arXiv subjects

Yutong Lu

Publications and source records attributed to Yutong Lu.

At least 55 records · Page 3Linked to original sources

arXiv2025

Solving the Catastrophic Forgetting Problem in Generalized Category Discovery

Generalized Category Discovery (GCD) aims to identify a mix of known and novel categories within unlabeled data sets, providing a more realistic setting for image recognition. Essentially, GCD needs to remember existing patterns thoroughly to recognize novel categories. Recent state-of-the-art method SimGCD transfers the knowledge from known-class data to the learning of novel classes through debiased learning. However, some patterns are catastrophically forgot during adaptation and thus lead to poor performance in novel categories classification. To address this issue, we propose a novel learning approach, LegoGCD, which is seamlessly integrated into previous methods to enhance the discrimination of novel classes while maintaining performance on previously encountered known classes. Specifically, we design two types of techniques termed as Local Entropy Regularization (LER) and Dual-views Kullback Leibler divergence constraint (DKL). The LER optimizes the distribution of potential known class samples in unlabeled data, thus ensuring the preservation of knowledge related to known categories while learning novel classes. Meanwhile, DKL introduces Kullback Leibler divergence to encourage the model to produce a similar prediction distribution of two view samples from the same image. In this way, it successfully avoids mismatched prediction and generates more reliable potential known class samples simultaneously. Extensive experiments validate that the proposed LegoGCD effectively addresses the known category forgetting issue across all datasets, eg, delivering a 7.74% and 2.51% accuracy boost on known and novel classes in CUB, respectively. Our code is available at: https://github.com/Cliffia123/LegoGCD.

cs.CV↗

arXiv2025

CL-MFAP: A Contrastive Learning-Based Multimodal Foundation Model for Molecular Property Prediction and Antibiotic Screening

Due to the rise in antimicrobial resistance, identifying novel compounds with antibiotic potential is crucial for combatting this global health issue. However, traditional drug development methods are costly and inefficient. Recognizing the pressing need for more effective solutions, researchers have turned to machine learning techniques to streamline the prediction and development of novel antibiotic compounds. While foundation models have shown promise in antibiotic discovery, current mainstream efforts still fall short of fully leveraging the potential of multimodal molecular data. Recent studies suggest that contrastive learning frameworks utilizing multimodal data exhibit excellent performance in representation learning across various domains. Building upon this, we introduce CL-MFAP, an unsupervised contrastive learning (CL)-based multimodal foundation (MF) model specifically tailored for discovering small molecules with potential antibiotic properties (AP) using three types of molecular data. This model employs 1.6 million bioactive molecules with drug-like properties from the ChEMBL dataset to jointly pretrain three encoders: (1) a transformer-based encoder with rotary position embedding for processing SMILES strings; (2) another transformer-based encoder, incorporating a novel bi-level routing attention mechanism to handle molecular graph representations; and (3) a Morgan fingerprint encoder using a multilayer perceptron, to achieve the contrastive learning purpose. The CL-MFAP outperforms baseline models in antibiotic property prediction by effectively utilizing different molecular modalities and demonstrates superior domain-specific performance when fine-tuned for antibiotic-related property prediction tasks.

q-bio.BM↗

arXiv2025

MolGraph-xLSTM: A graph-based dual-level xLSTM framework with multi-head mixture-of-experts for enhanced molecular representation and interpretability

Predicting molecular properties is essential for drug discovery, and computational methods can greatly enhance this process. Molecular graphs have become a focus for representation learning, with Graph Neural Networks (GNNs) widely used. However, GNNs often struggle with capturing long-range dependencies. To address this, we propose MolGraph-xLSTM, a novel graph-based xLSTM model that enhances feature extraction and effectively models molecule long-range interactions. Our approach processes molecular graphs at two scales: atom-level and motif-level. For atom-level graphs, a GNN-based xLSTM framework with jumping knowledge extracts local features and aggregates multilayer information to capture both local and global patterns effectively. Motif-level graphs provide complementary structural information for a broader molecular view. Embeddings from both scales are refined via a multi-head mixture of experts (MHMoE), further enhancing expressiveness and performance. We validate MolGraph-xLSTM on 10 molecular property prediction datasets, covering both classification and regression tasks. Our model demonstrates consistent performance across all datasets, with improvements of up to 7.03% on the BBBP dataset for classification and 7.54% on the ESOL dataset for regression compared to baselines. On average, MolGraph-xLSTM achieves an AUROC improvement of 3.18\% for classification tasks and an RMSE reduction of 3.83\% across regression datasets compared to the baseline methods. These results confirm the effectiveness of our model, offering a promising solution for molecular representation learning for drug discovery.

