papers

Publications (64)

cs.CL2025

Improved Personalized Headline Generation via Denoising Fake Interests from Implicit Feedback

Kejin Liu, Junhong Lian, Xiang Ao +5

Accurate personalized headline generation hinges on precisely capturing user interests from historical behaviors. However, existing methods neglect personalized-irrelevant click no…

physics.optics2026

Non-line-of-sight imaging with arbitrary relay surface geometries via 3D Gaussian Transient Rendering

Yi Wang, Ziyu Zhan, Yuran Wang +5

Imaging objects hidden outside the direct line of sight expands the effective field of view and is critical for applications such as autonomous driving and robotic perception. Desp…

cs.LG2024

Ultra-imbalanced classification guided by statistical information

Yin Jin, Ningtao Wang, Ruofan Wu +3

Imbalanced data are frequently encountered in real-world classification tasks. Previous works on imbalanced learning mostly focused on learning with a minority class of few samples…

cs.CL2025

ALPS: Attention Localization and Pruning Strategy for Efficient Alignment of Large Language Models

Hao Chen, Haoze Li, Zhiqing Xiao +6

Aligning general-purpose large language models (LLMs) to downstream tasks often incurs significant training adjustment costs. Prior research has explored various avenues to enhance…

physics.optics2024

Photonic Landau levels in a high-dimensional frequency-degenerate cavity

Jing Pan, Zhaoyang Wang, Yuan Meng +3

Topological orders emerge in both microscopic quantum dynamics and macroscopic materials as a fundamental principle to characterize intricate properties in nature with vital signif…

cs.LG2026

Training deep physical neural networks with local physical information bottleneck

Hao Wang, Ziao Wang, Xiangpeng Liang +8

Deep learning has revolutionized modern society but faces growing energy and latency constraints. Deep physical neural networks (PNNs) are interconnected computing systems that dir…

physics.optics2018

Sub-MHz Self-Q-switching in Nd:LuAG Laser

Guangju Zhang, Xing Fu, Yijie Shen +1

A compact pulsed Nd:LuAG laser at 1064 nm based on the self-Q-switching technique is reported, having the output power as high as 6.61 W at the incident pump power of 21.32 W, corr…

math.CA2015

Wavelet Characterizations of the Atomic Hardy Space on Spaces of Homogeneous Type

Xing Fu, Dachun Yang

Let be a metric measure space of homogeneous type in the sense of R. R. Coifman and G. Weiss and be the atomic Hardy space. Via o…

math.AP2026

Wolff potential estimates for elliptic obstacle problems with generalized Orlicz growth

Qi Xiong, Xing Fu

The paper studies elliptic obstacle problems with generalized Orlicz growth and measure data, proving existence of solutions in Musielak‑Orlicz spaces and obtaining gradient estima…

#elliptic obstacle problems#generalized orlicz growth#wolff potentials#regularity theory
cs.LG2025

Instruction-aware User Embedding via Synergistic Language and Representation Modeling

Ziyi Gao, Yike Xu, Jiahao Yuan +9

User representation modeling has become increasingly crucial for personalized applications, yet existing approaches struggle with generalizability across domains and sensitivity to…

physics.optics2021

Coherent ray-wave structured light based on (helical) Ince-Gaussian modes

Zhaoyang Wang, Yijie Shen, Qiang Liu +1

The topological evolution of classic eigenmodes including Hermite-Laguerre-Gaussian and (helical) InceGaussian modes is exploited to construct coherent state modes, which unifies t…

physics.optics2026

1.1 kW, 100 Hz room-temperature diode-pumped nanosecond laser by water immersion cooling

Xinxing Lei, Suyang Wang, Zichao Wang +3

We report a room-temperature diode-pumped solid-state laser by water immersion cooling, which delivers a pulse energy of 11 J at the repetition rate of 100 Hz and the pulse duratio…

physics.optics2021

Divergence-degenerated spatial multiplexing towards ultrahigh capacity, low bit-error-rate optical communications

