Federated learning (FL) algorithms usually sample a fraction of clients in each round (partial participation) when the number of participants is large and the server's communication bandwidth is limited. Recent works on the convergence analysis of FL have focused on unbiased client sampling, e.g., sampling uniformly at random, which suffers from slow wall-clock time for convergence due to high degrees of system heterogeneity and statistical heterogeneity. This paper aims to design an adaptive client sampling algorithm that tackles both system and statistical heterogeneity to minimize the wall-clock convergence time. We obtain a new tractable convergence bound for FL algorithms with arbitrary client sampling probabilities. Based on the bound, we analytically establish the relationship between the total learning time and sampling probabilities, which results in a non-convex optimization problem for training time minimization. We design an efficient algorithm for learning the unknown parameters in the convergence bound and develop a low-complexity algorithm to approximately solve the non-convex problem. Experimental results from both hardware prototype and simulation demonstrate that our proposed sampling scheme significantly reduces the convergence time compared to several baseline sampling schemes. Notably, our scheme in hardware prototype spends 73% less time than the uniform sampling baseline for reaching the same target loss.

Understanding the interactions formed between a ligand and its molecular target is key to guiding the optimization of molecules. Different experimental and computational methods have been applied to better understanding these intermolecular interactions. Here researchers report a method based on geometric deep learning that is capable of predicting the binding conformations of ligands to protein targets. The model learns a statistical potential based on the distance likelihood, which is tailor-made for each ligand–target pair. This potential can be coupled with global optimization algorithms to reproduce the experimental binding conformations of ligands.

Reinforcement learning agents perform well when presented with inputs within the distribution of those encountered during training. However, they are unable to respond effectively when faced with novel, out-of-distribution events, until they have undergone additional training. This paper presents an online, data-driven, emergency-response method that aims to provide autonomous agents the ability to react to unexpected situations that are very different from those it has been trained or designed to address. In such situations, learned policies cannot be expected to perform appropriately since the observations obtained in these novel situations would fall outside the distribution of inputs that the agent has been optimized to handle. The proposed approach devises a customized response to the unforeseen situation sequentially, by selecting actions that minimize the rate of increase of the reconstruction error from a variational auto-encoder. This optimization is achieved online in a data-efficient manner (on the order of 30 data-points) using a modified Bayesian optimization procedure. We demonstrate the potential of this approach in a simulated 3D car driving scenario, in which the agent devises a response in under 2 seconds to avoid collisions with objects it has not seen during training.

We present a federated learning framework that is designed to robustly deliver good predictive performance across individual clients with heterogeneous data. The proposed approach hinges upon a superquantile-based learning objective that captures the tail statistics of the error distribution over heterogeneous clients. We present a stochastic training algorithm which interleaves differentially private client reweighting steps with federated averaging steps. The proposed algorithm is supported with finite time convergence guarantees that cover both convex and non-convex settings. Experimental results on benchmark datasets for federated learning demonstrate that our approach is competitive with classical ones in terms of average error and outperforms them in terms of tail statistics of the error.

Reinforcement learning (RL) is a central problem in artificial intelligence. This problem consists of defining artificial agents that can learn optimal behaviour by interacting with an environment -- where the optimal behaviour is defined with respect to a reward signal that the agent seeks to maximize. Reward machines (RMs) provide a structured, automata-based representation of a reward function that enables an RL agent to decompose an RL problem into structured subproblems that can be efficiently learned via off-policy learning. Here we show that RMs can be learned from experience, instead of being specified by the user, and that the resulting problem decomposition can be used to effectively solve partially observable RL problems. We pose the task of learning RMs as a discrete optimization problem where the objective is to find an RM that decomposes the problem into a set of subproblems such that the combination of their optimal memoryless policies is an optimal policy for the original problem. We show the effectiveness of this approach on three partially observable domains, where it significantly outperforms A3C, PPO, and ACER, and discuss its advantages, limitations, and broader potential.

Policy Search and Model Predictive Control~(MPC) are two different paradigms for robot control: policy search has the strength of automatically learning complex policies using experienced data, while MPC can offer optimal control performance using models and trajectory optimization. An open research question is how to leverage and combine the advantages of both approaches. In this work, we provide an answer by using policy search for automatically choosing high-level decision variables for MPC, which leads to a novel policy-search-for-model-predictive-control framework. Specifically, we formulate the MPC as a parameterized controller, where the hard-to-optimize decision variables are represented as high-level policies. Such a formulation allows optimizing policies in a self-supervised fashion. We validate this framework by focusing on a challenging problem in agile drone flight: flying a quadrotor through fast-moving gates. Experiments show that our controller achieves robust and real-time control performance in both simulation and the real world. The proposed framework offers a new perspective for merging learning and control.

