Xin Liu

Xin Liu, Zhongdao Wang, Yali Li et al.

A mainstream type of current self-supervised learning methods pursues a general-purpose representation that can be well transferred to downstream tasks, typically by optimizing on a given pretext task such as instance discrimination. In this work, we argue that existing pretext tasks inevitably introduce biases into the learned representation, which in turn leads to biased transfer performance on various downstream tasks. To cope with this issue, we propose Maximum Entropy Coding (MEC), a more principled objective that explicitly optimizes on the structure of the representation, so that the learned representation is less biased and thus generalizes better to unseen downstream tasks. Inspired by the principle of maximum entropy in information theory, we hypothesize that a generalizable representation should be the one that admits the maximum entropy among all plausible representations. To make the objective end-to-end trainable, we propose to leverage the minimal coding length in lossy data coding as a computationally tractable surrogate for the entropy, and further derive a scalable reformulation of the objective that allows fast computation. Extensive experiments demonstrate that MEC learns a more generalizable representation than previous methods based on specific pretext tasks. It achieves state-of-the-art performance consistently on various downstream tasks, including not only ImageNet linear probe, but also semi-supervised classification, object detection, instance segmentation, and object tracking. Interestingly, we show that existing batch-wise and feature-wise self-supervised objectives could be seen equivalent to low-order approximations of MEC. Code and pre-trained models are available at https://github.com/xinliu20/MEC.

Huanle Zhang, Lei Fu, Mi Zhang et al.

Federated learning (FL) is a distributed model training paradigm that preserves clients' data privacy. It has gained tremendous attention from both academia and industry. FL hyper-parameters (e.g., the number of selected clients and the number of training passes) significantly affect the training overhead in terms of computation time, transmission time, computation load, and transmission load. However, the current practice of manually selecting FL hyper-parameters imposes a heavy burden on FL practitioners because applications have different training preferences. In this paper, we propose FedTune, an automatic FL hyper-parameter tuning algorithm tailored to applications' diverse system requirements in FL training. FedTune iteratively adjusts FL hyper-parameters during FL training and can be easily integrated into existing FL systems. Through extensive evaluations of FedTune for diverse applications and FL aggregation algorithms, we show that FedTune is lightweight and effective, achieving 8.48%-26.75% system overhead reduction compared to using fixed FL hyper-parameters. This paper assists FL practitioners in designing high-performance FL training solutions. The source code of FedTune is available at https://github.com/DataSysTech/FedTune.

Chao Huang, Jianwei Huang, Xin Liu

Federated learning (FL) is an emerging technology that enables the training of machine learning models from multiple clients while keeping the data distributed and private. Based on the participating clients and the model training scale, federated learning can be classified into two types: cross-device FL where clients are typically mobile devices and the client number can reach up to a scale of millions; cross-silo FL where clients are organizations or companies and the client number is usually small (e.g., within a hundred). While existing studies mainly focus on cross-device FL, this paper aims to provide an overview of the cross-silo FL. More specifically, we first discuss applications of cross-silo FL and outline its major challenges. We then provide a systematic overview of the existing approaches to the challenges in cross-silo FL by focusing on their connections and differences to cross-device FL. Finally, we discuss future directions and open issues that merit research efforts from the community.

Xin Liu, Mingchuan Zhang, Ziheng Jiang et al.

The growing need for technology that supports remote healthcare is being acutely highlighted by an aging population and the COVID-19 pandemic. In health-related machine learning applications the ability to learn predictive models without data leaving a private device is attractive, especially when these data might contain features (e.g., photographs or videos of the body) that make identifying a subject trivial and/or the training data volume is large (e.g., uncompressed video). Camera-based remote physiological sensing facilitates scalable and low-cost measurement, but is a prime example of a task that involves analysing high bit-rate videos containing identifiable images and sensitive health information. Federated learning enables privacy-preserving decentralized training which has several properties beneficial for camera-based sensing. We develop the first mobile federated learning camera-based sensing system and show that it can perform competitively with traditional state-of-the-art supervised approaches. However, in the presence of corrupted data (e.g., video or label noise) from a few devices the performance of weight averaging quickly degrades. To address this, we leverage knowledge about the expected noise profile within the video to intelligently adjust how the model weights are averaged on the server. Our results show that this significantly improves upon the robustness of models even when the signal-to-noise ratio is low

