arXiv subjects

Yong Li

Publications and source records attributed to Yong Li.

At least 379 records · Page 21Linked to original sources

arXiv2025

A Survey of Large Language Model-Powered Spatial Intelligence Across Scales: Advances in Embodied Agents, Smart Cities, and Earth Science

Over the past year, the development of large language models (LLMs) has brought spatial intelligence into focus, with much attention on vision-based embodied intelligence. However, spatial intelligence spans a broader range of disciplines and scales, from navigation and urban planning to remote sensing and earth science. What are the differences and connections between spatial intelligence across these fields? In this paper, we first review human spatial cognition and its implications for spatial intelligence in LLMs. We then examine spatial memory, knowledge representations, and abstract reasoning in LLMs, highlighting their roles and connections. Finally, we analyze spatial intelligence across scales -- from embodied to urban and global levels -- following a framework that progresses from spatial memory and understanding to spatial reasoning and intelligence. Through this survey, we aim to provide insights into interdisciplinary spatial intelligence research and inspire future studies.

cs.AI↗

arXiv2025

Multi-view Intent Learning and Alignment with Large Language Models for Session-based Recommendation

Session-based recommendation (SBR) methods often rely on user behavior data, which can struggle with the sparsity of session data, limiting performance. Researchers have identified that beyond behavioral signals, rich semantic information in item descriptions is crucial for capturing hidden user intent. While large language models (LLMs) offer new ways to leverage this semantic data, the challenges of session anonymity, short-sequence nature, and high LLM training costs have hindered the development of a lightweight, efficient LLM framework for SBR. To address the above challenges, we propose an LLM-enhanced SBR framework that integrates semantic and behavioral signals from multiple views. This two-stage framework leverages the strengths of both LLMs and traditional SBR models while minimizing training costs. In the first stage, we use multi-view prompts to infer latent user intentions at the session semantic level, supported by an intent localization module to alleviate LLM hallucinations. In the second stage, we align and unify these semantic inferences with behavioral representations, effectively merging insights from both large and small models. Extensive experiments on two real datasets demonstrate that the LLM4SBR framework can effectively improve model performance. We release our codes along with the baselines at https://github.com/tsinghua-fib-lab/LLM4SBR.

cs.IR↗

arXiv2025

A Survey of Large Language Model Empowered Agents for Recommendation and Search: Towards Next-Generation Information Retrieval

Information technology has profoundly altered the way humans interact with information. The vast amount of content created, shared, and disseminated online has made it increasingly difficult to access relevant information. Over the past two decades, recommender systems and search (collectively referred to as information retrieval systems) have evolved significantly to address these challenges. Recent advances in large language models (LLMs) have demonstrated capabilities that surpass human performance in various language-related tasks and exhibit general understanding, reasoning, and decision-making abilities. This paper explores the transformative potential of LLM agents in enhancing recommender and search systems. We discuss the motivations and roles of LLM agents, and establish a classification framework to elaborate on the existing research. We highlight the immense potential of LLM agents in addressing current challenges in recommendation and search, providing insights into future research directions. This paper is the first to systematically review and classify the research on LLM agents in these domains, offering a novel perspective on leveraging this advanced AI technology for information retrieval. To help understand the existing works, we list the existing papers on LLM agent based recommendation and search at this link: https://github.com/tsinghua-fib-lab/LLM-Agent-for-Recommendation-and-Search.

cs.IR↗

arXiv2025

An Evaluation of Cultural Value Alignment in LLM

LLMs as intelligent agents are being increasingly applied in scenarios where human interactions are involved, leading to a critical concern about whether LLMs are faithful to the variations in culture across regions. Several works have investigated this question in various ways, finding that there are biases present in the cultural representations of LLM outputs. To gain a more comprehensive view, in this work, we conduct the first large-scale evaluation of LLM culture assessing 20 countries' cultures and languages across ten LLMs. With a renowned cultural values questionnaire and by carefully analyzing LLM output with human ground truth scores, we thoroughly study LLMs' cultural alignment across countries and among individual models. Our findings show that the output over all models represents a moderate cultural middle ground. Given the overall skew, we propose an alignment metric, revealing that the United States is the best-aligned country and GLM-4 has the best ability to align to cultural values. Deeper investigation sheds light on the influence of model origin, prompt language, and value dimensions on cultural output. Specifically, models, regardless of where they originate, align better with the US than they do with China. The conclusions provide insight to how LLMs can be better aligned to various cultures as well as provoke further discussion of the potential for LLMs to propagate cultural bias and the need for more culturally adaptable models.