Many state-of-the-art adversarial training methods leverage upper bounds of the adversarial loss to provide security guarantees. Yet, these methods require computations at each training step that can not be incorporated in the gradient for backpropagation. We introduce a new, more principled approach to adversarial training based on a closed form solution of an upper bound of the adversarial loss, which can be effectively trained with backpropagation. This bound is facilitated by state-of-the-art tools from robust optimization. We derive two new methods with our approach. The first method (Approximated Robust Upper Bound or aRUB) uses the first order approximation of the network as well as basic tools from linear robust optimization to obtain an approximate upper bound of the adversarial loss that can be easily implemented. The second method (Robust Upper Bound or RUB), computes an exact upper bound of the adversarial loss. Across a variety of tabular and vision data sets we demonstrate the effectiveness of our more principled approach -- RUB is substantially more robust than state-of-the-art methods for larger perturbations, while aRUB matches the performance of state-of-the-art methods for small perturbations. Also, both RUB and aRUB run faster than standard adversarial training (at the expense of an increase in memory). All the code to reproduce the results can be found at https://github.com/kimvc7/Robustness.

Modern statistical applications often involve minimizing an objective function that may be nonsmooth and/or nonconvex. This paper focuses on a broad Bregman-surrogate algorithm framework including the local linear approximation, mirror descent, iterative thresholding, DC programming and many others as particular instances. The recharacterization via generalized Bregman functions enables us to construct suitable error measures and establish global convergence rates for nonconvex and nonsmooth objectives in possibly high dimensions. For sparse learning problems with a composite objective, under some regularity conditions, the obtained estimators as the surrogate's fixed points, though not necessarily local minimizers, enjoy provable statistical guarantees, and the sequence of iterates can be shown to approach the statistical truth within the desired accuracy geometrically fast. The paper also studies how to design adaptive momentum based accelerations without assuming convexity or smoothness by carefully controlling stepsize and relaxation parameters.

Outliers widely occur in big-data applications and may severely affect statistical estimation and inference. In this paper, a framework of outlier-resistant estimation is introduced to robustify an arbitrarily given loss function. It has a close connection to the method of trimming and includes explicit outlyingness parameters for all samples, which in turn facilitates computation, theory, and parameter tuning. To tackle the issues of nonconvexity and nonsmoothness, we develop scalable algorithms with implementation ease and guaranteed fast convergence. In particular, a new technique is proposed to alleviate the requirement on the starting point such that on regular datasets, the number of data resamplings can be substantially reduced. Based on combined statistical and computational treatments, we are able to perform nonasymptotic analysis beyond M-estimation. The obtained resistant estimators, though not necessarily globally or even locally optimal, enjoy minimax rate optimality in both low dimensions and high dimensions. Experiments in regression, classification, and neural networks show excellent performance of the proposed methodology at the occurrence of gross outliers.

Software configuration tuning is essential for optimizing a given performance objective (e.g., minimizing latency). Yet, due to the software's intrinsically complex configuration landscape and expensive measurement, there has been a rather mild success, particularly in preventing the search from being trapped in local optima. To address this issue, in this paper we take a different perspective. Instead of focusing on improving the optimizer, we work on the level of optimization model and propose a meta multi-objectivization (MMO) model that considers an auxiliary performance objective (e.g., throughput in addition to latency). What makes this model unique is that we do not optimize the auxiliary performance objective, but rather use it to make similarly-performing while different configurations less comparable (i.e. Pareto nondominated to each other), thus preventing the search from being trapped in local optima. Importantly through a new normalization method we show how to effectively use the MMO model without worrying about its weight -- the only yet highly sensitive parameter that can affect its effectiveness. Experiments on 22 cases from 11 real-world software systems/environments confirm that our MMO model with the new normalization performs better than its state-of-the-art single-objective counterparts on 82% cases while achieving up to 2.09x speedup. For 67% of the cases, the new normalization also enables the MMO model to outperform the instance when using it with the normalization used in our prior FSE work under pre-tuned best weights, saving a great amount of resources which would be otherwise necessary to find a good weight. We also demonstrate that the MMO model with the new normalization can consolidate Flash, a recent model-based tuning tool, on 68% of the cases with 1.22x speedup in general.