arXiv subjects

Xin Liu

Publications and source records attributed to Xin Liu.

At least 523 records · Page 29Linked to original sources

arXiv2024

Towards Subgraph Isomorphism Counting with Graph Kernels

Subgraph isomorphism counting is known as #P-complete and requires exponential time to find the accurate solution. Utilizing representation learning has been shown as a promising direction to represent substructures and approximate the solution. Graph kernels that implicitly capture the correlations among substructures in diverse graphs have exhibited great discriminative power in graph classification, so we pioneeringly investigate their potential in counting subgraph isomorphisms and further explore the augmentation of kernel capability through various variants, including polynomial and Gaussian kernels. Through comprehensive analysis, we enhance the graph kernels by incorporating neighborhood information. Finally, we present the results of extensive experiments to demonstrate the effectiveness of the enhanced graph kernels and discuss promising directions for future research.

cs.LG↗

arXiv2024

Disttack: Graph Adversarial Attacks Toward Distributed GNN Training

Graph Neural Networks (GNNs) have emerged as potent models for graph learning. Distributing the training process across multiple computing nodes is the most promising solution to address the challenges of ever-growing real-world graphs. However, current adversarial attack methods on GNNs neglect the characteristics and applications of the distributed scenario, leading to suboptimal performance and inefficiency in attacking distributed GNN training. In this study, we introduce Disttack, the first framework of adversarial attacks for distributed GNN training that leverages the characteristics of frequent gradient updates in a distributed system. Specifically, Disttack corrupts distributed GNN training by injecting adversarial attacks into one single computing node. The attacked subgraphs are precisely perturbed to induce an abnormal gradient ascent in backpropagation, disrupting gradient synchronization between computing nodes and thus leading to a significant performance decline of the trained GNN. We evaluate Disttack on four large real-world graphs by attacking five widely adopted GNNs. Compared with the state-of-the-art attack method, experimental results demonstrate that Disttack amplifies the model accuracy degradation by 2.75$\times$ and achieves speedup by 17.33$\times$ on average while maintaining unnoticeability.

cs.LG↗

arXiv2024

Provable Acceleration of Nesterov's Accelerated Gradient Method over Heavy Ball Method in Training Over-Parameterized Neural Networks

Due to its simplicity and efficiency, the first-order gradient method has been extensively employed in training neural networks. Although the optimization problem of the neural network is non-convex, recent research has proved that the first-order method is capable of attaining a global minimum during training over-parameterized neural networks, where the number of parameters is significantly larger than that of training instances. Momentum methods, including the heavy ball (HB) method and Nesterov's accelerated gradient (NAG) method, are the workhorse of first-order gradient methods owning to their accelerated convergence. In practice, NAG often exhibits superior performance than HB. However, current theoretical works fail to distinguish their convergence difference in training neural networks. To fill this gap, we consider the training problem of the two-layer ReLU neural network under over-parameterization and random initialization. Leveraging high-resolution dynamical systems and neural tangent kernel (NTK) theory, our result not only establishes tighter upper bounds of the convergence rate for both HB and NAG, but also provides the first theoretical guarantee for the acceleration of NAG over HB in training neural networks. Finally, we validate our theoretical results on three benchmark datasets.

cs.LG↗

arXiv2024

Projected gradient descent algorithm for $\textit{ab initio}$ crystal structure relaxation under a fixed unit cell volume

This paper is concerned with $\textit{ab initio}$ crystal structure relaxation under a fixed unit cell volume, which is a step in calculating the static equations of state and forms the basis of thermodynamic property calculations for materials. The task can be formulated as an energy minimization with a determinant constraint. Widely used line minimization-based methods (e.g., conjugate gradient method) lack both efficiency and convergence guarantees due to the nonconvex nature of the feasible region as well as the significant differences in the curvatures of the potential energy surface with respect to atomic and lattice components. To this end, we propose a projected gradient descent algorithm named PANBB. It is equipped with (i) search direction projections onto the tangent spaces of the nonconvex feasible region for lattice vectors, (ii) distinct curvature-aware initial trial step sizes for atomic and lattice updates, and (iii) a nonrestrictive line minimization criterion as the stopping rule for the inner loop. It can be proved that PANBB favors theoretical convergence to equilibrium states. Across a benchmark set containing 223 structures from various categories, PANBB achieves average speedup factors of approximately 1.41 and 1.45 over the conjugate gradient method and direct inversion in the iterative subspace implemented in off-the-shelf simulation software, respectively. Moreover, it normally converges on all the systems, manifesting its unparalleled robustness. As an application, we calculate the static equations of state for the high-entropy alloy AlCoCrFeNi, which remains elusive owing to 160 atoms representing both chemical and magnetic disorder and the strong local lattice distortion. The results are consistent with the previous calculations and are further validated by experimental thermodynamic data.