cs.CY↗

arXiv2025

Opening the black box of deep learning: Validating the statistical association between explainable artificial intelligence (XAI) and clinical domain knowledge in fundus image-based glaucoma diagnosis

While deep learning has exhibited remarkable predictive capabilities in various medical image tasks, its inherent black-box nature has hindered its widespread implementation in real-world healthcare settings. Our objective is to unveil the decision-making processes of deep learning models in the context of glaucoma classification by employing several Class Activation Map (CAM) techniques to generate model focus regions and comparing them with clinical domain knowledge of the anatomical area (optic cup, optic disk, and blood vessels). Four deep neural networks, including VGG-11, ResNet-18, DeiT-Tiny, and Swin Transformer-Tiny, were developed using binary diagnostic labels of glaucoma and five CAM methods (Grad-CAM, XGrad-CAM, Score-CAM, Eigen-CAM, and Layer-CAM) were employed to highlight the model focus area. We applied the paired-sample t-test to compare the percentage of anatomies in the model focus area to the proportion of anatomies in the entire image. After that, Pearson's and Spearman's correlation tests were implemented to examine the relationship between model predictive ability and the percentage of anatomical structures in the model focus area. On five public glaucoma datasets, all deep learning models consistently displayed statistically significantly higher percentages of anatomical structures in the focus area than the proportions of anatomical structures in the entire image. Also, we validated the positive relationship between the percentage of anatomical structures in the focus area and model predictive performance. Our study provides evidence of the convergence of decision logic between deep neural networks and human clinicians through rigorous statistical tests. We anticipate that it can help alleviate clinicians' concerns regarding the trustworthiness of deep learning in healthcare. For reproducibility, the code and dataset have been released at GitHub.

cs.CV↗

arXiv2025

UniFlow: A Foundation Model for Unified Urban Spatio-Temporal Flow Prediction

Urban spatio-temporal flow prediction, encompassing traffic flows and crowd flows, is crucial for optimizing city infrastructure and managing traffic and emergency responses. Traditional approaches have relied on separate models tailored to either grid-based data, representing cities as uniform cells, or graph-based data, modeling cities as networks of nodes and edges. In this paper, we build UniFlow, a foundational model for general urban flow prediction that unifies both grid-based and graphbased data. We first design a multi-view spatio-temporal patching mechanism to standardize different data into a consistent sequential format and then introduce a spatio-temporal transformer architecture to capture complex correlations and dynamics. To leverage shared spatio-temporal patterns across different data types and facilitate effective cross-learning, we propose SpatioTemporal Memory Retrieval Augmentation (ST-MRA). By creating structured memory modules to store shared spatio-temporal patterns, ST-MRA enhances predictions through adaptive memory retrieval. Extensive experiments demonstrate that UniFlow outperforms existing models in both grid-based and graph-based flow prediction, excelling particularly in scenarios with limited data availability, showcasing its superior performance and broad applicability. The datasets and code implementation have been released on https://github.com/YuanYuan98/UniFlow.

cs.LG↗

arXiv2025

The weak averaging principle of stochastic functional partial differential equations with H$\ddot{\text{o}}$lder continuous coefficients and infinite delay

In this paper, we establish the weak averaging principle for stochastic functional partial differential equations (in short, SFPDEs) with H$\ddot{\text{o}}$lder continuous coefficients and infinite delay by a new generalized coupling approach. Firstly, we rigorously establish the existence and uniqueness of weak solutions for a specific class of finite-dimensional systems by the generalized coupling approach. Then we extend these results to their infinite-dimensional counterparts using the variational approach and Galerkin projection technique. Subsequently, we establish the averaging principle for SFPDEs with infinite delay in the weak sense, i.e., we prove that the solution of the original system converges in law to that of the averaged system on a finite interval $[0,T]$ as the small parameter $\varepsilon\to 0$. To illustrate our findings, we present two applications: stochastic generalized porous media equations and stochastic reaction-diffusion equations.

math.PR↗

arXiv2025

DiffMove: Group Mobility Tendency Enhanced Trajectory Recovery via Diffusion Model