Zhensong Wan, Yijie Shen, Zhaoyang Wang +3

Spatial mode (de)multiplexing of orbital angular momentum (OAM) beams is a promising solution to address future bandwidth issues, but the rapidly increasing divergence with the mod…

cs.CR2024

Protecting Split Learning by Potential Energy Loss

Fei Zheng, Chaochao Chen, Lingjuan Lyu +5

As a practical privacy-preserving learning method, split learning has drawn much attention in academia and industry. However, its security is constantly being questioned since the…

math.CA2015

Products of Functions in and via Wavelets over Spaces of Homogeneous Type

Xing Fu, Dachun Yang, Yiyu Liang

Let be a metric measure space of homogeneous type in the sense of R. R. Coifman and G. Weiss and be the atomic Hardy space. Via o…

eess.IV2020

A One-Shot Learning Framework for Assessment of Fibrillar Collagen from Second Harmonic Generation Images of an Infarcted Myocardium

Qun Liu, Supratik Mukhopadhyay, Maria Ximena Bastidas Rodriguez +4

Myocardial infarction (MI) is a scientific term that refers to heart attack. In this study, we infer highly relevant second harmonic generation (SHG) cues from collagen fibers exhi…

math.DS2015

Multi-Resolution Dynamic Mode Decomposition

J. Nathan Kutz, Xing Fu, Steven L. Brunton

We demonstrate that the integration of the recently developed dynamic mode decomposition (DMD) with a multi-resolution analysis allows for a decomposition method capable of robustl…

physics.optics2023

Frequency-astigmatism asymmetric nonlinear conversion of structured light lasers

Jing Pan, Hao Wang, Zijian Shi +3

Nonlinear optics of structured light has recently delivered intriguing fundamental physical phenomena in light-matter interactions and advanced applications from classical imaging…

cs.AI2026

ConMem: Contribution-Aware Memory for Long-Horizon Manufacturing Inspection Logs

Bingchen Liu, Yuanyuan Fang, Lei Liu +6

ConMem is a memory framework that selects and retains the most diagnostically valuable segments of long‑horizon manufacturing inspection logs for LLM‑based analysis, using Shapley‑…

#manufacturing inspection#long-horizon reasoning#memory management#contribution estimation
cs.LG2023

Self-supervision meets kernel graph neural models: From architecture to augmentations

Jiawang Dan, Ruofan Wu, Yunpeng Liu +8

Graph representation learning has now become the de facto standard when handling graph-structured data, with the framework of message-passing graph neural networks (MPNN) being the…

physics.optics2020

High-dimensional classically entangled light from a laser

Yijie Shen, Isaac Nape, Xilin Yang +4

Vectorially structured light has emerged as an enabling tool in many diverse applications, from communication to imaging, exploiting quantum-like correlations courtesy of a non-sep…

physics.optics2022

Particle manipulation behind turbid medium based on intensity transmission matrix

Kaige Liu, Hengkang Zhang, Shanshan Du +4

Optical tweezers can manipulate tiny particles. However, the distortion caused by the scattering medium restricts the applications of optical tweezers. Wavefront shaping techniques…

eess.IV2022

Non-line-of-sight imaging with arbitrary illumination and detection pattern

Xintong Liu, Jianyu Wang, Leping Xiao +3

Non-line-of-sight (NLOS) imaging aims at reconstructing targets obscured from the direct line of sight. Existing NLOS imaging algorithms require dense measurements at rectangular g…

cs.CL2026

OP-Bench: Benchmarking Over-Personalization for Memory-Augmented Personalized Conversational Agents

