arXiv subjects

Xingjian Li

Publications and source records attributed to Xingjian Li.

87 records · Page 5Linked to original sources

arXiv2021

A Neural Network Approach Applied to Multi-Agent Optimal Control

We propose a neural network approach for solving high-dimensional optimal control problems. In particular, we focus on multi-agent control problems with obstacle and collision avoidance. These problems immediately become high-dimensional, even for moderate phase-space dimensions per agent. Our approach fuses the Pontryagin Maximum Principle and Hamilton-Jacobi-Bellman (HJB) approaches and parameterizes the value function with a neural network. Our approach yields controls in a feedback form for quick calculation and robustness to moderate disturbances to the system. We train our model using the objective function and optimality conditions of the control problem. Therefore, our training algorithm neither involves a data generation phase nor solutions from another algorithm. Our model uses empirically effective HJB penalizers for efficient training. By training on a distribution of initial states, we ensure the controls' optimality is achieved on a large portion of the state-space. Our approach is grid-free and scales efficiently to dimensions where grids become impractical or infeasible. We demonstrate our approach's effectiveness on a 150-dimensional multi-agent problem with obstacles.

math.OC↗

arXiv2021

OT-Flow: Fast and Accurate Continuous Normalizing Flows via Optimal Transport

A normalizing flow is an invertible mapping between an arbitrary probability distribution and a standard normal distribution; it can be used for density estimation and statistical inference. Computing the flow follows the change of variables formula and thus requires invertibility of the mapping and an efficient way to compute the determinant of its Jacobian. To satisfy these requirements, normalizing flows typically consist of carefully chosen components. Continuous normalizing flows (CNFs) are mappings obtained by solving a neural ordinary differential equation (ODE). The neural ODE's dynamics can be chosen almost arbitrarily while ensuring invertibility. Moreover, the log-determinant of the flow's Jacobian can be obtained by integrating the trace of the dynamics' Jacobian along the flow. Our proposed OT-Flow approach tackles two critical computational challenges that limit a more widespread use of CNFs. First, OT-Flow leverages optimal transport (OT) theory to regularize the CNF and enforce straight trajectories that are easier to integrate. Second, OT-Flow features exact trace computation with time complexity equal to trace estimators used in existing CNFs. On five high-dimensional density estimation and generative modeling tasks, OT-Flow performs competitively to state-of-the-art CNFs while on average requiring one-fourth of the number of weights with an 8x speedup in training time and 24x speedup in inference.

cs.LG↗

arXiv2020

Temporal Relational Modeling with Self-Supervision for Action Segmentation

Temporal relational modeling in video is essential for human action understanding, such as action recognition and action segmentation. Although Graph Convolution Networks (GCNs) have shown promising advantages in relation reasoning on many tasks, it is still a challenge to apply graph convolution networks on long video sequences effectively. The main reason is that large number of nodes (i.e., video frames) makes GCNs hard to capture and model temporal relations in videos. To tackle this problem, in this paper, we introduce an effective GCN module, Dilated Temporal Graph Reasoning Module (DTGRM), designed to model temporal relations and dependencies between video frames at various time spans. In particular, we capture and model temporal relations via constructing multi-level dilated temporal graphs where the nodes represent frames from different moments in video. Moreover, to enhance temporal reasoning ability of the proposed model, an auxiliary self-supervised task is proposed to encourage the dilated temporal graph reasoning module to find and correct wrong temporal relations in videos. Our DTGRM model outperforms state-of-the-art action segmentation models on three challenging datasets: 50Salads, Georgia Tech Egocentric Activities (GTEA), and the Breakfast dataset. The code is available at https://github.com/redwang/DTGRM.

cs.CV↗

arXiv2020

Measuring Information Transfer in Neural Networks

Quantifying the information content in a neural network model is essentially estimating the model's Kolmogorov complexity. Recent success of prequential coding on neural networks points to a promising path of deriving an efficient description length of a model. We propose a practical measure of the generalizable information in a neural network model based on prequential coding, which we term Information Transfer ($L_{IT}$). Theoretically, $L_{IT}$ is an estimation of the generalizable part of a model's information content. In experiments, we show that $L_{IT}$ is consistently correlated with generalizable information and can be used as a measure of patterns or "knowledge" in a model or a dataset. Consequently, $L_{IT}$ can serve as a useful analysis tool in deep learning. In this paper, we apply $L_{IT}$ to compare and dissect information in datasets, evaluate representation models in transfer learning, and analyze catastrophic forgetting and continual learning algorithms. $L_{IT}$ provides an information perspective which helps us discover new insights into neural network learning.