In the real world, trajectory data is often sparse and incomplete due to low collection frequencies or limited device coverage. Trajectory recovery aims to recover these missing trajectory points, making the trajectories denser and more complete. However, this task faces two key challenges: 1) The excessive sparsity of individual trajectories makes it difficult to effectively leverage historical information for recovery; 2) Sparse trajectories make it harder to capture complex individual mobility preferences. To address these challenges, we propose a novel method called DiffMove. Firstly, we harness crowd wisdom for trajectory recovery. Specifically, we construct a group tendency graph using the collective trajectories of all users and then integrate the group mobility trends into the location representations via graph embedding. This solves the challenge of sparse trajectories being unable to rely on individual historical trajectories for recovery. Secondly, we capture individual mobility preferences from both historical and current perspectives. Finally, we integrate group mobility tendencies and individual preferences into the spatiotemporal distribution of the trajectory to recover high-quality trajectories. Extensive experiments on two real-world datasets demonstrate that DiffMove outperforms existing state-of-the-art methods. Further analysis validates the robustness of our method.

cs.AI↗

arXiv2025

Simulating Filter Bubble on Short-video Recommender System with Large Language Model Agents

An increasing reliance on recommender systems has led to concerns about the creation of filter bubbles on social media, especially on short video platforms like TikTok. However, their formation is still not entirely understood due to the complex dynamics between recommendation algorithms and user feedback. In this paper, we aim to shed light on these dynamics using a large language model-based simulation framework. Our work employs real-world short-video data containing rich video content information and detailed user-agents to realistically simulate the recommendation-feedback cycle. Through large-scale simulations, we demonstrate that LLMs can replicate real-world user-recommender interactions, uncovering key mechanisms driving filter bubble formation. We identify critical factors, such as demographic features and category attraction that exacerbate content homogenization. To mitigate this, we design and test interventions including various cold-start and feedback weighting strategies, showing measurable reductions in filter bubble effects. Our framework enables rapid prototyping of recommendation strategies, offering actionable solutions to enhance content diversity in real-world systems. Furthermore, we analyze how LLM-inherent biases may propagate through recommendations, proposing safeguards to promote equity for vulnerable groups, such as women and low-income populations. By examining the interplay between recommendation and LLM agents, this work advances a deeper understanding of algorithmic bias and provides practical tools to promote inclusive digital spaces.

cs.IR↗

arXiv2025

Quantitative uniform exponential acceleration of averages along decaying waves

In this study, utilizing a specific exponential weighting function, we investigate the uniform exponential convergence of weighted Birkhoff averages along decaying waves and delve into several related variants. A key distinction from traditional scenarios is evident here: despite reduced regularity in observables, our method still maintains exponential convergence. In particular, we develop new techniques that yield very precise rates of exponential convergence, as evidenced by numerical simulations. Furthermore, this innovative approach extends to quantitative analyses involving different weighting functions employed by others, surpassing the limitations inherent in prior research. It also enhances the exponential convergence rates of weighted Birkhoff averages along quasi-periodic orbits via analytic observables. To the best of our knowledge, this is the first result on the uniform exponential acceleration beyond averages along quasi-periodic or almost periodic orbits, particularly from a quantitative perspective.

math.DS↗

arXiv2025

Flexible WMMSE Beamforming for MU-MIMO Movable Antenna Communications

Movable antennas offer new potential for wireless communication by introducing degrees of freedom in antenna positioning, which has recently been explored for improving sum rates. In this paper, we aim to fully leverage the capabilities of movable antennas (MAs) by assuming that both the transmitter and receiver can optimize their antenna positions in multi-user multiple-input multiple-output (MU-MIMO) communications. Recognizing that WMMSE beamforming is a highly effective method for maximizing the MU-MIMO sum rate, we modify it to integrate antenna position optimization for MA systems, which we refer to as flexible WMMSE (F-WMMSE) beamforming. Importantly, we reformulate the subproblems within WMMSE to develop regularized sparse optimization frameworks to achieve joint beamforming (antenna coefficient optimization) and element movement (antenna position optimization). We then propose a regularized least squares-based simultaneous orthogonal matching pursuit (RLS-SOMP) algorithm to address the resulting sparse optimization problem. To enhance practical applications, the low-complexity implementation of the proposed framework is developed based on the pre-calculations and matrix inverse lemma. The overall F-WMMSE algorithm converges similarly to WMMSE, and our findings indicate that F-WMMSE achieves a significant sum rate improvement compared to traditional WMMSE, exceeding 20% under appropriate simulation conditions

eess.SP↗

arXiv2025

The Onsager-Machlup functional for distribution dependent SDEs driven by fractional Brownian motion