Lei Wang, Xin Liu

Decentralized optimization with orthogonality constraints is found widely in scientific computing and data science. Since the orthogonality constraints are nonconvex, it is quite challenging to design efficient algorithms. Existing approaches leverage the geometric tools from Riemannian optimization to solve this problem at the cost of high sample and communication complexities. To relieve this difficulty, based on two novel techniques that can waive the orthogonality constraints, we propose a variance-reduced stochastic gradient tracking (VRSGT) algorithm with the convergence rate of $O(1 / k)$ to a stationary point. To the best of our knowledge, VRSGT is the first algorithm for decentralized optimization with orthogonality constraints that reduces both sampling and communication complexities simultaneously. In the numerical experiments, VRSGT has a promising performance in a real-world autonomous driving application.

Shuqi Ke, Chao Huang, Xin Liu

Federated Learning (FL) is a distributed machine learning paradigm where clients collaboratively train a model using their local (human-generated) datasets. While existing studies focus on FL algorithm development to tackle data heterogeneity across clients, the important issue of data quality (e.g., label noise) in FL is overlooked. This paper aims to fill this gap by providing a quantitative study on the impact of label noise on FL. We derive an upper bound for the generalization error that is linear in the clients' label noise level. Then we conduct experiments on MNIST and CIFAR-10 datasets using various FL algorithms. Our empirical results show that the global model accuracy linearly decreases as the noise level increases, which is consistent with our theoretical analysis. We further find that label noise slows down the convergence of FL training, and the global model tends to overfit when the noise level is high.

Kun He, Xin Liu, Yichen Yang et al.

Adversarial training has been demonstrated to be the most effective approach to defend against adversarial attacks. However, existing adversarial training methods show apparent oscillations and overfitting issue in the training process, degrading the defense efficacy. In this work, we propose a novel framework, termed Parameter Interpolation based Adversarial Training (PIAT), that makes full use of the historical information during training. Specifically, at the end of each epoch, PIAT tunes the model parameters as the interpolation of the parameters of the previous and current epochs. Besides, we suggest to use the Normalized Mean Square Error (NMSE) to further improve the robustness by aligning the clean and adversarial examples. Compared with other regularization methods, NMSE focuses more on the relative magnitude of the logits rather than the absolute magnitude. Extensive experiments on several benchmark datasets and various networks show that our method could prominently improve the model robustness and reduce the generalization error. Moreover, our framework is general and could further boost the robust accuracy when combined with other adversarial training methods.

Haoyi Xiu, Xin Liu, Weimin Wang et al.

We present a simple but effective attention named the unary-pairwise attention (UPA) for modeling the relationship between 3D point clouds. Our idea is motivated by the analysis that the standard self-attention (SA) that operates globally tends to produce almost the same attention maps for different query positions, revealing difficulties for learning query-independent and query-dependent information jointly. Therefore, we reformulate the SA and propose query-independent (Unary) and query-dependent (Pairwise) components to facilitate the learning of both terms. In contrast to the SA, the UPA ensures query dependence via operating locally. Extensive experiments show that the UPA outperforms the SA consistently on various point cloud understanding tasks including shape classification, part segmentation, and scene segmentation. Moreover, simply equipping the popular PointNet++ method with the UPA even outperforms or is on par with the state-of-the-art attention-based approaches. In addition, the UPA systematically boosts the performance of both standard and modern networks when it is integrated into them as a compositional module.

Lei Fu, Huanle Zhang, Ge Gao et al.