cs.LG↗

arXiv2020

XMixup: Efficient Transfer Learning with Auxiliary Samples by Cross-domain Mixup

Transferring knowledge from large source datasets is an effective way to fine-tune the deep neural networks of the target task with a small sample size. A great number of algorithms have been proposed to facilitate deep transfer learning, and these techniques could be generally categorized into two groups - Regularized Learning of the target task using models that have been pre-trained from source datasets, and Multitask Learning with both source and target datasets to train a shared backbone neural network. In this work, we aim to improve the multitask paradigm for deep transfer learning via Cross-domain Mixup (XMixup). While the existing multitask learning algorithms need to run backpropagation over both the source and target datasets and usually consume a higher gradient complexity, XMixup transfers the knowledge from source to target tasks more efficiently: for every class of the target task, XMixup selects the auxiliary samples from the source dataset and augments training samples via the simple mixup strategy. We evaluate XMixup over six real world transfer learning datasets. Experiment results show that XMixup improves the accuracy by 1.9% on average. Compared with other state-of-the-art transfer learning approaches, XMixup costs much less training time while still obtains higher accuracy.

cs.LG↗

arXiv2020

RIFLE: Backpropagation in Depth for Deep Transfer Learning through Re-Initializing the Fully-connected LayEr

Fine-tuning the deep convolution neural network(CNN) using a pre-trained model helps transfer knowledge learned from larger datasets to the target task. While the accuracy could be largely improved even when the training dataset is small, the transfer learning outcome is usually constrained by the pre-trained model with close CNN weights (Liu et al., 2019), as the backpropagation here brings smaller updates to deeper CNN layers. In this work, we propose RIFLE - a simple yet effective strategy that deepens backpropagation in transfer learning settings, through periodically Re-Initializing the Fully-connected LayEr with random scratch during the fine-tuning procedure. RIFLE brings meaningful updates to the weights of deep CNN layers and improves low-level feature learning, while the effects of randomization can be easily converged throughout the overall learning procedure. The experiments show that the use of RIFLE significantly improves deep transfer learning accuracy on a wide range of datasets, out-performing known tricks for the similar purpose, such as Dropout, DropConnect, StochasticDepth, Disturb Label and Cyclic Learning Rate, under the same settings with 0.5% -2% higher testing accuracy. Empirical cases and ablation studies further indicate RIFLE brings meaningful updates to deep CNN layers with accuracy improved.

cs.LG↗

arXiv2020

DELTA: DEep Learning Transfer using Feature Map with Attention for Convolutional Networks

Transfer learning through fine-tuning a pre-trained neural network with an extremely large dataset, such as ImageNet, can significantly accelerate training while the accuracy is frequently bottlenecked by the limited dataset size of the new target task. To solve the problem, some regularization methods, constraining the outer layer weights of the target network using the starting point as references (SPAR), have been studied. In this paper, we propose a novel regularized transfer learning framework DELTA, namely DEep Learning Transfer using Feature Map with Attention. Instead of constraining the weights of neural network, DELTA aims to preserve the outer layer outputs of the target network. Specifically, in addition to minimizing the empirical loss, DELTA intends to align the outer layer outputs of two networks, through constraining a subset of feature maps that are precisely selected by attention that has been learned in an supervised learning manner. We evaluate DELTA with the state-of-the-art algorithms, including L2 and L2-SP. The experiment results show that our proposed method outperforms these baselines with higher accuracy for new tasks.

cs.LG↗

arXiv2020

COLAM: Co-Learning of Deep Neural Networks and Soft Labels via Alternating Minimization

Softening labels of training datasets with respect to data representations has been frequently used to improve the training of deep neural networks (DNNs). While such a practice has been studied as a way to leverage privileged information about the distribution of the data, a well-trained learner with soft classification outputs should be first obtained as a prior to generate such privileged information. To solve such chicken-egg problem, we propose COLAM framework that Co-Learns DNNs and soft labels through Alternating Minimization of two objectives - (a) the training loss subject to soft labels and (b) the objective to learn improved soft labels - in one end-to-end training procedure. We performed extensive experiments to compare our proposed method with a series of baselines. The experiment results show that COLAM achieves improved performance on many tasks with better testing classification accuracy. We also provide both qualitative and quantitative analyses that explain why COLAM works well.

cs.LG↗

arXiv2019

Joint Active and Passive Beamforming for Intelligent Reflecting Surface-Assisted Massive MIMO Systems