cs.LG↗

arXiv2024

Sparkles: Unlocking Chats Across Multiple Images for Multimodal Instruction-Following Models

Large language models exhibit enhanced zero-shot performance on various tasks when fine-tuned with instruction-following data. Multimodal instruction-following models extend these capabilities by integrating both text and images. However, existing models such as MiniGPT-4 and LLaVA face challenges in maintaining dialogue coherence in scenarios involving multiple images. A primary reason is the lack of a specialized dataset for this critical application. To bridge these gaps, we introduce SparklesDialogue, the first machine-generated dialogue dataset tailored for word-level interleaved multi-image and text interactions. Furthermore, we construct SparklesEval, a GPT-assisted benchmark for quantitatively assessing a model's conversational competence across multiple images and dialogue turns. We then present SparklesChat, a multimodal instruction-following model for open-ended dialogues across multiple images. Our experiments validate the effectiveness of training SparklesChat with SparklesDialogue based on MiniGPT-4 and LLaVA-v1.5, which enhances comprehension across multiple images and dialogue turns, and does not compromise single-image understanding capabilities. Qualitative evaluations further demonstrate SparklesChat's generality in handling real-world applications. All resources related to this study are publicly available at https://github.com/HYPJUDY/Sparkles.

cs.CV↗

arXiv2024

Causality-Inspired Models for Financial Time Series Forecasting

We introduce a novel framework to financial time series forecasting that leverages causality-inspired models to balance the trade-off between invariance to distributional changes and minimization of prediction errors. To the best of our knowledge, this is the first study to conduct a comprehensive comparative analysis among state-of-the-art causal discovery algorithms, benchmarked against non-causal feature selection techniques, in the application of forecasting asset returns. Empirical evaluations demonstrate the efficacy of our approach in yielding stable and accurate predictions, outperforming baseline models, particularly in tumultuous market conditions.

q-fin.CP↗

arXiv2024

UNR: Unified Notifiable RMA Library for HPC

Remote Memory Access (RMA) enables direct access to remote memory to achieve high performance for HPC applications. However, most modern parallel programming models lack schemes for the remote process to detect the completion of RMA operations. Many previous works have proposed programming models and extensions to notify the communication peer, but they did not solve the multi-NIC aggregation, portability, hardware-software co-design, and usability problems. In this work, we proposed a Unified Notifiable RMA (UNR) library for HPC to address these challenges. In addition, we demonstrate the best practice of utilizing UNR within a real-world scientific application, PowerLLEL. We deployed UNR across four HPC systems, each with a different interconnect. The results show that PowerLLEL powered by UNR achieves up to a 36% acceleration on 1728 nodes of the Tianhe-Xingyi supercomputing system.

cs.DC↗

arXiv2024

Solving Partial Differential Equations in Different Domains by Operator Learning method Based on Boundary Integral Equations

This article explores operator learning models that can deduce solutions to partial differential equations (PDEs) on arbitrary domains without requiring retraining. We introduce two innovative models rooted in boundary integral equations (BIEs): the Boundary Integral Type Deep Operator Network (BI-DeepONet) and the Boundary Integral Trigonometric Deep Operator Neural Network (BI-TDONet), which are crafted to address PDEs across diverse domains. Once fully trained, these BIE-based models adeptly predict the solutions of PDEs in any domain without the need for additional training. BI-TDONet notably enhances its performance by employing the singular value decomposition (SVD) of bounded linear operators, allowing for the efficient distribution of input functions across its modules. Furthermore, to tackle the issue of function sampling values that do not effectively capture oscillatory and impulse signal characteristics, trigonometric coefficients are utilized as both inputs and outputs in BI-TDONet. Our numerical experiments robustly support and confirm the efficacy of this theoretical framework.

math-ph↗

arXiv2024

Galaxy: A Resource-Efficient Collaborative Edge AI System for In-situ Transformer Inference