As a privacy-preserving paradigm for training Machine Learning (ML) models, Federated Learning (FL) has received tremendous attention from both industry and academia. In a typical FL scenario, clients exhibit significant heterogeneity in terms of data distribution and hardware configurations. Thus, randomly sampling clients in each training round may not fully exploit the local updates from heterogeneous clients, resulting in lower model accuracy, slower convergence rate, degraded fairness, etc. To tackle the FL client heterogeneity problem, various client selection algorithms have been developed, showing promising performance improvement. In this paper, we systematically present recent advances in the emerging field of FL client selection and its challenges and research opportunities. We hope to facilitate practitioners in choosing the most suitable client selection mechanisms for their applications, as well as inspire researchers and newcomers to better understand this exciting research topic.

Xin Liu, Wei Tao, Zhisong Pan

Momentum methods, including heavy-ball~(HB) and Nesterov's accelerated gradient~(NAG), are widely used in training neural networks for their fast convergence. However, there is a lack of theoretical guarantees for their convergence and acceleration since the optimization landscape of the neural network is non-convex. Nowadays, some works make progress towards understanding the convergence of momentum methods in an over-parameterized regime, where the number of the parameters exceeds that of the training instances. Nonetheless, current results mainly focus on the two-layer neural network, which are far from explaining the remarkable success of the momentum methods in training deep neural networks. Motivated by this, we investigate the convergence of NAG with constant learning rate and momentum parameter in training two architectures of deep linear networks: deep fully-connected linear neural networks and deep linear ResNets. Based on the over-parameterization regime, we first analyze the residual dynamics induced by the training trajectory of NAG for a deep fully-connected linear neural network under the random Gaussian initialization. Our results show that NAG can converge to the global minimum at a $(1 - \mathcal{O}(1/\sqrtκ))^t$ rate, where $t$ is the iteration number and $κ> 1$ is a constant depending on the condition number of the feature matrix. Compared to the $(1 - \mathcal{O}(1/κ))^t$ rate of GD, NAG achieves an acceleration over GD. To the best of our knowledge, this is the first theoretical guarantee for the convergence of NAG to the global minimum in training deep neural networks. Furthermore, we extend our analysis to deep linear ResNets and derive a similar convergence result.

Xin Ma, Renyi Bao, Jinpeng Jiang et al.

In this work, we propose FedSSO, a server-side second-order optimization method for federated learning (FL). In contrast to previous works in this direction, we employ a server-side approximation for the Quasi-Newton method without requiring any training data from the clients. In this way, we not only shift the computation burden from clients to server, but also eliminate the additional communication for second-order updates between clients and server entirely. We provide theoretical guarantee for convergence of our novel method, and empirically demonstrate our fast convergence and communication savings in both convex and non-convex settings.

Yuting Zhang, Hao Lu, Xin Liu et al.

Remote photoplethysmography (rPPG) is a promising technology that captures physiological signals from face videos, with potential applications in medical health, emotional computing, and biosecurity recognition. The demand for rPPG tasks has expanded from demonstrating good performance on intra-dataset testing to cross-dataset testing (i.e., domain generalization). However, most existing methods have overlooked the prior knowledge of rPPG, resulting in poor generalization ability. In this paper, we propose a novel framework that simultaneously utilizes explicit and implicit prior knowledge in the rPPG task. Specifically, we systematically analyze the causes of noise sources (e.g., different camera, lighting, skin types, and movement) across different domains and incorporate these prior knowledge into the network. Additionally, we leverage a two-branch network to disentangle the physiological feature distribution from noises through implicit label correlation. Our extensive experiments demonstrate that the proposed method not only outperforms state-of-the-art methods on RGB cross-dataset evaluation but also generalizes well from RGB datasets to NIR datasets. The code is available at https://github.com/keke-nice/Greip.

Siyuan Zhang, Nachuan Xiao, Xin Liu

In this paper, we focus on the decentralized stochastic subgradient-based methods in minimizing nonsmooth nonconvex functions without Clarke regularity, especially in the decentralized training of nonsmooth neural networks. We propose a general framework that unifies various decentralized subgradient-based methods, such as decentralized stochastic subgradient descent (DSGD), DSGD with gradient-tracking technique (DSGD-T), and DSGD with momentum (DSGD-M). To establish the convergence properties of our proposed framework, we relate the discrete iterates to the trajectories of a continuous-time differential inclusion, which is assumed to have a coercive Lyapunov function with a stable set $\mathcal{A}$. We prove the asymptotic convergence of the iterates to the stable set $\mathcal{A}$ with sufficiently small and diminishing step-sizes. These results provide first convergence guarantees for some well-recognized of decentralized stochastic subgradient-based methods without Clarke regularity of the objective function. Preliminary numerical experiments demonstrate that our proposed framework yields highly efficient decentralized stochastic subgradient-based methods with convergence guarantees in the training of nonsmooth neural networks.