physics.comp-ph↗

arXiv2024

Enhancing Micro Gesture Recognition for Emotion Understanding via Context-aware Visual-Text Contrastive Learning

Psychological studies have shown that Micro Gestures (MG) are closely linked to human emotions. MG-based emotion understanding has attracted much attention because it allows for emotion understanding through nonverbal body gestures without relying on identity information (e.g., facial and electrocardiogram data). Therefore, it is essential to recognize MG effectively for advanced emotion understanding. However, existing Micro Gesture Recognition (MGR) methods utilize only a single modality (e.g., RGB or skeleton) while overlooking crucial textual information. In this letter, we propose a simple but effective visual-text contrastive learning solution that utilizes text information for MGR. In addition, instead of using handcrafted prompts for visual-text contrastive learning, we propose a novel module called Adaptive prompting to generate context-aware prompts. The experimental results show that the proposed method achieves state-of-the-art performance on two public datasets. Furthermore, based on an empirical study utilizing the results of MGR for emotion understanding, we demonstrate that using the textual results of MGR significantly improves performance by 6%+ compared to directly using video as input.

cs.CV↗

arXiv2024

Exploring Changing-look Active Galactic Nuclei with the Sloan Digital Sky Survey V: First Year Results

"Changing-look" active galactic nuclei (CL-AGNs) challenge our basic ideas about the physics of accretion flows and circumnuclear gas around supermassive black holes. Using first-year Sloan Digital Sky Survey V (SDSS-V) repeated spectroscopy of nearly 29,000 previously known AGNs, combined with dedicated follow-up spectroscopy, and publicly available optical light curves, we have identified 116 CL-AGNs where (at least) one broad emission line has essentially (dis-)appeared, as well as 88 other extremely variable systems. Our CL-AGN sample, with 107 newly identified cases, is the largest reported to date, and includes $\sim0.4\%$ of the AGNs reobserved in first-year SDSS-V operations. Among our CL-AGNs, 67% exhibit dimming while 33% exhibit brightening. Our sample probes extreme AGN spectral variability on months to decades timescales, including some cases of recurring transitions on surprisingly short timescales ($\lesssim 2$ months in the rest frame). We find that CL events are preferentially found in lower-Eddington-ratio ($f_{Edd}$) systems: Our CL-AGNs have a $f_{Edd}$ distribution that significantly differs from that of a carefully constructed, redshift- and luminosity-matched control sample (Anderson-Darling test yielding $p_{\rm AD}\approx 6\times10^{-5}$; median $f_{Edd}\approx0.025$ vs. $0.043$). This preference for low $f_{Edd}$ strengthens previous findings of higher CL-AGN incidence at lower $f_{Edd}$, found in smaller samples. Finally, we show that the broad MgII emission line in our CL-AGN sample tends to vary significantly less than the broad H$β$ emission line. Our large CL-AGN sample demonstrates the advantages and challenges in using multi-epoch spectroscopy from large surveys to study extreme AGN variability and physics.

astro-ph.GA↗

arXiv2024

RAGCache: Efficient Knowledge Caching for Retrieval-Augmented Generation

Retrieval-Augmented Generation (RAG) has shown significant improvements in various natural language processing tasks by integrating the strengths of large language models (LLMs) and external knowledge databases. However, RAG introduces long sequence generation and leads to high computation and memory costs. We propose RAGCache, a novel multilevel dynamic caching system tailored for RAG. Our analysis benchmarks current RAG systems, pinpointing the performance bottleneck (i.e., long sequence due to knowledge injection) and optimization opportunities (i.e., caching knowledge's intermediate states). Based on these insights, we design RAGCache, which organizes the intermediate states of retrieved knowledge in a knowledge tree and caches them in the GPU and host memory hierarchy. RAGCache proposes a replacement policy that is aware of LLM inference characteristics and RAG retrieval patterns. It also dynamically overlaps the retrieval and inference steps to minimize the end-to-end latency. We implement RAGCache and evaluate it on vLLM, a state-of-the-art LLM inference system and Faiss, a state-of-the-art vector database. The experimental results show that RAGCache reduces the time to first token (TTFT) by up to 4x and improves the throughput by up to 2.1x compared to vLLM integrated with Faiss.

