arXiv subjects

Yan Xie

Publications and source records attributed to Yan Xie.

At least 55 records · Page 3Linked to original sources

arXiv2025

GSTBench: A Benchmark Study on the Transferability of Graph Self-Supervised Learning

Self-supervised learning (SSL) has shown great promise in graph representation learning. However, most existing graph SSL methods are developed and evaluated under a single-dataset setting, leaving their cross-dataset transferability largely unexplored and limiting their ability to leverage knowledge transfer and large-scale pretraining, factors that are critical for developing generalized intelligence beyond fitting training data. To address this gap and advance foundation model research for graphs, we present GSTBench, the first systematic benchmark for evaluating the transferability of graph SSL methods. We conduct large-scale pretraining on ogbn-papers100M and evaluate five representative SSL methods across a diverse set of target graphs. Our standardized experimental setup decouples confounding factors such as model architecture, dataset characteristics, and adaptation protocols, enabling rigorous comparisons focused solely on pretraining objectives. Surprisingly, we observe that most graph SSL methods struggle to generalize, with some performing worse than random initialization. In contrast, GraphMAE, a masked autoencoder approach, consistently improves transfer performance. We analyze the underlying factors that drive these differences and offer insights to guide future research on transferable graph SSL, laying a solid foundation for the "pretrain-then-transfer" paradigm in graph learning. Our code is available at https://github.com/SongYYYY/GSTBench.

cs.LG↗

arXiv2025

A Scalable Pretraining Framework for Link Prediction with Efficient Adaptation

Link Prediction (LP) is a critical task in graph machine learning. While Graph Neural Networks (GNNs) have significantly advanced LP performance recently, existing methods face key challenges including limited supervision from sparse connectivity, sensitivity to initialization, and poor generalization under distribution shifts. We explore pretraining as a solution to address these challenges. Unlike node classification, LP is inherently a pairwise task, which requires the integration of both node- and edge-level information. In this work, we present the first systematic study on the transferability of these distinct modules and propose a late fusion strategy to effectively combine their outputs for improved performance. To handle the diversity of pretraining data and avoid negative transfer, we introduce a Mixture-of-Experts (MoE) framework that captures distinct patterns in separate experts, facilitating seamless application of the pretrained model on diverse downstream datasets. For fast adaptation, we develop a parameter-efficient tuning strategy that allows the pretrained model to adapt to unseen datasets with minimal computational overhead. Experiments on 16 datasets across two domains demonstrate the effectiveness of our approach, achieving state-of-the-art performance on low-resource link prediction while obtaining competitive results compared to end-to-end trained methods, with over 10,000x lower computational overhead.

cs.LG↗

arXiv2025

QR-LoRA: Efficient and Disentangled Fine-tuning via QR Decomposition for Customized Generation

Existing text-to-image models often rely on parameter fine-tuning techniques such as Low-Rank Adaptation (LoRA) to customize visual attributes. However, when combining multiple LoRA models for content-style fusion tasks, unstructured modifications of weight matrices often lead to undesired feature entanglement between content and style attributes. We propose QR-LoRA, a novel fine-tuning framework leveraging QR decomposition for structured parameter updates that effectively separate visual attributes. Our key insight is that the orthogonal Q matrix naturally minimizes interference between different visual features, while the upper triangular R matrix efficiently encodes attribute-specific transformations. Our approach fixes both Q and R matrices while only training an additional task-specific $ΔR$ matrix. This structured design reduces trainable parameters to half of conventional LoRA methods and supports effective merging of multiple adaptations without cross-contamination due to the strong disentanglement properties between $ΔR$ matrices. Experiments demonstrate that QR-LoRA achieves superior disentanglement in content-style fusion tasks, establishing a new paradigm for parameter-efficient, disentangled fine-tuning in generative models. The project page is available at: https://luna-ai-lab.github.io/QR-LoRA/.

cs.CV↗

arXiv2025

Higher-order Structure Boosts Link Prediction on Temporal Graphs

Temporal Graph Neural Networks (TGNNs) have gained growing attention for modeling and predicting structures in temporal graphs. However, existing TGNNs primarily focus on pairwise interactions while overlooking higher-order structures that are integral to link formation and evolution in real-world temporal graphs. Meanwhile, these models often suffer from efficiency bottlenecks, further limiting their expressive power. To tackle these challenges, we propose a Higher-order structure Temporal Graph Neural Network, which incorporates hypergraph representations into temporal graph learning. In particular, we develop an algorithm to identify the underlying higher-order structures, enhancing the model's ability to capture the group interactions. Furthermore, by aggregating multiple edge features into hyperedge representations, HTGN effectively reduces memory cost during training. We theoretically demonstrate the enhanced expressiveness of our approach and validate its effectiveness and efficiency through extensive experiments on various real-world temporal graphs. Experimental results show that HTGN achieves superior performance on dynamic link prediction while reducing memory costs by up to 50\% compared to existing methods.