Yulin Hu, Zimo Long, Jiahe Guo +5

Memory-augmented conversational agents enable personalized interactions using long-term user memory and have gained substantial traction. However, existing benchmarks primarily foc…

cs.CV2026

Text-Driven Emotionally Continuous Talking Face Generation

Hao Yang, Yanyan Zhao, Tian Zheng +7

Talking Face Generation (TFG) strives to create realistic and emotionally expressive digital faces. While previous TFG works have mastered the creation of naturalistic facial movem…

cs.CL2025

Chinese ModernBERT with Whole-Word Masking

Zeyu Zhao, Ningtao Wang, Xing Fu +1

Encoder-only Transformers have advanced along three axes -- architecture, data, and systems -- yielding Pareto gains in accuracy, speed, and memory efficiency. Yet these improvemen…

physics.optics2021

Classical structured light analogy of quantum squeezed state

Zhaoyang Wang, Ziyu Zhan, Xing Fu +1

Much of the richness in nature arises due to the connection between classical and quantum mechanics. In advanced science, the tools of quantum mechanics was not only applied in mic…

cs.CL2026

How Do Decoder-Only LLMs Perceive Users? Rethinking Attention Masking for User Representation Learning

Jiahao Yuan, Yike Xu, Jinyong Wen +8

Decoder-only large language models are increasingly used as behavioral encoders for user representation learning, yet the impact of attention masking on the quality of user embeddi…

cs.CV2025

Fast and Memory-efficient Non-line-of-sight Imaging with Quasi-Fresnel Transform

Yijun Wei, Jianyu Wang, Leping Xiao +3

Non-line-of-sight (NLOS) imaging seeks to reconstruct hidden objects by analyzing reflections from intermediary surfaces. Existing methods typically model both the measurement data…

physics.optics2022

Structured light analogy of squeezed state

Zhaoyang Wang, Ziyu Zhan, Anton N. Vetlugin +3

Control of structured light is of great importance to explore fundamental physical effects and extend practical scientific applications, which has been advanced by accepting method…

cs.CL2026

ENPMR-Bench: Benchmarking Proactive Memory Retrieval for Emotional Support Agents

Xing Fu, Yulin Hu, Mengtong Ji +5

Memory-augmented language agents are increasingly deployed in affective applications such as emotional support, where understanding and responding to users' latent emotional needs…

cs.LG2024

Estimating Conditional Average Treatment Effects via Sufficient Representation Learning

Pengfei Shi, Wei Zhong, Xinyu Zhang +4

Estimating the conditional average treatment effects (CATE) is very important in causal inference and has a wide range of applications across many fields. In the estimation process…

cs.LG2023

Differentially Private Learning with Per-Sample Adaptive Clipping

Tianyu Xia, Shuheng Shen, Su Yao +4

Privacy in AI remains a topic that draws attention from researchers and the general public in recent years. As one way to implement privacy-preserving AI, differentially private le…

physics.optics2018

Polygonal vortex beams in quasi-frequency-degenerate states

Yijie Shen, Zhensong Wan, Yuan Meng +2

We originally demonstrate the vortex beams with patterns of closed polygons [namely polygonal vortex beams (PVBs)] generated by a quasi-frequency-degenerate (QFD) Yb:CALGO laser re…

physics.optics2022

3D inhomogeneous self-accelerating beams

Jing Pan, Hao Wang, Yijie Shen +2

We propose and generate a new class of structured light fulfilling quantum-like coherent states based on a set of circular Airy vortex modes. Such coherent-state wave packets posse…

math.CA2014

Generalized Fractional Integrals and Their Commutators over Non-homogeneous Metric Measure Spaces

Xing Fu, Dachun Yang, Wen Yuan

Let be a metric measure space satisfying both the upper doubling and the geometrically doubling conditions. In this paper, the authors establish some equivale…

cs.CL2026

Query as Anchor: Scenario-Adaptive User Representation via Large Language Model

Jiahao Yuan, Yike Xu, Jinyong Wen +9

Industrial-scale user representation learning requires balancing robust universality with acute task-sensitivity. However, existing paradigms primarily yield static, task-agnostic…

cs.CL2025

CARE-Bench: A Benchmark of Diverse Client Simulations Guided by Expert Principles for Evaluating LLMs in Psychological Counseling

Bichen Wang, Yixin Sun, Junzhe Wang +6

The mismatch between the growing demand for psychological counseling and the limited availability of services has motivated research into the application of Large Language Models (…

cs.CV2024

Clean-image Backdoor Attacks

Dazhong Rong, Guoyao Yu, Shuheng Shen +6

To gather a significant quantity of annotated training data for high-performance image classification models, numerous companies opt to enlist third-party providers to label their…

cs.LG2025

Beyond Tree Models: A Hybrid Model of KAN and gMLP for Large-Scale Financial Tabular Data