cs.DC↗

arXiv2024

Real-Time 4K Super-Resolution of Compressed AVIF Images. AIS 2024 Challenge Survey

This paper introduces a novel benchmark as part of the AIS 2024 Real-Time Image Super-Resolution (RTSR) Challenge, which aims to upscale compressed images from 540p to 4K resolution (4x factor) in real-time on commercial GPUs. For this, we use a diverse test set containing a variety of 4K images ranging from digital art to gaming and photography. The images are compressed using the modern AVIF codec, instead of JPEG. All the proposed methods improve PSNR fidelity over Lanczos interpolation, and process images under 10ms. Out of the 160 participants, 25 teams submitted their code and models. The solutions present novel designs tailored for memory-efficiency and runtime on edge devices. This survey describes the best solutions for real-time SR of compressed high-resolution images.

cs.CV↗

arXiv2024

A cost-efficient quantum access network with qubit-based synchronization

Quantum Key Distribution (QKD) is a physical layer encryption technique that enables two distant parties to exchange secure keys with information-theoretic security. In the last two decades, QKD has transitioned from laboratory research to real-world applications, including multi-user quantum access networks (QANs). This network structure allows users to share a single-photon detector at a network node through time-division multiplexing, thereby significantly reducing the network cost. However, current QAN implementations require additional hardware for auxiliary tasks such as time synchronization. To address this issue, we propose a cost-efficient QAN that uses qubit-based synchronization. In this approach, the transmitted qubits facilitate time synchronization, eliminating the need for additional synchronization hardware. We tested our scheme by implementing a network for two users and successfully achieved average secure key rates of $53.84$ kbps and $71.90$ kbps for each user over a 50-km commercial fiber spool. In addition, we investigated the capacity of the access network under cross-talk and loss conditions. The simulation results demonstrate that this scheme can support a QAN with 64 users with key rates up to 1070~bps. Our work provides a feasible and cost-effective way to implement a multi-user QKD network, further promoting the widespread application of QKD.

quant-ph↗

arXiv2024

SPIRiT-Diffusion: Self-Consistency Driven Diffusion Model for Accelerated MRI

Diffusion models have emerged as a leading methodology for image generation and have proven successful in the realm of magnetic resonance imaging (MRI) reconstruction. However, existing reconstruction methods based on diffusion models are primarily formulated in the image domain, making the reconstruction quality susceptible to inaccuracies in coil sensitivity maps (CSMs). k-space interpolation methods can effectively address this issue but conventional diffusion models are not readily applicable in k-space interpolation. To overcome this challenge, we introduce a novel approach called SPIRiT-Diffusion, which is a diffusion model for k-space interpolation inspired by the iterative self-consistent SPIRiT method. Specifically, we utilize the iterative solver of the self-consistent term (i.e., k-space physical prior) in SPIRiT to formulate a novel stochastic differential equation (SDE) governing the diffusion process. Subsequently, k-space data can be interpolated by executing the diffusion process. This innovative approach highlights the optimization model's role in designing the SDE in diffusion models, enabling the diffusion process to align closely with the physics inherent in the optimization model, a concept referred to as model-driven diffusion. We evaluated the proposed SPIRiT-Diffusion method using a 3D joint intracranial and carotid vessel wall imaging dataset. The results convincingly demonstrate its superiority over image-domain reconstruction methods, achieving high reconstruction quality even at a substantial acceleration rate of 10.