cs.LG↗

arXiv2025

Discovering Fine-Grained Visual-Concept Relations by Disentangled Optimal Transport Concept Bottleneck Models

Concept Bottleneck Models (CBMs) try to make the decision-making process transparent by exploring an intermediate concept space between the input image and the output prediction. Existing CBMs just learn coarse-grained relations between the whole image and the concepts, less considering local image information, leading to two main drawbacks: i) they often produce spurious visual-concept relations, hence decreasing model reliability; and ii) though CBMs could explain the importance of every concept to the final prediction, it is still challenging to tell which visual region produces the prediction. To solve these problems, this paper proposes a Disentangled Optimal Transport CBM (DOT-CBM) framework to explore fine-grained visual-concept relations between local image patches and concepts. Specifically, we model the concept prediction process as a transportation problem between the patches and concepts, thereby achieving explicit fine-grained feature alignment. We also incorporate orthogonal projection losses within the modality to enhance local feature disentanglement. To further address the shortcut issues caused by statistical biases in the data, we utilize the visual saliency map and concept label statistics as transportation priors. Thus, DOT-CBM can visualize inversion heatmaps, provide more reliable concept predictions, and produce more accurate class predictions. Comprehensive experiments demonstrate that our proposed DOT-CBM achieves SOTA performance on several tasks, including image classification, local part detection and out-of-distribution generalization.

cs.CV↗

arXiv2024

HICEScore: A Hierarchical Metric for Image Captioning Evaluation

Image captioning evaluation metrics can be divided into two categories, reference-based metrics and reference-free metrics. However, reference-based approaches may struggle to evaluate descriptive captions with abundant visual details produced by advanced multimodal large language models, due to their heavy reliance on limited human-annotated references. In contrast, previous reference-free metrics have been proven effective via CLIP cross-modality similarity. Nonetheless, CLIP-based metrics, constrained by their solution of global image-text compatibility, often have a deficiency in detecting local textual hallucinations and are insensitive to small visual objects. Besides, their single-scale designs are unable to provide an interpretable evaluation process such as pinpointing the position of caption mistakes and identifying visual regions that have not been described. To move forward, we propose a novel reference-free metric for image captioning evaluation, dubbed Hierarchical Image Captioning Evaluation Score (HICE-S). By detecting local visual regions and textual phrases, HICE-S builds an interpretable hierarchical scoring mechanism, breaking through the barriers of the single-scale structure of existing reference-free metrics. Comprehensive experiments indicate that our proposed metric achieves the SOTA performance on several benchmarks, outperforming existing reference-free metrics like CLIP-S and PAC-S, and reference-based metrics like METEOR and CIDEr. Moreover, several case studies reveal that the assessment process of HICE-S on detailed captions closely resembles interpretable human judgments.Our code is available at https://github.com/joeyz0z/HICE.

cs.CV↗

arXiv2024

A Scalable and Effective Alternative to Graph Transformers

Graph Neural Networks (GNNs) have shown impressive performance in graph representation learning, but they face challenges in capturing long-range dependencies due to their limited expressive power. To address this, Graph Transformers (GTs) were introduced, utilizing self-attention mechanism to effectively model pairwise node relationships. Despite their advantages, GTs suffer from quadratic complexity w.r.t. the number of nodes in the graph, hindering their applicability to large graphs. In this work, we present Graph-Enhanced Contextual Operator (GECO), a scalable and effective alternative to GTs that leverages neighborhood propagation and global convolutions to effectively capture local and global dependencies in quasilinear time. Our study on synthetic datasets reveals that GECO reaches 169x speedup on a graph with 2M nodes w.r.t. optimized attention. Further evaluations on diverse range of benchmarks showcase that GECO scales to large graphs where traditional GTs often face memory and time limitations. Notably, GECO consistently achieves comparable or superior quality compared to baselines, improving the SOTA up to 4.5%, and offering a scalable and effective solution for large-scale graph learning.