Transformer-based models have unlocked a plethora of powerful intelligent applications at the edge, such as voice assistant in smart home. Traditional deployment approaches offload the inference workloads to the remote cloud server, which would induce substantial pressure on the backbone network as well as raise users' privacy concerns. To address that, in-situ inference has been recently recognized for edge intelligence, but it still confronts significant challenges stemming from the conflict between intensive workloads and limited on-device computing resources. In this paper, we leverage our observation that many edge environments usually comprise a rich set of accompanying trusted edge devices with idle resources and propose Galaxy, a collaborative edge AI system that breaks the resource walls across heterogeneous edge devices for efficient Transformer inference acceleration. Galaxy introduces a novel hybrid model parallelism to orchestrate collaborative inference, along with a heterogeneity-aware parallelism planning for fully exploiting the resource potential. Furthermore, Galaxy devises a tile-based fine-grained overlapping of communication and computation to mitigate the impact of tensor synchronizations on inference latency under bandwidth-constrained edge environments. Extensive evaluation based on prototype implementation demonstrates that Galaxy remarkably outperforms state-of-the-art approaches under various edge environment setups, achieving up to 2.5x end-to-end latency reduction.

cs.DC↗

arXiv2024

Co-trading networks for modeling dynamic interdependency structures and estimating high-dimensional covariances in US equity markets

The time proximity of trades across stocks reveals interesting topological structures of the equity market in the United States. In this article, we investigate how such concurrent cross-stock trading behaviors, which we denote as co-trading, shape the market structures and affect stock price co-movements. By leveraging a co-trading-based pairwise similarity measure, we propose a novel method to construct dynamic networks of stocks. Our empirical studies employ high-frequency limit order book data from 2017-01-03 to 2019-12-09. By applying spectral clustering on co-trading networks, we uncover economically meaningful clusters of stocks. Beyond the static Global Industry Classification Standard (GICS) sectors, our data-driven clusters capture the time evolution of the dependency among stocks. Furthermore, we demonstrate statistically significant positive relations between low-latency co-trading and return covariance. With the aid of co-trading networks, we develop a robust estimator for high-dimensional covariance matrix, which yields superior economic value on portfolio allocation. The mean-variance portfolios based on our covariance estimates achieve both lower volatility and higher Sharpe ratios than standard benchmarks.

q-fin.TR↗

arXiv2024

Trade Co-occurrence, Trade Flow Decomposition, and Conditional Order Imbalance in Equity Markets

The time proximity of high-frequency trades can contain a salient signal. In this paper, we propose a method to classify every trade, based on its proximity with other trades in the market within a short period of time, into five types. By means of a suitably defined normalized order imbalance associated to each type of trade, which we denote as conditional order imbalance (COI), we investigate the price impact of the decomposed trade flows. Our empirical findings indicate strong positive correlations between contemporaneous returns and COIs. In terms of predictability, we document that associations with future returns are positive for COIs of trades which are isolated from trades of stocks other than themselves, and negative otherwise. Furthermore, trading strategies which we develop using COIs achieve conspicuous returns and Sharpe ratios, in an extensive experimental setup on a universe of 457 stocks using daily data for a period of four years.

q-fin.TR↗

arXiv2024

Crystal Structure Prediction by Joint Equivariant Diffusion

Crystal Structure Prediction (CSP) is crucial in various scientific disciplines. While CSP can be addressed by employing currently-prevailing generative models (e.g. diffusion models), this task encounters unique challenges owing to the symmetric geometry of crystal structures -- the invariance of translation, rotation, and periodicity. To incorporate the above symmetries, this paper proposes DiffCSP, a novel diffusion model to learn the structure distribution from stable crystals. To be specific, DiffCSP jointly generates the lattice and atom coordinates for each crystal by employing a periodic-E(3)-equivariant denoising model, to better model the crystal geometry. Notably, different from related equivariant generative approaches, DiffCSP leverages fractional coordinates other than Cartesian coordinates to represent crystals, remarkably promoting the diffusion and the generation process of atom positions. Extensive experiments verify that our DiffCSP significantly outperforms existing CSP methods, with a much lower computation cost in contrast to DFT-based methods. Moreover, the superiority of DiffCSP is also observed when it is extended for ab initio crystal generation.