cs.CV↗

arXiv2024

Predicting Emergent Abilities with Infinite Resolution Evaluation

The scientific scale-up of large language models (LLMs) necessitates a comprehensive understanding of their scaling properties. However, the existing literature on the scaling properties only yields an incomplete answer: optimization loss decreases predictably as the model size increases, in line with established scaling law; yet no scaling law for task has been established and the task performances are far from predictable during scaling. Task performances typically show minor gains on small models until they improve dramatically once models exceed a size threshold, exemplifying the ``emergent abilities''. In this study, we discover that small models, although they exhibit minor performance, demonstrate critical and consistent task performance improvements that are not captured by conventional evaluation strategies due to insufficient measurement resolution. To measure such improvements, we introduce PassUntil, an evaluation strategy with theoretically infinite resolution, through massive sampling in the decoding phase. With PassUntil, we conduct a quantitative investigation into the scaling law of task performance. The investigation contains two parts. Firstly, a strict task scaling law that is not conventionally known to exist, is identified, enhancing the predictability of task performances. Remarkably, we are able to predict the performance of the 2.4B model on code generation with merely 0.05\% deviation before training starts, which is the first systematic attempt to verify predictable scaling proposed by GPT-4's report. Secondly, we are able to study emergent abilities quantitatively. We identify a kind of accelerated emergence whose scaling curve cannot be fitted by standard scaling law function and has a increasing speed. We then examine two hypothesis and imply that the ``multiple circuits hypothesis'' might be responsible for the accelerated emergence.

cs.CL↗

arXiv2024

Kilometer-Level Coupled Modeling Using 40 Million Cores: An Eight-Year Journey of Model Development

With current and future leading systems adopting heterogeneous architectures, adapting existing models for heterogeneous supercomputers is of urgent need for improving model resolution and reducing modeling uncertainty. This paper presents our three-week effort on porting a complex earth system model, CESM 2.2, to a 40-million-core Sunway supercomputer. Taking a non-intrusive approach that tries to minimizes manual code modifications, our project tries to achieve both improvement of performance and consistency of the model code. By using a hierarchical grid system and an OpenMP-based offloading toolkit, our porting and parallelization effort covers over 80% of the code, and achieves a simulation speed of 340 SDPD (simulated days per day) for 5-km atmosphere, 265 SDPD for 3-km ocean, and 222 SDPD for a coupled model, thus making multi-year or even multi-decadal experiments at such high resolution possible.

cs.DC↗

arXiv2024

DEGNN: Dual Experts Graph Neural Network Handling Both Edge and Node Feature Noise

Graph Neural Networks (GNNs) have achieved notable success in various applications over graph data. However, recent research has revealed that real-world graphs often contain noise, and GNNs are susceptible to noise in the graph. To address this issue, several Graph Structure Learning (GSL) models have been introduced. While GSL models are tailored to enhance robustness against edge noise through edge reconstruction, a significant limitation surfaces: their high reliance on node features. This inherent dependence amplifies their susceptibility to noise within node features. Recognizing this vulnerability, we present DEGNN, a novel GNN model designed to adeptly mitigate noise in both edges and node features. The core idea of DEGNN is to design two separate experts: an edge expert and a node feature expert. These experts utilize self-supervised learning techniques to produce modified edges and node features. Leveraging these modified representations, DEGNN subsequently addresses downstream tasks, ensuring robustness against noise present in both edges and node features of real-world graphs. Notably, the modification process can be trained end-to-end, empowering DEGNN to adjust dynamically and achieves optimal edge and node representations for specific tasks. Comprehensive experiments demonstrate DEGNN's efficacy in managing noise, both in original real-world graphs and in graphs with synthetic noise.

cs.LG↗

arXiv2024

A Framework for Simulating the Path-level Residual Stress in the Laser Powder Bed Fusion Process

Laser Powder Bed Fusion (LPBF) additive manufacturing has revolutionized industries with its capability to create intricate and customized components. The LPBF process uses moving heat sources to melt and solidify metal powders. The fast melting and cooling leads to residual stress, which critically affects the part quality. Currently, the computational intensity of accurately simulating the residual stress on the path scale remains a significant challenge, limiting our understanding of the LPBF processes. This paper presents a framework for simulating the LPBF process residual stress based on the path-level thermal history. Compared with the existing approaches, the path-level simulation requires discretization only to capture the scanning path rather than the details of the melt pools, thus requiring less dense mesh and is more computationally efficient. We develop this framework by introducing a new concept termed effective thermal strain to capture the anisotropic thermal strain near and around the melt pool. We validate our approach with the high-fidelity results from the literature. We use the proposed approach to simulate various single-island scanning patterns and layers with multiple full and trimmed islands. We further investigate the influence of the path-level thermal history and the layer shape on the residual stress by analyzing their simulation results.

