arXiv subjects

Wei Hu

Publications and source records attributed to Wei Hu.

At least 181 records · Page 10Linked to original sources

arXiv2023

The Law of Parsimony in Gradient Descent for Learning Deep Linear Networks

Over the past few years, an extensively studied phenomenon in training deep networks is the implicit bias of gradient descent towards parsimonious solutions. In this work, we investigate this phenomenon by narrowing our focus to deep linear networks. Through our analysis, we reveal a surprising "law of parsimony" in the learning dynamics when the data possesses low-dimensional structures. Specifically, we show that the evolution of gradient descent starting from orthogonal initialization only affects a minimal portion of singular vector spaces across all weight matrices. In other words, the learning process happens only within a small invariant subspace of each weight matrix, despite the fact that all weight parameters are updated throughout training. This simplicity in learning dynamics could have significant implications for both efficient training and a better understanding of deep networks. First, the analysis enables us to considerably improve training efficiency by taking advantage of the low-dimensional structure in learning dynamics. We can construct smaller, equivalent deep linear networks without sacrificing the benefits associated with the wider counterparts. Second, it allows us to better understand deep representation learning by elucidating the linear progressive separation and concentration of representations from shallow to deep layers. We also conduct numerical experiments to support our theoretical results. The code for our experiments can be found at https://github.com/cjyaras/lawofparsimony.

cs.LG↗

arXiv2023

Serial Contrastive Knowledge Distillation for Continual Few-shot Relation Extraction

Continual few-shot relation extraction (RE) aims to continuously train a model for new relations with few labeled training data, of which the major challenges are the catastrophic forgetting of old relations and the overfitting caused by data sparsity. In this paper, we propose a new model, namely SCKD, to accomplish the continual few-shot RE task. Specifically, we design serial knowledge distillation to preserve the prior knowledge from previous models and conduct contrastive learning with pseudo samples to keep the representations of samples in different relations sufficiently distinguishable. Our experiments on two benchmark datasets validate the effectiveness of SCKD for continual few-shot RE and its superiority in knowledge transfer and memory utilization over state-of-the-art models.

cs.CL↗

arXiv2023

Improving Continual Relation Extraction by Distinguishing Analogous Semantics

Continual relation extraction (RE) aims to learn constantly emerging relations while avoiding forgetting the learned relations. Existing works store a small number of typical samples to re-train the model for alleviating forgetting. However, repeatedly replaying these samples may cause the overfitting problem. We conduct an empirical study on existing works and observe that their performance is severely affected by analogous relations. To address this issue, we propose a novel continual extraction model for analogous relations. Specifically, we design memory-insensitive relation prototypes and memory augmentation to overcome the overfitting problem. We also introduce integrated training and focal knowledge distillation to enhance the performance on analogous relations. Experimental results show the superiority of our model and demonstrate its effectiveness in distinguishing analogous relations and overcoming overfitting.

cs.CL↗

arXiv2023

Using a Bayesian-Inference Approach to Calibrating Models for Simulation in Robotics

In robotics, simulation has the potential to reduce design time and costs, and lead to a more robust engineered solution and a safer development process. However, the use of simulators is predicated on the availability of good models. This contribution is concerned with improving the quality of these models via calibration, which is cast herein in a Bayesian framework. First, we discuss the Bayesian machinery involved in model calibration. Then, we demonstrate it in one example: calibration of a vehicle dynamics model that has low degree of freedom count and can be used for state estimation, model predictive control, or path planning. A high fidelity simulator is used to emulate the ``experiments'' and generate the data for the calibration. The merit of this work is not tied to a new Bayesian methodology for calibration, but to the demonstration of how the Bayesian machinery can establish connections among models in computational dynamics, even when the data in use is noisy. The software used to generate the results reported herein is available in a public repository for unfettered use and distribution.

cs.RO↗

arXiv2023

Single-shot spatial instability and electric control of polariton condensates at room temperature