cs.LG↗

arXiv2024

Degree of Irrationality: Sentiment and Implied Volatility Surface

In this study, we constructed daily high-frequency sentiment data and used the VAR method to attempt to predict the next day's implied volatility surface. We utilized 630,000 text data entries from the East Money Stock Forum from 2014 to 2023 and employed deep learning methods such as BERT and LSTM to build daily market sentiment indicators. By applying FFT and EMD methods for sentiment decomposition, we found that high-frequency sentiment had a stronger correlation with at-the-money (ATM) options' implied volatility, while low-frequency sentiment was more strongly correlated with deep out-of-the-money (DOTM) options' implied volatility. Further analysis revealed that the shape of the implied volatility surface contains richer market sentiment information beyond just market panic. We demonstrated that incorporating this sentiment information can improve the accuracy of implied volatility surface predictions.

cs.LG↗

arXiv2024

MGCNN: a learnable multigrid solver for sparse linear systems from PDEs on structured grids

This paper presents a learnable solver tailored to iteratively solve sparse linear systems from discretized partial differential equations (PDEs). Unlike traditional approaches relying on specialized expertise, our solver streamlines the algorithm design process for a class of PDEs through training, which requires only training data of coefficient distributions. The proposed method is anchored by three core principles: (1) a multilevel hierarchy to promote rapid convergence, (2) adherence to linearity concerning the right-hand-side of equations, and (3) weights sharing across different levels to facilitate adaptability to various problem sizes. Built on these foundational principles and considering the similar computation pattern of the convolutional neural network (CNN) as multigrid components, we introduce a network adept at solving linear systems from PDEs with heterogeneous coefficients, discretized on structured grids. Notably, our proposed solver possesses the ability to generalize over right-hand-side terms, PDE coefficients, and grid sizes, thereby ensuring its training is purely offline. To evaluate its effectiveness, we train the solver on convection-diffusion equations featuring heterogeneous diffusion coefficients. The solver exhibits swift convergence to high accuracy over a range of grid sizes, extending from $31 \times 31$ to $4095 \times 4095$. Remarkably, our method outperforms the classical Geometric Multigrid (GMG) solver, demonstrating a speedup of approximately 3 to 8 times. Furthermore, our numerical investigation into the solver's capacity to generalize to untrained coefficient distributions reveals promising outcomes.

math.NA↗

arXiv2024

VCR-Graphormer: A Mini-batch Graph Transformer via Virtual Connections

Graph transformer has been proven as an effective graph learning method for its adoption of attention mechanism that is capable of capturing expressive representations from complex topological and feature information of graphs. Graph transformer conventionally performs dense attention (or global attention) for every pair of nodes to learn node representation vectors, resulting in quadratic computational costs that are unaffordable for large-scale graph data. Therefore, mini-batch training for graph transformers is a promising direction, but limited samples in each mini-batch can not support effective dense attention to encode informative representations. Facing this bottleneck, (1) we start by assigning each node a token list that is sampled by personalized PageRank (PPR) and then apply standard multi-head self-attention only on this list to compute its node representations. This PPR tokenization method decouples model training from complex graph topological information and makes heavy feature engineering offline and independent, such that mini-batch training of graph transformers is possible by loading each node's token list in batches. We further prove this PPR tokenization is viable as a graph convolution network with a fixed polynomial filter and jumping knowledge. However, only using personalized PageRank may limit information carried by a token list, which could not support different graph inductive biases for model training. To this end, (2) we rewire graphs by introducing multiple types of virtual connections through structure- and content-based super nodes that enable PPR tokenization to encode local and global contexts, long-range interaction, and heterophilous information into each node's token list, and then formalize our Virtual Connection Ranking based Graph Transformer (VCR-Graphormer).

cs.LG↗

arXiv2024

MeaCap: Memory-Augmented Zero-shot Image Captioning

Zero-shot image captioning (IC) without well-paired image-text data can be divided into two categories, training-free and text-only-training. Generally, these two types of methods realize zero-shot IC by integrating pretrained vision-language models like CLIP for image-text similarity evaluation and a pre-trained language model (LM) for caption generation. The main difference between them is whether using a textual corpus to train the LM. Though achieving attractive performance w.r.t. some metrics, existing methods often exhibit some common drawbacks. Training-free methods tend to produce hallucinations, while text-only-training often lose generalization capability. To move forward, in this paper, we propose a novel Memory-Augmented zero-shot image Captioning framework (MeaCap). Specifically, equipped with a textual memory, we introduce a retrieve-then-filter module to get key concepts that are highly related to the image. By deploying our proposed memory-augmented visual-related fusion score in a keywords-to-sentence LM, MeaCap can generate concept-centered captions that keep high consistency with the image with fewer hallucinations and more world-knowledge. The framework of MeaCap achieves the state-of-the-art performance on a series of zero-shot IC settings. Our code is available at https://github.com/joeyz0z/MeaCap.