Mingming Zhang, Jiahao Hu, Pengfei Shi +8

Tabular data plays a critical role in real-world financial scenarios. Traditionally, tree models have dominated in handling tabular data. However, financial datasets in the industr…

physics.optics2024

Scalable photonic diffractive generators through sampling noises from scattering medium

Ziyu Zhan, Hao Wang, Qiang Liu +1

Photonic computing, with potentials of high parallelism, low latency and high energy efficiency, have gained progressive interest at the forefront of neural network (NN) accelerato…

cs.CV2025

Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation

Xuechao Zou, Shun Zhang, Xing Fu +6

Controllable face generation poses critical challenges in generative modeling due to the intricate balance required between semantic controllability and photorealism. While existin…

cs.CV2022

Few-shot Non-line-of-sight Imaging with Signal-surface Collaborative Regularization

Xintong Liu, Jianyu Wang, Leping Xiao +3

The non-line-of-sight imaging technique aims to reconstruct targets from multiply reflected light. For most existing methods, dense points on the relay surface are raster scanned t…

physics.optics2018

Hybrid topological evolution of multi-singularity vortex beams: Generalized nature for helical-Ince-Gaussian and Hermite-Laguerre-Gaussian modes

Yijie Shen, Yuan Meng, Xing Fu +1

A generalized family of scalar structured Gaussian modes including helical-Ince--Gaussian (HIG) and Hermite--Laguerre--Gaussian (HLG) beams is presented with physical insight upon…

cs.AI2026

InCarEmo: A Multimodal Dataset for In-Cabin Emotion Recognition and Driver State Monitoring

Hao Yang, Yanyan Zhao, Kewei Zhao +11

The paper presents InCarEmo, a multimodal dataset that combines RGB and infrared video, audio, and dialogue text for in-cabin emotion recognition, fatigue detection, and distractio…

#in-cabin emotion recognition#multimodal dataset#driver state monitoring#fatigue detection
math.CA2016

Riesz Transform Characterization and Fefferman-Stein Decomposition of Triebel-Lizorkin Spaces

Xing Fu, Dachun Yang, Qixiang Yang

Let , and be the Euclidean space equipped with the -dimensional Lebesgue measure. In this article, via an auxiliary…

physics.optics2024

Vector Angular Spectrum Model for light travelling in scattering media

Kaige Liu, Hengkang Zhang, Zeqi Liu +4

Strongly scattering media disrupt both the wavefront distribution and the polarization state of the incident light field. Controlling and effectively utilizing depolarization effec…

cs.CV2026

Multivariate Diffusion Transformer with Decoupled Attention for High-Fidelity Mask-Text Collaborative Facial Generation

Yushe Cao, Dianxi Shi, Xing Fu +5

While significant progress has been achieved in multimodal facial generation using semantic masks and textual descriptions, conventional feature fusion approaches often fail to ena…

cs.LG2026

FOUNDv2: Learning Unified User Quantized Tokenizers for User Representation

Chuan He, Yang Chen, Bin Dou +10

User representation learning serves as a fundamental pillar for personalized services on large-scale web platforms. Despite its importance, conventional continuous embedding method…

cs.LG2026

From Parameters to Data: A Task-Parameter-Guided Fine-Tuning Pipeline for Efficient LLM Alignment

Hao Chen, Qi Zhang, Liyao Li +7

Adapting Large Language Models (LLMs) to specialized domains typically incurs high data and computational overhead. While prior efficiency efforts have largely treated data selecti…

cs.CL2025

Psychological Counseling Cannot Be Achieved Overnight: Automated Psychological Counseling Through Multi-Session Conversations

Junzhe Wang, Bichen Wang, Xing Fu +3

In recent years, Large Language Models (LLMs) have made significant progress in automated psychological counseling. However, current research focuses on single-session counseling,…

physics.optics2020

Two-dimensional-controlled high-order modes and vortex beams from an intracavity mode converter laser