In planar microcavities, the transverse-electric and transverse-magnetic (TE-TM) mode splitting of cavity photons arises due to their different penetration into the Bragg mirrors and can result in optical spin-orbit coupling (SOC). In this work, we find that in a liquid crystal (LC) microcavity filled with perovskite microplates, the pronounced TE-TM splitting gives rise to a strong SOC that leads to the spatial instability of microcavity polariton condensates under single-shot excitation. Spatially varying hole burning and mode competition occurs between polarization components leading to different condensate profiles from shot to shot. The single-shot polariton condensates become stable when the SOC vanishes as the TE and TM modes are spectrally well separated from each other, which can be achieved by application of an electric field to our LC microcavity with electrically tunable anisotropy. Our findings are well reproduced and traced back to their physical origin by our detailed numerical simulations. With the electrical manipulation our work reveals how the shot-to-shot spatial instability of spatial polariton profiles can be engineered in anisotropic microcavities at room temperature, which will benefit the development of stable polariton-based optoeletronic and light-emitting devices.

physics.optics↗

arXiv2023

Density-Insensitive Unsupervised Domain Adaption on 3D Object Detection

3D object detection from point clouds is crucial in safety-critical autonomous driving. Although many works have made great efforts and achieved significant progress on this task, most of them suffer from expensive annotation cost and poor transferability to unknown data due to the domain gap. Recently, few works attempt to tackle the domain gap in objects, but still fail to adapt to the gap of varying beam-densities between two domains, which is critical to mitigate the characteristic differences of the LiDAR collectors. To this end, we make the attempt to propose a density-insensitive domain adaption framework to address the density-induced domain gap. In particular, we first introduce Random Beam Re-Sampling (RBRS) to enhance the robustness of 3D detectors trained on the source domain to the varying beam-density. Then, we take this pre-trained detector as the backbone model, and feed the unlabeled target domain data into our newly designed task-specific teacher-student framework for predicting its high-quality pseudo labels. To further adapt the property of density-insensitivity into the target domain, we feed the teacher and student branches with the same sample of different densities, and propose an Object Graph Alignment (OGA) module to construct two object-graphs between the two branches for enforcing the consistency in both the attribute and relation of cross-density objects. Experimental results on three widely adopted 3D object detection datasets demonstrate that our proposed domain adaption method outperforms the state-of-the-art methods, especially over varying-density data. Code is available at https://github.com/WoodwindHu/DTS}{https://github.com/WoodwindHu/DTS.

cs.CV↗

arXiv2023

Lifelong Embedding Learning and Transfer for Growing Knowledge Graphs

Existing knowledge graph (KG) embedding models have primarily focused on static KGs. However, real-world KGs do not remain static, but rather evolve and grow in tandem with the development of KG applications. Consequently, new facts and previously unseen entities and relations continually emerge, necessitating an embedding model that can quickly learn and transfer new knowledge through growth. Motivated by this, we delve into an expanding field of KG embedding in this paper, i.e., lifelong KG embedding. We consider knowledge transfer and retention of the learning on growing snapshots of a KG without having to learn embeddings from scratch. The proposed model includes a masked KG autoencoder for embedding learning and update, with an embedding transfer strategy to inject the learned knowledge into the new entity and relation embeddings, and an embedding regularization method to avoid catastrophic forgetting. To investigate the impacts of different aspects of KG growth, we construct four datasets to evaluate the performance of lifelong KG embedding. Experimental results show that the proposed model outperforms the state-of-the-art inductive and lifelong embedding baselines.

cs.CL↗

arXiv2023

Heterogeneous Federated Knowledge Graph Embedding Learning and Unlearning

Federated Learning (FL) recently emerges as a paradigm to train a global machine learning model across distributed clients without sharing raw data. Knowledge Graph (KG) embedding represents KGs in a continuous vector space, serving as the backbone of many knowledge-driven applications. As a promising combination, federated KG embedding can fully take advantage of knowledge learned from different clients while preserving the privacy of local data. However, realistic problems such as data heterogeneity and knowledge forgetting still remain to be concerned. In this paper, we propose FedLU, a novel FL framework for heterogeneous KG embedding learning and unlearning. To cope with the drift between local optimization and global convergence caused by data heterogeneity, we propose mutual knowledge distillation to transfer local knowledge to global, and absorb global knowledge back. Moreover, we present an unlearning method based on cognitive neuroscience, which combines retroactive interference and passive decay to erase specific knowledge from local clients and propagate to the global model by reusing knowledge distillation. We construct new datasets for assessing realistic performance of the state-of-the-arts. Extensive experiments show that FedLU achieves superior results in both link prediction and knowledge forgetting.