cs.CV↗

arXiv2023

Cross-hospital Sepsis Early Detection via Semi-supervised Optimal Transport with Self-paced Ensemble

Leveraging machine learning techniques for Sepsis early detection and diagnosis has attracted increasing interest in recent years. However, most existing methods require a large amount of labeled training data, which may not be available for a target hospital that deploys a new Sepsis detection system. More seriously, as treated patients are diversified between hospitals, directly applying a model trained on other hospitals may not achieve good performance for the target hospital. To address this issue, we propose a novel semi-supervised transfer learning framework based on optimal transport theory and self-paced ensemble for Sepsis early detection, called SPSSOT, which can efficiently transfer knowledge from the source hospital (with rich labeled data) to the target hospital (with scarce labeled data). Specifically, SPSSOT incorporates a new optimal transport-based semi-supervised domain adaptation component that can effectively exploit all the unlabeled data in the target hospital. Moreover, self-paced ensemble is adapted in SPSSOT to alleviate the class imbalance issue during transfer learning. In a nutshell, SPSSOT is an end-to-end transfer learning method that automatically selects suitable samples from two domains (hospitals) respectively and aligns their feature spaces. Extensive experiments on two open clinical datasets, MIMIC-III and Challenge, demonstrate that SPSSOT outperforms state-of-the-art transfer learning methods by improving 1-3% of AUC.

cs.LG↗

arXiv2023

Generation of highly retrievable atom photon entanglement with a millisecond lifetime via a spatially multiplexed cavity

Qubit memory that is entangled with photonic qubit is the building block for long distance quantum repeaters. Cavity enhanced and long lived spin wave photon entanglement has been demonstrated by applying dual laser beams onto optical-lattice atoms. However, owing to cross readouts by two beams, retrieval efficiency of spin wave qubit is decreased by one quarter compared to that of single mode spin wave at all storage times. Here, by coupling cold atoms to two modes of a polarization interferometer based cavity, we achieve perfect qubit retrieval in cavity enhanced and long lived atom photon entanglement. A write laser beam is applied onto cold atoms, we then create a magnetic field insensitive spin wave qubit that is entangled with the photonic qubit encoded onto two arms of the interferometer. The spin wave qubit is retrieved by a read beam, which avoids the cross readouts. Our experiment demonstrates 540μs storage time at 50% intrinsic qubit retrieval efficiency, which is 13.5 times longer than the best reported result.

quant-ph↗

arXiv2021

Generation of perfect-cavity-enhanced atom-photon entanglement with a millisecond lifetime via a spatially-multiplexed cavity

A qubit memory is the building block for quantum information. Cavity-enhanced spin-wave-photon entanglement has been achieved by applying dual-control modes. However, owing to cross readouts between the modes, the qubit retrieval efficiency is about one quarter lower than that for a single spin-wave mode at all storage times. Here, we overcome cross readouts using a multiplexed ring cavity. The cavity is embedded with a polarization interferometer, and we create a write-out photonic qubit entangled with a magnetic-field-insensitive spin-wave qubit by applying a single-mode write-laser beam to cold atoms. The spin-wave qubit is retrieved with a single-mode read-laser beam, and the quarter retrieval-efficiency loss is avoided at all storage times. Our experiment demonstrates 50% intrinsic retrieval efficiency for 540 microsecond storage time, which is 13.5 times longer than the best reported result. Importantly, our multiplexed-cavity scheme paves one road to generate perfect-cavity-enhanced and large-scale multiplexed spin-wave-photon entanglement with a long lifetime.

quant-ph↗

arXiv2021

Local large temperature difference and ultra-wideband photothermoelectric response of the silver nanostructure film/carbon nanotube film heterostructure