Honghao Wei, Xiyue Peng, Arnob Ghosh et al.

We propose WSAC (Weighted Safe Actor-Critic), a novel algorithm for Safe Offline Reinforcement Learning (RL) under functional approximation, which can robustly optimize policies to improve upon an arbitrary reference policy with limited data coverage. WSAC is designed as a two-player Stackelberg game to optimize a refined objective function. The actor optimizes the policy against two adversarially trained value critics with small importance-weighted Bellman errors, which focus on scenarios where the actor's performance is inferior to the reference policy. In theory, we demonstrate that when the actor employs a no-regret optimization oracle, WSAC achieves a number of guarantees: (i) For the first time in the safe offline RL setting, we establish that WSAC can produce a policy that outperforms any reference policy while maintaining the same level of safety, which is critical to designing a safe algorithm for offline RL. (ii) WSAC achieves the optimal statistical convergence rate of $1/\sqrt{N}$ to the reference policy, where $N$ is the size of the offline dataset. (iii) We theoretically show that WSAC guarantees a safe policy improvement across a broad range of hyperparameters that control the degree of pessimism, indicating its practical robustness. Additionally, we offer a practical version of WSAC and compare it with existing state-of-the-art safe offline RL algorithms in several continuous control environments. WSAC outperforms all baselines across a range of tasks, supporting the theoretical results.

Huanle Zhang, Mi Zhang, Xin Liu et al.

Federated learning (FL) hyper-parameters significantly affect the training overheads in terms of computation time, transmission time, computation load, and transmission load. However, the current practice of manually selecting FL hyper-parameters puts a high burden on FL practitioners since various applications prefer different training preferences. In this paper, we propose FedTune, an automatic FL hyper-parameter tuning algorithm tailored to applications' diverse system requirements of FL training. FedTune is lightweight and flexible, achieving 8.48%-26.75% improvement for different datasets compared to fixed FL hyper-parameters.

Xin Liu, Zhisong Pan, Wei Tao

Momentum methods, such as heavy ball method~(HB) and Nesterov's accelerated gradient method~(NAG), have been widely used in training neural networks by incorporating the history of gradients into the current updating process. In practice, they often provide improved performance over (stochastic) gradient descent~(GD) with faster convergence. Despite these empirical successes, theoretical understandings of their accelerated convergence rates are still lacking. Recently, some attempts have been made by analyzing the trajectories of gradient-based methods in an over-parameterized regime, where the number of the parameters is significantly larger than the number of the training instances. However, the majority of existing theoretical work is mainly concerned with GD and the established convergence result of NAG is inferior to HB and GD, which fails to explain the practical success of NAG. In this paper, we take a step towards closing this gap by analyzing NAG in training a randomly initialized over-parameterized two-layer fully connected neural network with ReLU activation. Despite the fact that the objective function is non-convex and non-smooth, we show that NAG converges to a global minimum at a non-asymptotic linear rate $(1-Θ(1/\sqrtκ))^t$, where $κ> 1$ is the condition number of a gram matrix and $t$ is the number of the iterations. Compared to the convergence rate $(1-Θ(1/κ))^t$ of GD, our result provides theoretical guarantees for the acceleration of NAG in neural network training. Furthermore, our findings suggest that NAG and HB have similar convergence rate. Finally, we conduct extensive experiments on six benchmark datasets to validate the correctness of our theoretical results.