cs.LG↗

arXiv2023

LS-DYNA Machine Learning-based Multiscale Method for Nonlinear Modeling of Short Fiber-Reinforced Composites

Short-fiber-reinforced composites (SFRC) are high-performance engineering materials for lightweight structural applications in the automotive and electronics industries. Typically, SFRC structures are manufactured by injection molding, which induces heterogeneous microstructures, and the resulting nonlinear anisotropic behaviors are challenging to predict by conventional micromechanical analyses. In this work, we present a machine learning-based multiscale method by integrating injection molding-induced microstructures, material homogenization, and Deep Material Network (DMN) in the finite element simulation software LS-DYNA for structural analysis of SFRC. DMN is a physics-embedded machine learning model that learns the microscale material morphologies hidden in representative volume elements of composites through offline training. By coupling DMN with finite elements, we have developed a highly accurate and efficient data-driven approach, which predicts nonlinear behaviors of composite materials and structures at a computational speed orders-of-magnitude faster than the high-fidelity direct numerical simulation. To model industrial-scale SFRC products, transfer learning is utilized to generate a unified DMN database, which effectively captures the effects of injection molding-induced fiber orientations and volume fractions on the overall composite properties. Numerical examples are presented to demonstrate the promising performance of this LS-DYNA machine learning-based multiscale method for SFRC modeling.

cs.CE↗

arXiv2022

Exploring the Devil in Graph Spectral Domain for 3D Point Cloud Attacks

With the maturity of depth sensors, point clouds have received increasing attention in various applications such as autonomous driving, robotics, surveillance, etc., while deep point cloud learning models have shown to be vulnerable to adversarial attacks. Existing attack methods generally add/delete points or perform point-wise perturbation over point clouds to generate adversarial examples in the data space, which may neglect the geometric characteristics of point clouds. Instead, we propose point cloud attacks from a new perspective -- Graph Spectral Domain Attack (GSDA), aiming to perturb transform coefficients in the graph spectral domain that corresponds to varying certain geometric structure. In particular, we naturally represent a point cloud over a graph, and adaptively transform the coordinates of points into the graph spectral domain via graph Fourier transform (GFT) for compact representation. We then analyze the influence of different spectral bands on the geometric structure of the point cloud, based on which we propose to perturb the GFT coefficients in a learnable manner guided by an energy constraint loss function. Finally, the adversarial point cloud is generated by transforming the perturbed spectral representation back to the data domain via the inverse GFT (IGFT). Experimental results demonstrate the effectiveness of the proposed GSDA in terms of both imperceptibility and attack success rates under a variety of defense strategies. The code is available at https://github.com/WoodwindHu/GSDA.

cs.CV↗

arXiv2022

Exploration of growth conditions of TaAs Weyl semimetal thin film by pulsed laser deposition

TaAs, the first experimentally discovered Weyl semimetal material, has attracted a lot of attention due to its high carrier mobility, high anisotropy, nonmagnetic and strong interaction with light. These make it an ideal candidate for the study of Weyl fermions and the applications in quantum computation, thermoelectric devices, and photodetection. For further basic physics studies and potential applications, large-size and high-quality TaAs films are urgently needed. However, it is difficult to grow As-stoichiometry TaAs films due to the volatilization of As during the growth. To solve this problem, the TaAs films were attempted to grow on different substrates using targets with different As stoichiometric ratios by pulsed laser deposition (PLD). In this work, we have found that partial As ions of the GaAs substrate are likely to diffuse into the TaAs films during growth, which was preliminarily confirmed by the structural characterization, surface topography and composition analysis. As a result, the As content in the TaAs film is improved and the TaAs phase is achieved. Our work presents an effective method to fabricate the TaAs films by PLD, providing the possible use of the Weyl semimetal film for functional devices.