cond-mat.mtrl-sci↗

arXiv2024

Intensive Vision-guided Network for Radiology Report Generation

Automatic radiology report generation is booming due to its huge application potential for the healthcare industry. However, existing computer vision and natural language processing approaches to tackle this problem are limited in two aspects. First, when extracting image features, most of them neglect multi-view reasoning in vision and model single-view structure of medical images, such as space-view or channel-view. However, clinicians rely on multi-view imaging information for comprehensive judgment in daily clinical diagnosis. Second, when generating reports, they overlook context reasoning with multi-modal information and focus on pure textual optimization utilizing retrieval-based methods. We aim to address these two issues by proposing a model that better simulates clinicians' perspectives and generates more accurate reports. Given the above limitation in feature extraction, we propose a Globally-intensive Attention (GIA) module in the medical image encoder to simulate and integrate multi-view vision perception. GIA aims to learn three types of vision perception: depth view, space view, and pixel view. On the other hand, to address the above problem in report generation, we explore how to involve multi-modal signals to generate precisely matched reports, i.e., how to integrate previously predicted words with region-aware visual content in next word prediction. Specifically, we design a Visual Knowledge-guided Decoder (VKGD), which can adaptively consider how much the model needs to rely on visual information and previously predicted text to assist next word prediction. Hence, our final Intensive Vision-guided Network (IVGN) framework includes a GIA-guided Visual Encoder and the VKGD. Experiments on two commonly-used datasets IU X-Ray and MIMIC-CXR demonstrate the superior ability of our method compared with other state-of-the-art approaches.

cs.CV↗

arXiv2024

Exploring Low-Resource Medical Image Classification with Weakly Supervised Prompt Learning

Most advances in medical image recognition supporting clinical auxiliary diagnosis meet challenges due to the low-resource situation in the medical field, where annotations are highly expensive and professional. This low-resource problem can be alleviated by leveraging the transferable representations of large-scale pre-trained vision-language models via relevant medical text prompts. However, existing pre-trained vision-language models require domain experts to carefully design the medical prompts, which greatly increases the burden on clinicians. To address this problem, we propose a weakly supervised prompt learning method MedPrompt to automatically generate medical prompts, which includes an unsupervised pre-trained vision-language model and a weakly supervised prompt learning model. The unsupervised pre-trained vision-language model utilizes the natural correlation between medical images and corresponding medical texts for pre-training, without any manual annotations. The weakly supervised prompt learning model only utilizes the classes of images in the dataset to guide the learning of the specific class vector in the prompt, while the learning of other context vectors in the prompt requires no manual annotations for guidance. To the best of our knowledge, this is the first model to automatically generate medical prompts. With these prompts, the pre-trained vision-language model can be freed from the strong expert dependency of manual annotation and manual prompt design. Experimental results show that the model using our automatically generated prompts outperforms its full-shot learning hand-crafted prompts counterparts with only a minimal number of labeled samples for few-shot learning, and reaches superior or comparable accuracy on zero-shot image classification. The proposed prompt generator is lightweight and therefore can be embedded into any network architecture.

cs.CV↗

arXiv2023

Searching High Temperature Superconductors with the assistance of Graph Neural Networks

Predicting high temperature superconductors has long been a great challenge. A major difficulty is how to predict the transition temperature Tc of superconductors. Recently, progress in material informatics has led to a number of machine learning models predicting Tc, which greatly improves the efficiency of prediction. Unfortunately, prevailing models have not shown adequate physical rationality and generalization ability to find new high temperature superconductors, yet. In this work, in order to give a trustable prediction on the unexplored materials, we built a bond-sensitive graph neural network (BSGNN), which is optimized to process the information of chemical bond and electron interaction in the crystal lattice, to predict the Tc maximum of each type of superconducting materials. On the basis of the domain knowledge considered in the data preparation and algorithm design, our model revealed a relevance between the Tc-Tc maximum and chemical bonds. The results indicate that shorter bond length is favored by high Tc, which is in accordance with previous human experience. Moreover, it also shows that some specific chemical elements are favored by high Tc, which is beyond what human experts already knew. It gives a convenient guidance for searching high temperature superconductors in materials database, by ruling out the materials that could never have high Tc.

cond-mat.supr-con↗

arXiv2023

PSLT: A Light-weight Vision Transformer with Ladder Self-Attention and Progressive Shift