Honghao Wei, Xin Liu, Lei Ying

This paper presents the first model-free, simulator-free reinforcement learning algorithm for Constrained Markov Decision Processes (CMDPs) with sublinear regret and zero constraint violation. The algorithm is named Triple-Q because it includes three key components: a Q-function (also called action-value function) for the cumulative reward, a Q-function for the cumulative utility for the constraint, and a virtual-Queue that (over)-estimates the cumulative constraint violation. Under Triple-Q, at each step, an action is chosen based on the pseudo-Q-value that is a combination of the three "Q" values. The algorithm updates the reward and utility Q-values with learning rates that depend on the visit counts to the corresponding (state, action) pairs and are periodically reset. In the episodic CMDP setting, Triple-Q achieves $\tilde{\cal O}\left(\frac{1 }δH^4 S^{\frac{1}{2}}A^{\frac{1}{2}}K^{\frac{4}{5}} \right)$ regret, where $K$ is the total number of episodes, $H$ is the number of steps in each episode, $S$ is the number of states, $A$ is the number of actions, and $δ$ is Slater's constant. Furthermore, Triple-Q guarantees zero constraint violation, both on expectation and with a high probability, when $K$ is sufficiently large. Finally, the computational complexity of Triple-Q is similar to SARSA for unconstrained MDPs and is computationally efficient.

Shahbaz Rezaei, Zubair Shafiq, Xin Liu

Deep ensemble learning has been shown to improve accuracy by training multiple neural networks and averaging their outputs. Ensemble learning has also been suggested to defend against membership inference attacks that undermine privacy. In this paper, we empirically demonstrate a trade-off between these two goals, namely accuracy and privacy (in terms of membership inference attacks), in deep ensembles. Using a wide range of datasets and model architectures, we show that the effectiveness of membership inference attacks increases when ensembling improves accuracy. We analyze the impact of various factors in deep ensembles and demonstrate the root cause of the trade-off. Then, we evaluate common defenses against membership inference attacks based on regularization and differential privacy. We show that while these defenses can mitigate the effectiveness of membership inference attacks, they simultaneously degrade ensemble accuracy. We illustrate similar trade-off in more advanced and state-of-the-art ensembling techniques, such as snapshot ensembles and diversified ensemble networks. Finally, we propose a simple yet effective defense for deep ensembles to break the trade-off and, consequently, improve the accuracy and privacy, simultaneously.

Quentin Jodelet, Xin Liu, Tsuyoshi Murata

When incrementally trained on new classes, deep neural networks are subject to catastrophic forgetting which leads to an extreme deterioration of their performance on the old classes while learning the new ones. Using a small memory containing few samples from past classes has shown to be an effective method to mitigate catastrophic forgetting. However, due to the limited size of the replay memory, there is a large imbalance between the number of samples for the new and the old classes in the training dataset resulting in bias in the final model. To address this issue, we propose to use the Balanced Softmax Cross-Entropy and show that it can be seamlessly combined with state-of-the-art approaches for class-incremental learning in order to improve their accuracy while also potentially decreasing the computational cost of the training procedure. We further extend this approach to the more demanding class-incremental learning without memory setting and achieve competitive results with memory-based approaches. Experiments on the challenging ImageNet, ImageNet-Subset and CIFAR100 benchmarks with various settings demonstrate the benefits of our approach.

Yi Wei, Xue Mei, Xin Liu et al.

Training a deep neural network heavily relies on a large amount of training data with accurate annotations. To alleviate this problem, various methods have been proposed to annotate the data automatically. However, automatically generating annotations will inevitably yields noisy labels. In this paper, we propose a Data Selection and joint Training (DST) method to automatically select training samples with accurate annotations. Specifically, DST fits a mixture model according to the original annotation as well as the predicted label for each training sample, and the mixture model is utilized to dynamically divide the training dataset into a correctly labeled dataset, a correctly predicted set and a wrong dataset. Then, DST is trained with these datasets in a supervised manner. Due to confirmation bias problem, we train the two networks alternately, and each network is tasked to establish the data division to teach another network. For each iteration, the correctly labeled and predicted labels are reweighted respectively by the probabilities from the mixture model, and a uniform distribution is used to generate the probabilities of the wrong samples. Experiments on CIFAR-10, CIFAR-100 and Clothing1M demonstrate that DST is the comparable or superior to the state-of-the-art methods.