cond-mat.mtrl-sci↗

arXiv2022

Learning Latent Part-Whole Hierarchies for Point Clouds

Strong evidence suggests that humans perceive the 3D world by parsing visual scenes and objects into part-whole hierarchies. Although deep neural networks have the capability of learning powerful multi-level representations, they can not explicitly model part-whole hierarchies, which limits their expressiveness and interpretability in processing 3D vision data such as point clouds. To this end, we propose an encoder-decoder style latent variable model that explicitly learns the part-whole hierarchies for the multi-level point cloud segmentation. Specifically, the encoder takes a point cloud as input and predicts the per-point latent subpart distribution at the middle level. The decoder takes the latent variable and the feature from the encoder as an input and predicts the per-point part distribution at the top level. During training, only annotated part labels at the top level are provided, thus making the whole framework weakly supervised. We explore two kinds of approximated inference algorithms, i.e., most-probable-latent and Monte Carlo methods, and three stochastic gradient estimations for learning discrete latent variables, i.e., straight-through, REINFORCE, and pathwise estimators. Experimental results on the PartNet dataset show that the proposed method achieves state-of-the-art performance in not only top-level part segmentation but also middle-level latent subpart segmentation.

cs.CV↗

arXiv2022

EventEA: Benchmarking Entity Alignment for Event-centric Knowledge Graphs

Entity alignment is to find identical entities in different knowledge graphs (KGs) that refer to the same real-world object. Embedding-based entity alignment techniques have been drawing a lot of attention recently because they can help solve the issue of symbolic heterogeneity in different KGs. However, in this paper, we show that the progress made in the past was due to biased and unchallenging evaluation. We highlight two major flaws in existing datasets that favor embedding-based entity alignment techniques, i.e., the isomorphic graph structures in relation triples and the weak heterogeneity in attribute triples. Towards a critical evaluation of embedding-based entity alignment methods, we construct a new dataset with heterogeneous relations and attributes based on event-centric KGs. We conduct extensive experiments to evaluate existing popular methods, and find that they fail to achieve promising performance. As a new approach to this difficult problem, we propose a time-aware literal encoder for entity alignment. The dataset and source code are publicly available to foster future research. Our work calls for more effective and practical embedding-based solutions to entity alignment.

cs.CL↗

arXiv2022

Sparse Graph Learning with Spectrum Prior for Deep Graph Convolutional Networks

A graph convolutional network (GCN) employs a graph filtering kernel tailored for data with irregular structures. However, simply stacking more GCN layers does not improve performance; instead, the output converges to an uninformative low-dimensional subspace, where the convergence rate is characterized by the graph spectrum -- this is the known over-smoothing problem in GCN. In this paper, we propose a sparse graph learning algorithm incorporating a new spectrum prior to compute a graph topology that circumvents over-smoothing while preserving pairwise correlations inherent in data. Specifically, based on a spectral analysis of multilayer GCN output, we derive a spectrum prior for the graph Laplacian matrix $\mathbf{L}$ to robustify the model expressiveness against over-smoothing. Then, we formulate a sparse graph learning problem with the spectrum prior, solved efficiently via block coordinate descent (BCD). Moreover, we optimize the weight parameter trading off the fidelity term with the spectrum prior, based on data smoothness on the original graph learned without spectrum manipulation. The output $\mathbf{L}$ is then normalized for supervised GCN training. Experiments show that our proposal produced deeper GCNs and higher prediction accuracy for regression and classification tasks compared to competing schemes.

cs.LG↗

arXiv2022

Museformer: Transformer with Fine- and Coarse-Grained Attention for Music Generation