In this paper, we study the problem of joint active and passive beamforming for intelligent reflecting surface (IRS)-assisted massive MIMO systems, where multiple IRSs equipped with a large number of passive elements are deployed to assist a base station (BS) to simultaneously serve a small number of single-antenna users in the same time-frequency resource. Our objective is to maximize the minimum signal to interference plus noise (SINR) at users by jointly optimizing the transmit precoding vector at the BS and phase shift parameters at IRSs. We show that an interesting automatic interference cancelation (AIC) property holds asymptotically as the number of passive elements approaches infinity, i.e., when an IRS is optimally tuned to serve a certain user, this IRS will become interference-free to other users. By utilizing this property, the max-min problem can be converted into an IRS-user association problem, where the objective is to determine which IRSs are assigned for each user. An exhaustive search scheme and a greedy search scheme are proposed to solve the IRS-user association problem. Our theoretical analysis reveals that our proposed solution attains an SINR that scales quadratically with the number of reflecting elements. Also, our theoretical result suggests that even with a moderate number of active antennas at the BS, a massive MIMO like gain can be achieved by increasing the number of passive reflecting elements, thus significantly reducing the energy consumption at the BS. Simulation results are provided to corroborate our theoretical results and to illustrate the effectiveness of our proposed solution.

cs.IT↗

arXiv2019

Towards Making Deep Transfer Learning Never Hurt

Transfer learning have been frequently used to improve deep neural network training through incorporating weights of pre-trained networks as the starting-point of optimization for regularization. While deep transfer learning can usually boost the performance with better accuracy and faster convergence, transferring weights from inappropriate networks hurts training procedure and may lead to even lower accuracy. In this paper, we consider deep transfer learning as minimizing a linear combination of empirical loss and regularizer based on pre-trained weights, where the regularizer would restrict the training procedure from lowering the empirical loss, with conflicted descent directions (e.g., derivatives). Following the view, we propose a novel strategy making regularization-based Deep Transfer learning Never Hurt (DTNH) that, for each iteration of training procedure, computes the derivatives of the two terms separately, then re-estimates a new descent direction that does not hurt the empirical loss minimization while preserving the regularization affects from the pre-trained weights. Extensive experiments have been done using common transfer learning regularizers, such as L2-SP and knowledge distillation, on top of a wide range of deep transfer learning benchmarks including Caltech, MIT indoor 67, CIFAR-10 and ImageNet. The empirical results show that the proposed descent direction estimation strategy DTNH can always improve the performance of deep transfer learning tasks based on all above regularizers, even when transferring pre-trained weights from inappropriate networks. All in all, DTNH strategy can improve state-of-the-art regularizers in all cases with 0.1%--7% higher accuracy in all experiments.

cs.LG↗

arXiv2019

An Empirical Study on Regularization of Deep Neural Networks by Local Rademacher Complexity

Regularization of Deep Neural Networks (DNNs) for the sake of improving their generalization capability is important and challenging. The development in this line benefits theoretical foundation of DNNs and promotes their usability in different areas of artificial intelligence. In this paper, we investigate the role of Rademacher complexity in improving generalization of DNNs and propose a novel regularizer rooted in Local Rademacher Complexity (LRC). While Rademacher complexity is well known as a distribution-free complexity measure of function class that help boost generalization of statistical learning methods, extensive study shows that LRC, its counterpart focusing on a restricted function class, leads to sharper convergence rates and potential better generalization given finite training sample. Our LRC based regularizer is developed by estimating the complexity of the function class centered at the minimizer of the empirical loss of DNNs. Experiments on various types of network architecture demonstrate the effectiveness of LRC regularization in improving generalization. Moreover, our method features the state-of-the-art result on the CIFAR-$10$ dataset with network architecture found by neural architecture search.

cs.LG↗

arXiv2018

Fast Beam Alignment for Millimeter Wave Communications: A Sparse Encoding and Phaseless Decoding Approach