Photothermoelectric materials have important applications in many fields. Here, we joined a silver nanostructure film (AgNSF) and a carbon nanotube film (CNTF) by van der Waals force to form a AgNSF/CNTF heterojunction, which shows excellent photothermal and photoelectric conversion properties. The local temperature difference and the output photovoltage increase rapidly when the heterojunction is irradiated by lasers with wavelengths ranging from ultraviolet to terahertz. The maximum of the local temperature difference reaches 205.9 K, which is significantly higher than that of other photothermoelectric materials reported in literatures. The photothermal and photoelectric responsivity depend on the wavelength of lasers, which are 175-601 K/W and 9.35-40.4 mV/W, respectively. We demonstrate that light absorption of the carbon nanotube is enhanced by local surface plasmons, and the output photovoltage is dominated by Seebeck effect. The AgNSF/CNTF heterostructure can be used as high-efficiency sensitive photothermal materials or as ultra-wideband fast-response photoelectric material.

physics.app-ph↗

arXiv2021

Ultrabroadband THz/IR upconversion and photovoltaic response in semi-conductor ratchet based upconverter

An ultrabroadband upconversion device is demonstrated by direct tandem integration of a p-type GaAs/AlxGa1-xAs ratchet photodetector (RP) with a GaAs double heterojunction LED (DH-LED) using the molecular beam epitaxy (MBE). An ultrabroadband photoresponse from terahertz (THz) to near infrared (NIR) region (4-200 THz) was realized that covers a much wider frequency range com-pared with the existing upconversion devices. Broadband IR/THz radiation from 1000 K blackbody is successfully upconverted into NIR photons which can be detected by commercial Si-based device. The normal incidence absorption of the RP simplifies the structure of the RP-LED device and make it more compact compared with the inter-subband transition based upconverters. In addition to the up-conversion function, the proposed upconverter is also tested as photovoltaic detectors in the infrared region (15-200 THz) without an applied bias voltage due to the ratchet effect.

physics.app-ph↗

arXiv2021

Realization of ultrabroadband THz/IR photoresponse in a bias-tunable ratchet photodetector

High performance Terahertz (THz) photodetector has drawn wide attention and got great improvement due to its significant application in biomedical, astrophysics, nondestructive inspection, 6th generation communication system as well as national security application. Here we demonstrate a novel broadband photon-type THz/infrared (IR) photodetector based on the GaAs/AlxGa1-xAs ratchet structure. This kind of photodetector realizes a THz photon-response based on the electrically pumped hot hole injection and overcomes the internal workfunction related spectral response limit. An ultrabroadband photoresponse from 4 THz to 300 THz and a peak responsivity of 50.3 mA/W are realized at negative bias voltage of -1 V. The photodetector also presents a bias-tunable photon-response characteristic due to the asymmetric structure. The ratchet structure also induces an evident photocurrent even at zero bias voltage, which indicates the detector can be regard as a broadband photovoltaic-like detector. The rectification characteristic and high temperature operation possibility of the photodetector are also discussed. This work not only demonstrates a novel ultrabroadband THz/IR photodetector, but also provides a new method to study the light-responsive ratchet.

physics.app-ph↗

arXiv2021

Dynamic Balancing of Humanoid Robot Walker3 with Proprioceptive Actuation: Systematic Design of Algorithm, Software and Hardware

Dynamic balancing under uncertain disturbances is important for a humanoid robot, which requires a good capability of coordinating the entire body redundancy to execute multi tasks. Whole-body control (WBC) based on hierarchical optimization has been generally accepted and utilized in torque-controlled robots. A good hierarchy is the prerequisite for WBC and can be predefined according to prior knowledge. However, the real-time computation would be problematic in the physical applications considering the computational complexity of WBC. For robots with proprioceptive actuation, the joint friction in gear reducer would also degrade the torque tracking performance. In our paper, a reasonable hierarchy of tasks and constraints is first customized for robot dynamic balancing. Then a real-time WBC is implemented via a computationally efficient WBC software. Such a method is solved on a modular master control system UBTMaster characterized by the real-time communication and powerful computing capability. After the joint friction being well covered by the model identification, extensive experiments on various balancing scenarios are conducted on a humanoid Walker3 with proprioceptive actuation. The robot shows an outstanding balance performance even under external impulses as well as the two feet of the robot suffering the inclination and shift disturbances independently. The results demonstrate that with the strict hierarchy, real-time computation and joint friction being handled carefully, the robot with proprioceptive actuation can manage the dynamic physical interactions with the unstructured environments well.

cs.RO↗