Jing Pan, Yijie Shen, Zhensong Wan +3

We present a novel scheme of structured light laser with an astigmatic mode converter (AMC) as intracavity element, first enabling the generation of Hermite-Gaussian (HG) modes wit…

cs.LG2026

KMLP: A Scalable Hybrid Architecture for Web-Scale Tabular Data Modeling

Mingming Zhang, Pengfei Shi, Zhiqing Xiao +8

Predictive modeling on web-scale tabular data with billions of instances and hundreds of heterogeneous numerical features faces significant scalability challenges. These features e…

physics.optics2018

Dual-wavelength vortex beam with high stability in diode-pumped Yb:CaGdAlO4 Laser

Yijie Shen, Yuan Meng, Xing Fu +1

We present a stable dual-wavelength vortex beam carrying orbital angular momentum (OAM) with two spectral peaks separated by a few terahertz in diode-pumped Yb:CaGdAlO4 (CALGO) las…

physics.optics2021

Deep-learning-based recognition of multi-singularity structured light

Hao Wang, Xilin Yang, Zeqi Liu +8

Structured light with customized complex topological pattern inspires diverse classical and quantum investigations underpinned by accurate detection techniques. However, the curren…

cs.AI2024

AIGT: AI Generative Table Based on Prompt

Mingming Zhang, Zhiqing Xiao, Guoshan Lu +5

Tabular data, which accounts for over 80% of enterprise data assets, is vital in various fields. With growing concerns about privacy protection and data-sharing restrictions, gener…

cs.ET2022

Intelligent optoelectronic processor for orbital angular momentum spectrum measurement

Hao Wang, Ziyu Zhan, Futai Hu +4

Orbital angular momentum (OAM) detection underpins almost all aspects of vortex beams' advances such as communication and quantum analogy. Conventional schemes are frustrated by lo…

cs.LG2024

Revisiting Modularity Maximization for Graph Clustering: A Contrastive Learning Perspective

Yunfei Liu, Jintang Li, Yuehe Chen +9

Graph clustering, a fundamental and challenging task in graph mining, aims to classify nodes in a graph into several disjoint clusters. In recent years, graph contrastive learning…

cs.CL2026

TabEmbed: Benchmarking and Learning Generalist Embeddings for Tabular Understanding

Minjie Qiang, Mingming Zhang, Xiaoyi Bao +5

Foundation models have established unified representations for natural language processing, yet this paradigm remains largely unexplored for tabular data. Existing methods face fun…

math.FA2025

Characterization of Vanishing Campanato Spaces via Ball Banach Function Spaces and Its Applications

Xing Fu, Yoshihiro Sawano, Jin Tao +1

In this article, the authors provide some new characterizations of several vanishing Campanato spaces using a type of oscillation defined within the general framework of ball Banac…

cs.LG2021

SHORING: Design Provable Conditional High-Order Interaction Network via Symbolic Testing

Hui Li, Xing Fu, Ruofan Wu +8

Deep learning provides a promising way to extract effective representations from raw data in an end-to-end fashion and has proven its effectiveness in various domains such as compu…

math.CA2014

Hardy spaces over non-homogeneous metric measure spaces and their applications

Xing Fu, Haibo Lin, Dachun Yang +1

Let be a metric measure space satisfying both the geometrically doubling and the upper doubling conditions. Let , , $…

cs.CL2026

Table as a Modality for Large Language Models

Liyao Li, Chao Ye, Wentao Ye +9

To migrate the remarkable successes of Large Language Models (LLMs), the community has made numerous efforts to generalize them to the table reasoning tasks for the widely deployed…

physics.optics2018

Truncated triangular diffraction lattices and orbital-angular-momentum detection of vortex SU(2) geometric modes

Yijie Shen, Xing Fu, Mali Gong

We for the first time report the truncated diffraction with a triangular aperture of the SU(2) geometric modes and propose a method to detect the complicated orbital angular moment…