In this paper, we studied the problem of beam alignment for millimeter wave (mmWave) communications, in which we assume a hybrid analog and digital beamforming structure is employed at the transmitter (i.e. base station), and an omni-directional antenna or an antenna array is used at the receiver (i.e. user). By exploiting the sparse scattering nature of mmWave channels, the beam alignment problem is formulated as a sparse encoding and phaseless decoding problem. More specifically, the problem of interest involves finding a sparse sensing matrix and an efficient recovery algorithm to recover the support and magnitude of the sparse signal from compressive phaseless measurements. A sparse bipartite graph coding (SBG-Coding) algorithm is developed for sparse encoding and phaseless decoding. Our theoretical analysis shows that, in the noiseless case, our proposed algorithm can perfectly recover the support and magnitude of the sparse signal with probability exceeding a pre-specified value from $\mathcal{O}(K^2)$ measurements, where $K$ is the number of nonzero entries of the sparse signal. The proposed algorithm has a simple decoding procedure which is computationally efficient and noise-robust. Simulation results show that our proposed method renders a reliable beam alignment in the low and moderate signal-to-noise ratio (SNR) regimes and presents a clear performance advantage over existing methods.

cs.IT↗

arXiv2018

Intelligent Power Control for Spectrum Sharing in Cognitive Radios: A Deep Reinforcement Learning Approach

We consider the problem of spectrum sharing in a cognitive radio system consisting of a primary user and a secondary user. The primary user and the secondary user work in a non-cooperative manner. Specifically, the primary user is assumed to update its transmit power based on a pre-defined power control policy. The secondary user does not have any knowledge about the primary user's transmit power, or its power control strategy. The objective of this paper is to develop a learning-based power control method for the secondary user in order to share the common spectrum with the primary user. To assist the secondary user, a set of sensor nodes are spatially deployed to collect the received signal strength information at different locations in the wireless environment. We develop a deep reinforcement learning-based method, which the secondary user can use to intelligently adjust its transmit power such that after a few rounds of interaction with the primary user, both users can transmit their own data successfully with required qualities of service. Our experimental results show that the secondary user can interact with the primary user efficiently to reach a goal state (defined as a state in which both users can successfully transmit their data) from any initial states within a few number of steps.

cs.IT↗

arXiv2017

Millimeter Wave Channel Estimation via Exploiting Joint Sparse and Low-Rank Structures

We consider the problem of channel estimation for millimeter wave (mmWave) systems, where, to minimize the hardware complexity and power consumption, an analog transmit beamforming and receive combining structure with only one radio frequency (RF) chain at the base station (BS) and mobile station (MS) is employed. Most existing works for mmWave channel estimation exploit sparse scattering characteristics of the channel. In addition to sparsity, mmWave channels may exhibit angular spreads over the angle of arrival (AoA), angle of departure (AoD), and elevation domains. In this paper, we show that angular spreads give rise to a useful low-rank structure that, along with the sparsity, can be simultaneously utilized to reduce the sample complexity, i.e. the number of samples needed to successfully recover the mmWave channel. Specifically, to effectively leverage the joint sparse and low-rank structure, we develop a two-stage compressed sensing method for mmWave channel estimation, where the sparse and low-rank properties are respectively utilized in two consecutive stages, namely, a matrix completion stage and a sparse recovery stage. Our theoretical analysis reveals that the proposed two-stage scheme can achieve a lower sample complexity than a direct compressed sensing method that exploits only the sparse structure of the mmWave channel. Simulation results are provided to corroborate our theoretical results and to show the superiority of the proposed two-stage method.

cs.IT↗

arXiv2016

Low-Rank Covariance-Assisted Downlink Training and Channel Estimation for FDD Massive MIMO Systems

We consider the problem of downlink training and channel estimation in frequency division duplex (FDD) massive MIMO systems, where the base station (BS) equipped with a large number of antennas serves a number of single-antenna users simultaneously. To obtain the channel state information (CSI) at the BS in FDD systems, the downlink channel has to be estimated by users via downlink training and then fed back to the BS. For FDD large-scale MIMO systems, the overhead for downlink training and CSI uplink feedback could be prohibitively high, which presents a significant challenge. In this paper, we study the behavior of the minimum mean-squared error (MMSE) estimator when the channel covariance matrix has a low-rank or an approximate low-rank structure. Our theoretical analysis reveals that the amount of training overhead can be substantially reduced by exploiting the low-rank property of the channel covariance matrix. In particular, we show that the MMSE estimator is able to achieve exact channel recovery in the asymptotic low-noise regime, provided that the number of pilot symbols in time is no less than the rank of the channel covariance matrix. We also present an optimal pilot design for the single-user case, and an asymptotic optimal pilot design for the multi-user scenario. Lastly, we develop a simple model-based scheme to estimate the channel covariance matrix, based on which the MMSE estimator can be employed to estimate the channel. The proposed scheme does not need any additional training overhead. Simulation results are provided to verify our theoretical results and illustrate the effectiveness of the proposed estimated covariance-assisted MMSE estimator.

cs.IT↗