Vision Transformer (ViT) has shown great potential for various visual tasks due to its ability to model long-range dependency. However, ViT requires a large amount of computing resource to compute the global self-attention. In this work, we propose a ladder self-attention block with multiple branches and a progressive shift mechanism to develop a light-weight transformer backbone that requires less computing resources (e.g. a relatively small number of parameters and FLOPs), termed Progressive Shift Ladder Transformer (PSLT). First, the ladder self-attention block reduces the computational cost by modelling local self-attention in each branch. In the meanwhile, the progressive shift mechanism is proposed to enlarge the receptive field in the ladder self-attention block by modelling diverse local self-attention for each branch and interacting among these branches. Second, the input feature of the ladder self-attention block is split equally along the channel dimension for each branch, which considerably reduces the computational cost in the ladder self-attention block (with nearly 1/3 the amount of parameters and FLOPs), and the outputs of these branches are then collaborated by a pixel-adaptive fusion. Therefore, the ladder self-attention block with a relatively small number of parameters and FLOPs is capable of modelling long-range interactions. Based on the ladder self-attention block, PSLT performs well on several vision tasks, including image classification, objection detection and person re-identification. On the ImageNet-1k dataset, PSLT achieves a top-1 accuracy of 79.9% with 9.2M parameters and 1.9G FLOPs, which is comparable to several existing models with more than 20M parameters and 4G FLOPs. Code is available at https://isee-ai.cn/wugaojie/PSLT.html.

cs.CV↗

arXiv2023

Company Competition Graph

Financial market participants frequently rely on numerous business relationships to make investment decisions. Investors can learn about potential risks and opportunities associated with other connected entities through these corporate connections. Nonetheless, human annotation of a large corpus to extract such relationships is highly time-consuming, not to mention that it requires a considerable amount of industry expertise and professional training. Meanwhile, we have yet to observe means to generate reliable knowledge graphs of corporate relationships due to the lack of impartial and granular data sources. This study proposes a system to process financial reports and construct the public competitor graph to fill the void. Our method can retrieve more than 83\% competition relationship of the S\&P 500 index companies. Based on the output from our system, we construct a knowledge graph with more than 700 nodes and 1200 edges. A demo interactive graph interface is available.

q-fin.GN↗

arXiv2023

Form 10-K Itemization

Form 10-K report is a financial report disclosing the annual financial state of a public company. It is an important evidence to conduct financial analysis, i.e., asset pricing, corporate finance. Practitioners and researchers are constantly designing algorithms to better conduct analysis on information in the Form 10-K report. The vast majority of previous works focus on quantitative data. With recent advancement on natural language processing (NLP), textual data in financial filing attracts more attention. However, to incorporate textual data for analyzing, Form 10-K Itemization is a necessary pre-process step. It aims to segment the whole document into several Item sections, where each Item section focuses on a specific financial aspect of the company. With the segmented Item sections, NLP techniques can directly apply on those Item sections related to downstream tasks. In this paper, we develop a Form 10-K Itemization system which can automatically segment all the Item sections in 10-K documents. The system is both effective and efficient. It reaches a retrieval rate of 93%.

cs.IR↗

arXiv2022

SAIH: A Scalable Evaluation Methodology for Understanding AI Performance Trend on HPC Systems

Novel artificial intelligence (AI) technology has expedited various scientific research, e.g., cosmology, physics and bioinformatics, inevitably becoming a significant category of workload on high performance computing (HPC) systems. Existing AI benchmarks tend to customize well-recognized AI applications, so as to evaluate the AI performance of HPC systems under predefined problem size, in terms of datasets and AI models. Due to lack of scalability on the problem size, static AI benchmarks might be under competent to help understand the performance trend of evolving AI applications on HPC systems, in particular, the scientific AI applications on large-scale systems. In this paper, we propose a scalable evaluation methodology (SAIH) for analyzing the AI performance trend of HPC systems with scaling the problem sizes of customized AI applications. To enable scalability, SAIH builds a set of novel mechanisms for augmenting problem sizes. As the data and model constantly scale, we can investigate the trend and range of AI performance on HPC systems, and further diagnose system bottlenecks. To verify our methodology, we augment a cosmological AI application to evaluate a real HPC system equipped with GPUs as a case study of SAIH.

cs.DC↗