Symbolic music generation aims to generate music scores automatically. A recent trend is to use Transformer or its variants in music generation, which is, however, suboptimal, because the full attention cannot efficiently model the typically long music sequences (e.g., over 10,000 tokens), and the existing models have shortcomings in generating musical repetition structures. In this paper, we propose Museformer, a Transformer with a novel fine- and coarse-grained attention for music generation. Specifically, with the fine-grained attention, a token of a specific bar directly attends to all the tokens of the bars that are most relevant to music structures (e.g., the previous 1st, 2nd, 4th and 8th bars, selected via similarity statistics); with the coarse-grained attention, a token only attends to the summarization of the other bars rather than each token of them so as to reduce the computational cost. The advantages are two-fold. First, it can capture both music structure-related correlations via the fine-grained attention, and other contextual information via the coarse-grained attention. Second, it is efficient and can model over 3X longer music sequences compared to its full-attention counterpart. Both objective and subjective experimental results demonstrate its ability to generate long music sequences with high quality and better structures.

cs.SD↗

arXiv2022

RVE Analysis in LS-DYNA for High-fidelity Multiscale Material Modeling

In modern engineering designs, advanced materials (e.g., fiber/particle-reinforced polymers, metallic alloys, laminar composites, etc.) are widely used, where microscale heterogeneities such as grains, inclusions, voids, micro-cracks, and interfaces significantly affect the macroscopic constitutive behaviors. Obviously, an accurate description of the multiscale material behaviors is of great importance to the success of material design and structural analysis. The Representative Volume Element (RVE) analysis method provides a rigorous means to obtain homogenized macroscopic material properties at the upper length scale from the properties of the material constituents and structures at a lower length scale. Recently, we have developed an RVE module (keyword: *RVE_ANALYSIS_FEM) in the multiphysics simulation software LS-DYNA to enable high-fidelity virtual testing of numerically re-constructed material samples at user-specified characteristic length scales. In this article, a brief introduction to this new feature will be given.

cs.CE↗

arXiv2022

Implicit Bias in Leaky ReLU Networks Trained on High-Dimensional Data

The implicit biases of gradient-based optimization algorithms are conjectured to be a major factor in the success of modern deep learning. In this work, we investigate the implicit bias of gradient flow and gradient descent in two-layer fully-connected neural networks with leaky ReLU activations when the training data are nearly-orthogonal, a common property of high-dimensional data. For gradient flow, we leverage recent work on the implicit bias for homogeneous neural networks to show that asymptotically, gradient flow produces a neural network with rank at most two. Moreover, this network is an $\ell_2$-max-margin solution (in parameter space), and has a linear decision boundary that corresponds to an approximate-max-margin linear predictor. For gradient descent, provided the random initialization variance is small enough, we show that a single step of gradient descent suffices to drastically reduce the rank of the network, and that the rank remains small throughout training. We provide experiments which suggest that a small initialization scale is important for finding low-rank neural networks with gradient descent.

cs.LG↗

arXiv2022

Automated Extraction of Energy Systems Information from Remotely Sensed Data: A Review and Analysis

High quality energy systems information is a crucial input to energy systems research, modeling, and decision-making. Unfortunately, actionable information about energy systems is often of limited availability, incomplete, or only accessible for a substantial fee or through a non-disclosure agreement. Recently, remotely sensed data (e.g., satellite imagery, aerial photography) have emerged as a potentially rich source of energy systems information. However, the use of these data is frequently challenged by its sheer volume and complexity, precluding manual analysis. Recent breakthroughs in machine learning have enabled automated and rapid extraction of useful information from remotely sensed data, facilitating large-scale acquisition of critical energy system variables. Here we present a systematic review of the literature on this emerging topic, providing an in-depth survey and review of papers published within the past two decades. We first taxonomize the existing literature into ten major areas, spanning the energy value chain. Within each research area, we distill and critically discuss major features that are relevant to energy researchers, including, for example, key challenges regarding the accessibility and reliability of the methods. We then synthesize our findings to identify limitations and trends in the literature as a whole, and discuss opportunities for innovation. These include the opportunity to extend the methods beyond electricity to broader energy systems and wider geographic areas; and the ability to expand the use of these methods in research and decision making as satellite data become cheaper and easier to access. We also find that there are persistent challenges: limited standardization and rigor of performance assessments; limited sharing of code, which would improve replicability; and a limited consideration of the ethics and privacy of data.

eess.SP↗