In this paper, we compute the Onsager-Machlup functional for distribution dependent SDEs driven by fractional Brownian motions with Hurst parameter $H\in (\frac{1}{4},1)$. In the case $ \frac{1}{4} < H < \frac{1}{2} $, the norm can be either the supremum norm or Hölder norms of order $ β$ with $ 0 < β< H - \frac{1}{4} $. In the case $\frac{1}{2} < H < 1 $, the norms can be a Hölder norm of order $ β$ with $ H - \frac{1}{2} < β< H - \frac{1}{4} $. As an example, we compute the Onsager-Machlup functional for the stochastic pendulum equation

math.DS↗

arXiv2025

Semi-Supervised Learning for AVO Inversion with Strong Spatial Feature Constraints

One-dimensional convolution is a widely used deep learning technique in prestack amplitude variation with offset (AVO) inversion; however, it lacks lateral continuity. Although two-dimensional convolution improves lateral continuity, due to the sparsity of well-log data, the model only learns weak spatial features and fails to explore the spatial correlations in seismic data fully. To overcome these challenges, we propose a novel AVO inversion method based on semi-supervised learning with strong spatial feature constraints (SSFC-SSL). First, two-dimensional predicted values are obtained through the inversion network, and the predicted values at well locations are sparsely represented using well-log labels. Subsequently, a label-annihilation operator is introduced, enabling the predicted values at non-well locations to learn the spatial features of well locations through the neural network. Ultimately, a two-way strong spatial feature mapping between non-well locations and well locations is achieved. Additionally, to reduce the dependence on well-log labels, we combine the semi-supervised learning strategy with a low-frequency model, further enhancing the robustness of the method. Experimental results on both synthetic example and field data demonstrate that the proposed method significantly improves lateral continuity and inversion accuracy compared to one- and two-dimensional deep learning techniques.

physics.geo-ph↗

arXiv2025

The Nash-MTL-STCN Method For Prestack Three-Parameter Inversion

Deep learning (DL) techniques have been widely used in prestack three-parameter inversion to address its ill-posed problems. Among these DL techniques, Multi-task learning (MTL) methods can simultaneously train multiple tasks, thereby enhancing model generalization and predictive performance. However, existing MTL methods typically adopt heuristic or non-heuristic approaches to jointly update the gradient of each task, leading to gradient conflicts between different tasks and reducing inversion accuracy. To address this issue, we propose a semi-supervised temporal convolutional network (STCN) based on Nash equilibrium (Nash-MTL-STCN). Firstly, temporal convolutional networks (TCN) with non-causal convolution and convolutional neural networks (CNNs) are used as multi-task layers to extract the shared features from partial angle stack seismic data, with CNNs serving as the single-task layer. Subsequently, the feature mechanism is utilized to extract shared features in the multi-task layer through hierarchical processing, and the gradient combination of these shared features is treated as a Nash game for strategy optimization and joint updates. Ultimately, the overall utility of the three-parameter is maximized, and gradient conflicts are alleviated. In addition, to enhance the network's generalization and stability, we have incorporated geophysical forward modeling and low-frequency models into the network. Experimental results demonstrate that the proposed method overcomes the gradient conflict issue of the conventional MTL methods with constant weights (CW) and achieves higher precision than four widely used non-heuristic MTL methods. Further field data experiments also validate the method's effectiveness.

physics.geo-ph↗

arXiv2025

Resurgence in the Universal Structures in B-model Topological String Theory

We propose a systematic analysis of Alim-Yau-Zhou's double scaling limit and Couso-Santamaría's large radius limit for the perturbative free energies in B-model topological string theory based on Écalle's Resurgence Theory. Taking advantage of the known resurgent properties of the formal solutions to the Airy equation and of the stability of resurgent series under exponential/logarithm and nonlinear changes of variable, we show how to rigorously derive the non-perturbative information from the perturbative one by means of alien calculus in this context, spelling out the notions of formal integral and Bridge Equation, typical of the resurgent approach to ordinary differential equations. We also discuss the Borel-Laplace summation of the obtained resurgent transseries, including a study of real analyticity based on the connection formulas stemming from the resummation of the Bridge Equation.