cs.CE↗

arXiv2024

Simultaneous Localization and Recognition of Subwavelength Non-Cooperative Entities Based on SISO Time Reversal and Neural Networks

The simultaneous localization and recognition of subwavelength non-cooperative entities within complex multi-scattering environments using a simplified system continues to pose a substantial challenge. This letter addresses this challenge by synergistically integrating time reversal time-frequency phase prints (TRTFPPs) and neural networks. Initially, a time reversal (TR) single-input single-output (SISO) framework is employed to generate TRTFPPs. To enhance the models' adaptability, particularly in the presence of noise, data augmentation techniques are applied. Subsequently, neural networks are employed to comprehend the TRTFPPs. Specifically, a cascaded neural network structure is embraced, encompassing both a recognition neural network and distinct neural networks for localizing different entities. Through the devised approach, two types of subwavelength entities are successfully identified and precisely localized through numerical simulations and experimental verification in laboratory environment. The proposed methodology holds applicability across various electromagnetic systems, including but not limited to detection, imaging, human-computer interaction, and the Internet of Things (IoT).

physics.app-ph↗

arXiv2024

Recent Advances, Applications, and Open Challenges in Machine Learning for Health: Reflections from Research Roundtables at ML4H 2023 Symposium

The third ML4H symposium was held in person on December 10, 2023, in New Orleans, Louisiana, USA. The symposium included research roundtable sessions to foster discussions between participants and senior researchers on timely and relevant topics for the \ac{ML4H} community. Encouraged by the successful virtual roundtables in the previous year, we organized eleven in-person roundtables and four virtual roundtables at ML4H 2022. The organization of the research roundtables at the conference involved 17 Senior Chairs and 19 Junior Chairs across 11 tables. Each roundtable session included invited senior chairs (with substantial experience in the field), junior chairs (responsible for facilitating the discussion), and attendees from diverse backgrounds with interest in the session's topic. Herein we detail the organization process and compile takeaways from these roundtable discussions, including recent advances, applications, and open challenges for each topic. We conclude with a summary and lessons learned across all roundtables. This document serves as a comprehensive review paper, summarizing the recent advancements in machine learning for healthcare as contributed by foremost researchers in the field.

cs.LG↗

arXiv2024

AbsPyramid: Benchmarking the Abstraction Ability of Language Models with a Unified Entailment Graph

Cognitive research indicates that abstraction ability is essential in human intelligence, which remains under-explored in language models. In this paper, we present AbsPyramid, a unified entailment graph of 221K textual descriptions of abstraction knowledge. While existing resources only touch nouns or verbs within simplified events or specific domains, AbsPyramid collects abstract knowledge for three components of diverse events to comprehensively evaluate the abstraction ability of language models in the open domain. Experimental results demonstrate that current LLMs face challenges comprehending abstraction knowledge in zero-shot and few-shot settings. By training on our rich abstraction knowledge, we find LLMs can acquire basic abstraction abilities and generalize to unseen events. In the meantime, we empirically show that our benchmark is comprehensive to enhance LLMs across two previous abstraction tasks.

cs.CL↗

arXiv2024

Mipha: A Comprehensive Overhaul of Multimodal Assistant with Small Language Models

Multimodal Large Language Models (MLLMs) have showcased impressive skills in tasks related to visual understanding and reasoning. Yet, their widespread application faces obstacles due to the high computational demands during both the training and inference phases, restricting their use to a limited audience within the research and user communities. In this paper, we investigate the design aspects of Multimodal Small Language Models (MSLMs) and propose an efficient multimodal assistant named Mipha, which is designed to create synergy among various aspects: visual representation, language models, and optimization strategies. We show that without increasing the volume of training data, our Mipha-3B outperforms the state-of-the-art large MLLMs, especially LLaVA-1.5-13B, on multiple benchmarks. Through detailed discussion, we provide insights and guidelines for developing strong MSLMs that rival the capabilities of MLLMs. Our code is available at https://github.com/zhuyiche/llava-phi.

cs.CV↗