BiX-NAS: Searching Efficient Bi-directional Architecture for Medical Image Segmentation

The recurrent mechanism has recently been introduced into U-Net in various medical image segmentation tasks. Existing studies have focused on promoting network recursion via reusing building blocks. Although network parameters could be greatly saved, computational costs still increase inevitably in accordance with the pre-set iteration time. In this work, we study a multi-scale upgrade of a bi-directional skip connected network and then automatically discover an efficient architecture by a novel two-phase Neural Architecture Search (NAS) algorithm, namely BiX-NAS. Our proposed method reduces the network computational cost by sifting out ineffective multi-scale features at different levels and iterations. We evaluate BiX-NAS on two segmentation tasks using three different medical image datasets, and the experimental results show that our BiX-NAS searched architecture achieves the state-of-the-art performance with significantly lower computational cost.

BiAdam: Fast Adaptive Bilevel Optimization Methods

Bilevel optimization recently has attracted increased interest in machine learning due to its many applications such as hyper-parameter optimization and policy optimization. Although some methods recently have been proposed to solve the bilevel problems, these methods do not consider using adaptive learning rates. To fill this gap, in the paper, we propose a class of fast and effective adaptive methods for solving bilevel optimization problems that the outer problem is possibly nonconvex and the inner problem is strongly-convex. Specifically, we propose a fast single-loop BiAdam algorithm based on the basic momentum technique, which achieves a sample complexity of $\tilde{O}(\epsilon^{-4})$ for finding an $\epsilon$-stationary point. At the same time, we propose an accelerated version of BiAdam algorithm (VR-BiAdam) by using variance reduced technique, which reaches the best known sample complexity of $\tilde{O}(\epsilon^{-3})$. To further reduce computation in estimating derivatives, we propose a fast single-loop stochastic approximated BiAdam algorithm (saBiAdam) by avoiding the Hessian inverse, which still achieves a sample complexity of $\tilde{O}(\epsilon^{-4})$ without large batches. We further present an accelerated version of saBiAdam algorithm (VR-saBiAdam), which also reaches the best known sample complexity of $\tilde{O}(\epsilon^{-3})$. We apply the unified adaptive matrices to our methods as the SUPER-ADAM \citep{huang2021super}, which including many types of adaptive learning rates. Moreover, our framework can flexibly use the momentum and variance reduced techniques. In particular, we provide a useful convergence analysis framework for both the constrained and unconstrained bilevel optimization. To the best of our knowledge, we first study the adaptive bilevel optimization methods with adaptive learning rates.

SUPER-ADAM: Faster and Universal Framework of Adaptive Gradients

Adaptive gradient methods have shown excellent performance for solving many machine learning problems. Although multiple adaptive methods were recently studied, they mainly focus on either empirical or theoretical aspects and also only work for specific problems by using specific adaptive learning rates. It is desired to design a universal framework for practical algorithms of adaptive gradients with theoretical guarantee to solve general problems. To fill this gap, we propose a faster and universal framework of adaptive gradients (i.e., SUPER-ADAM) by introducing a universal adaptive matrix that includes most existing adaptive gradient forms. Moreover, our framework can flexibly integrates the momentum and variance reduced techniques. In particular, our novel framework provides the convergence analysis support for adaptive gradient methods under the nonconvex setting. In theoretical analysis, we prove that our new algorithm can achieve the best known complexity of $\tilde{O}(\epsilon^{-3})$ for finding an $\epsilon$-stationary point of nonconvex optimization, which matches the lower bound for stochastic smooth nonconvex optimization. In numerical experiments, we employ various deep learning tasks to validate that our algorithm consistently outperforms the existing adaptive algorithms.

