of ~7 deg. Importantly, the twist-angle evolution of superconducting interference measurements reveals that efficient nonreciprocity arises from asymmetric edge supercurrents, whereas bulk transport suppresses the effect. These findings establish edge states as the driving mechanism of the unconventional SDE, linking it to exotic pairing and topology in multiband iron-based superconductors. Our findings reveal intricate physics involving novel pairing symmetry, magnetism, and topology in the multiband iron-based superconductor, and offer a new route to high-performance superconducting diodes.

cond-mat.supr-con↗

arXiv2025

Observation of non-Hermitian topology in cold Rydberg quantum gases

The pursuit of topological phenomena in non-Hermitian systems has unveiled new physics beyond the conventional Hermitian paradigm, yet their realization in interacting many-body platforms remains a critical challenge. Exploring this interplay is essential to understand how strong interactions and dissipation collectively shape topological phases in open quantum systems. Here, we experimentally demonstrate non-Hermitian spectra topology in a dissipative Rydberg atomic gas and characterize parameters-dependent winding numbers. By increasing the interaction strength, the system evolves from Hermitian to non-Hermitian regime, accompanying emergence of trajectory loop in the complex energy plane. As the scanning time is varied, the spectra topology becomes twisted in the complex energy plane manifesting as a topology phase transition with the sign winding number changed. When preparing the system in different initial states, we can access a nontrivial fractional phase within a parameter space that globally possesses an integer winding. Furthermore, by changing the scanning direction, we observe the differentiated loops, revealing the breaking of chirality symmetry. This work establishes cold Rydberg gases as a versatile platform for exploring the rich interplay between non-Hermitian topology, strong interactions, and dissipative quantum dynamics.

cond-mat.quant-gas↗

arXiv2025

To Remember, To Adapt, To Preempt: A Stable Continual Test-Time Adaptation Framework for Remote Physiological Measurement in Dynamic Domain Shifts

Remote photoplethysmography (rPPG) aims to extract non-contact physiological signals from facial videos and has shown great potential. However, existing rPPG approaches struggle to bridge the gap between source and target domains. Recent test-time adaptation (TTA) solutions typically optimize rPPG model for the incoming test videos using self-training loss under an unrealistic assumption that the target domain remains stationary. However, time-varying factors like weather and lighting in dynamic environments often cause continual domain shifts. The erroneous gradients accumulation from these shifts may corrupt the model's key parameters for physiological information, leading to catastrophic forgetting. Therefore, We propose a physiology-related parameters freezing strategy to retain such knowledge. It isolates physiology-related and domain-related parameters by assessing the model's uncertainty to current domain and freezes the physiology-related parameters during adaptation to prevent catastrophic forgetting. Moreover, the dynamic domain shifts with various non-physiological characteristics may lead to conflicting optimization objectives during TTA, which is manifested as the over-adapted model losing its adaptability to future domains. To fix over-adaptation, we propose a preemptive gradient modification strategy. It preemptively adapts to future domains and uses the acquired gradients to modify current adaptation, thereby preserving the model's adaptability. In summary, we propose a stable continual test-time adaptation (CTTA) framework for rPPG measurement, called \textbf{PhysRAP}, which \textbf{R}emembers the past, \textbf{A}dapts to the present, and \textbf{P}reempts the future. Extensive experiments show its state-of-the-art performance, especially in domain shifts. The code is available at https://github.com/xjtucsy/PhysRAP.

q-bio.QM↗

arXiv2025

Bidirectional Uncertainty-Aware Region Learning for Semi-Supervised Medical Image Segmentation

In semi-supervised medical image segmentation, the poor quality of unlabeled data and the uncertainty in the model's predictions lead to models that inevitably produce erroneous pseudo-labels. These errors accumulate throughout model training, thereby weakening the model's performance. We found that these erroneous pseudo-labels are typically concentrated in high-uncertainty regions. Traditional methods improve performance by directly discarding pseudo-labels in these regions, which can also result in neglecting potentially valuable training data. To alleviate this problem, we propose a bidirectional uncertainty-aware region learning strategy to fully utilize the precise supervision provided by labeled data and stabilize the training of unlabeled data. Specifically, in the training labeled data, we focus on high-uncertainty regions, using precise label information to guide the model's learning in potentially uncontrollable areas. Meanwhile, in the training of unlabeled data, we concentrate on low-uncertainty regions to reduce the interference of erroneous pseudo-labels on the model. Through this bidirectional learning strategy, the model's overall performance has significantly improved. Extensive experiments show that our proposed method achieves significant performance improvement on different medical image segmentation tasks.

cs.CV↗