math-ph↗

arXiv2025

EmT: A Novel Transformer for Generalized Cross-subject EEG Emotion Recognition

Integrating prior knowledge of neurophysiology into neural network architecture enhances the performance of emotion decoding. While numerous techniques emphasize learning spatial and short-term temporal patterns, there has been limited emphasis on capturing the vital long-term contextual information associated with emotional cognitive processes. In order to address this discrepancy, we introduce a novel transformer model called emotion transformer (EmT). EmT is designed to excel in both generalized cross-subject EEG emotion classification and regression tasks. In EmT, EEG signals are transformed into a temporal graph format, creating a sequence of EEG feature graphs using a temporal graph construction module (TGC). A novel residual multi-view pyramid GCN module (RMPG) is then proposed to learn dynamic graph representations for each EEG feature graph within the series, and the learned representations of each graph are fused into one token. Furthermore, we design a temporal contextual transformer module (TCT) with two types of token mixers to learn the temporal contextual information. Finally, the task-specific output module (TSO) generates the desired outputs. Experiments on four publicly available datasets show that EmT achieves higher results than the baseline methods for both EEG emotion classification and regression tasks. The code is available at https://github.com/yi-ding-cs/EmT.

cs.LG↗

arXiv2025

Beyond Overfitting: Doubly Adaptive Dropout for Generalizable AU Detection

Facial Action Units (AUs) are essential for conveying psychological states and emotional expressions. While automatic AU detection systems leveraging deep learning have progressed, they often overfit to specific datasets and individual features, limiting their cross-domain applicability. To overcome these limitations, we propose a doubly adaptive dropout approach for cross-domain AU detection, which enhances the robustness of convolutional feature maps and spatial tokens against domain shifts. This approach includes a Channel Drop Unit (CD-Unit) and a Token Drop Unit (TD-Unit), which work together to reduce domain-specific noise at both the channel and token levels. The CD-Unit preserves domain-agnostic local patterns in feature maps, while the TD-Unit helps the model identify AU relationships generalizable across domains. An auxiliary domain classifier, integrated at each layer, guides the selective omission of domain-sensitive features. To prevent excessive feature dropout, a progressive training strategy is used, allowing for selective exclusion of sensitive features at any model layer. Our method consistently outperforms existing techniques in cross-domain AU detection, as demonstrated by extensive experimental evaluations. Visualizations of attention maps also highlight clear and meaningful patterns related to both individual and combined AUs, further validating the approach's effectiveness.

cs.CV↗

arXiv2025

Decoupled Doubly Contrastive Learning for Cross Domain Facial Action Unit Detection

Despite the impressive performance of current vision-based facial action unit (AU) detection approaches, they are heavily susceptible to the variations across different domains and the cross-domain AU detection methods are under-explored. In response to this challenge, we propose a decoupled doubly contrastive adaptation (D$^2$CA) approach to learn a purified AU representation that is semantically aligned for the source and target domains. Specifically, we decompose latent representations into AU-relevant and AU-irrelevant components, with the objective of exclusively facilitating adaptation within the AU-relevant subspace. To achieve the feature decoupling, D$^2$CA is trained to disentangle AU and domain factors by assessing the quality of synthesized faces in cross-domain scenarios when either AU or domain attributes are modified. To further strengthen feature decoupling, particularly in scenarios with limited AU data diversity, D$^2$CA employs a doubly contrastive learning mechanism comprising image and feature-level contrastive learning to ensure the quality of synthesized faces and mitigate feature ambiguities. This new framework leads to an automatically learned, dedicated separation of AU-relevant and domain-relevant factors, and it enables intuitive, scale-specific control of the cross-domain facial image synthesis. Extensive experiments demonstrate the efficacy of D$^2$CA in successfully decoupling AU and domain factors, yielding visually pleasing cross-domain synthesized facial images. Meanwhile, D$^2$CA consistently outperforms state-of-the-art cross-domain AU detection approaches, achieving an average F1 score improvement of 6\%-14\% across various cross-domain scenarios.

cs.CV↗