Bregman Gradient Policy Optimization

In this paper, we design a novel Bregman gradient policy optimization framework for reinforcement learning based on Bregman divergences and momentum techniques. Specifically, we propose a Bregman gradient policy optimization (BGPO) algorithm based on the basic momentum technique and mirror descent iteration. At the same time, we present an accelerated Bregman gradient policy optimization (VR-BGPO) algorithm based on a momentum variance-reduced technique. Moreover, we introduce a convergence analysis framework for our Bregman gradient policy optimization under the nonconvex setting. Specifically, we prove that BGPO achieves the sample complexity of $\tilde{O}(\epsilon^{-4})$ for finding $\epsilon$-stationary point only requiring one trajectory at each iteration, and VR-BGPO reaches the best known sample complexity of $\tilde{O}(\epsilon^{-3})$ for finding an $\epsilon$-stationary point, which also only requires one trajectory at each iteration. In particular, by using different Bregman divergences, our methods unify many existing policy optimization algorithms and their new variants such as the existing (variance-reduced) policy gradient algorithms and (variance-reduced) natural policy gradient algorithms. Extensive experimental results on multiple reinforcement learning tasks demonstrate the efficiency of our new algorithms.

Compositional Federated Learning: Applications in Distributionally Robust Averaging and Meta Learning

In the paper, we propose an effective and efficient Compositional Federated Learning (ComFedL) algorithm for solving a new compositional Federated Learning (FL) framework, which frequently appears in many machine learning problems with a hierarchical structure such as distributionally robust federated learning and model-agnostic meta learning (MAML). Moreover, we study the convergence analysis of our ComFedL algorithm under some mild conditions, and prove that it achieves a fast convergence rate of $O(\frac{1}{\sqrt{T}})$, where $T$ denotes the number of iteration. To the best of our knowledge, our algorithm is the first work to bridge federated learning with composition stochastic optimization. In particular, we first transform the distributionally robust FL (i.e., a minimax optimization problem) into a simple composition optimization problem by using KL divergence regularization. At the same time, we also first transform the distribution-agnostic MAML problem (i.e., a minimax optimization problem) into a simple composition optimization problem. Finally, we apply two popular machine learning tasks, i.e., distributionally robust FL and MAML to demonstrate the effectiveness of our algorithm.

Learning Sampling Policy for Faster Derivative Free Optimization

Zeroth-order (ZO, also known as derivative-free) methods, which estimate the gradient only by two function evaluations, have attracted much attention recently because of its broad applications in machine learning community. The two function evaluations are normally generated with random perturbations from standard Gaussian distribution. To speed up ZO methods, many methods, such as variance reduced stochastic ZO gradients and learning an adaptive Gaussian distribution, have recently been proposed to reduce the variances of ZO gradients. However, it is still an open problem whether there is a space to further improve the convergence of ZO methods. To explore this problem, in this paper, we propose a new reinforcement learning based ZO algorithm (ZO-RL) with learning the sampling policy for generating the perturbations in ZO optimization instead of using random sampling. To find the optimal policy, an actor-critic RL algorithm called deep deterministic policy gradient (DDPG) with two neural network function approximators is adopted. The learned sampling policy guides the perturbed points in the parameter space to estimate a more accurate ZO gradient. To the best of our knowledge, our ZO-RL is the first algorithm to learn the sampling policy using reinforcement learning for ZO optimization which is parallel to the existing methods. Especially, our ZO-RL can be combined with existing ZO algorithms that could further accelerate the algorithms. Experimental results for different ZO optimization problems show that our ZO-RL algorithm can effectively reduce the variances of ZO gradient by learning a sampling policy, and converge faster than existing ZO algorithms in different scenarios.

Data Augmentation for Object Detection via Differentiable Neural Rendering

It is challenging to train a robust object detector under the supervised learning setting when the annotated data are scarce. Thus, previous approaches tackling this problem are in two categories: semi-supervised learning models that interpolate labeled data from unlabeled data, and self-supervised learning approaches that exploit signals within unlabeled data via pretext tasks. To seamlessly integrate and enhance existing supervised object detection methods, in this work, we focus on addressing the data scarcity problem from a fundamental viewpoint without changing the supervised learning paradigm. We propose a new offline data augmentation method for object detection, which semantically interpolates the training data with novel views. Specifically, our new system generates controllable views of training images based on differentiable neural rendering, together with corresponding bounding box annotations which involve no human intervention. Firstly, we extract and project pixel-aligned image features into point clouds while estimating depth maps. We then re-project them with a target camera pose and render a novel-view 2d image. Objects in the form of keypoints are marked in point clouds to recover annotations in new views. Our new method is fully compatible with online data augmentation methods, such as affine transform, image mixup, etc. Extensive experiments show that our method, as a cost-free tool to enrich images and labels, can significantly boost the performance of object detection systems with scarce training data. Code is available at \url{https://github.com/Guanghan/DANR}.

Doubly Contrastive Deep Clustering

Deep clustering successfully provides more effective features than conventional ones and thus becomes an important technique in current unsupervised learning. However, most deep clustering methods ignore the vital positive and negative pairs introduced by data augmentation and further the significance of contrastive learning, which leads to suboptimal performance. In this paper, we present a novel Doubly Contrastive Deep Clustering (DCDC) framework, which constructs contrastive loss over both sample and class views to obtain more discriminative features and competitive results. Specifically, for the sample view, we set the class distribution of the original sample and its augmented version as positive sample pairs and set one of the other augmented samples as negative sample pairs. After that, we can adopt the sample-wise contrastive loss to pull positive sample pairs together and push negative sample pairs apart. Similarly, for the class view, we build the positive and negative pairs from the sample distribution of the class. In this way, two contrastive losses successfully constrain the clustering results of mini-batch samples in both sample and class level. Extensive experimental results on six benchmark datasets demonstrate the superiority of our proposed model against state-of-the-art methods. Particularly in the challenging dataset Tiny-ImageNet, our method leads 5.6\% against the latest comparison method. Our code will be available at \url{https://github.com/ZhiyuanDang/DCDC}.

Secure Bilevel Asynchronous Vertical Federated Learning with Backward Updating

Vertical federated learning (VFL) attracts increasing attention due to the emerging demands of multi-party collaborative modeling and concerns of privacy leakage. In the real VFL applications, usually only one or partial parties hold labels, which makes it challenging for all parties to collaboratively learn the model without privacy leakage. Meanwhile, most existing VFL algorithms are trapped in the synchronous computations, which leads to inefficiency in their real-world applications. To address these challenging problems, we propose a novel {\bf VF}L framework integrated with new {\bf b}ackward updating mechanism and {\bf b}ilevel asynchronous parallel architecture (VF{${\textbf{B}}^2$}), under which three new algorithms, including VF{${\textbf{B}}^2$}-SGD, -SVRG, and -SAGA, are proposed. We derive the theoretical results of the convergence rates of these three algorithms under both strongly convex and nonconvex conditions. We also prove the security of VF{${\textbf{B}}^2$} under semi-honest threat models. Extensive experiments on benchmark datasets demonstrate that our algorithms are efficient, scalable and lossless.

Optimizing Large-Scale Hyperparameters via Automated Learning Algorithm

Modern machine learning algorithms usually involve tuning multiple (from one to thousands) hyperparameters which play a pivotal role in terms of model generalizability. Black-box optimization and gradient-based algorithms are two dominant approaches to hyperparameter optimization while they have totally distinct advantages. How to design a new hyperparameter optimization technique inheriting all benefits from both approaches is still an open problem. To address this challenging problem, in this paper, we propose a new hyperparameter optimization method with zeroth-order hyper-gradients (HOZOG). Specifically, we first exactly formulate hyperparameter optimization as an A-based constrained optimization problem, where A is a black-box optimization algorithm (such as deep neural network). Then, we use the average zeroth-order hyper-gradients to update hyperparameters. We provide the feasibility analysis of using HOZOG to achieve hyperparameter optimization. Finally, the experimental results on three representative hyperparameter (the size is from 1 to 1250) optimization tasks demonstrate the benefits of HOZOG in terms of simplicity, scalability, flexibility, effectiveness and efficiency compared with the state-of-the-art hyperparameter optimization methods.