Deep Reinforcement Learning (DRL) has led to many recent breakthroughs on complex control tasks, such as defeating the best human player in the game of Go. However, decisions made by the DRL agent are not explainable, hindering its applicability in safety-critical settings. Viper, a recently proposed technique, constructs a decision tree policy by mimicking the DRL agent. Decision trees are interpretable as each action made can be traced back to the decision rule path that lead to it. However, one global decision tree approximating the DRL policy has significant limitations with respect to the geometry of decision boundaries. We propose Mo\"ET, a more expressive, yet still interpretable model based on Mixture of Experts, consisting of a gating function that partitions the state space, and multiple decision tree experts that specialize on different partitions. We propose a training procedure to support non-differentiable decision tree experts and integrate it into imitation learning procedure of Viper. We evaluate our algorithm on four OpenAI gym environments, and show that the policy constructed in such a way is more performant and better mimics the DRL agent by lowering mispredictions and increasing the reward. We also show that Mo\"ET policies are amenable for verification using off-the-shelf automated theorem provers such as Z3.

When function approximation is deployed in reinforcement learning (RL), the same problem may be formulated in different ways, often by treating a pre-processing step as a part of the environment or as part of the agent. As a consequence, fundamental concepts in RL, such as (optimal) value functions, are not uniquely defined as they depend on where we draw this agent-environment boundary, causing problems in theoretical analyses that provide optimality guarantees. We address this issue via a simple and novel boundary-invariant analysis of Fitted Q-Iteration, a representative RL algorithm, where the assumptions and the guarantees are invariant to the choice of boundary. We also discuss closely related issues on state resetting and Monte-Carlo Tree Search, deterministic vs stochastic systems, imitation learning, and the verifiability of theoretical assumptions from data.

We provide a process to modify a neural network to an equivariant one, which we call {\em equivarification}. As an illustration, we build an equivariant neural network for image classification by equivarifying a convolutional neural network.

In this work we examine the problematic associated to the development of machine learning models to achieve robust generalization capabilities on common-task multiple-database scenarios. Referred as the ''database variability problem'', we focus on a specific medical domain (sleep staging in Sleep Medicine) to show the non-triviality of translating the estimated model's local generalization capabilities to independent external databases. We analyze some of the scalability problems when multiple-database data are used as input to train a single learning model. Then, we introduce a novel approach based on an ensemble of local models, and we show its advantages in terms of inter-database generalization performance and data scalability. Further on, we analyze different model configurations and data pre-processing techniques to evaluate their effects over the overall generalization performance. For this purpose we carry out experimentation involving several sleep databases evaluating different machine learning models based on Convolutional Neural Networks

Background Based Conversations (BBCs) have been developed to make dialogue systems generate more informative and natural responses by leveraging background knowledge. Existing methods for BBCs can be grouped into two categories: extraction-based methods and generation-based methods. The former extract spans frombackground material as responses that are not necessarily natural. The latter generate responses thatare natural but not necessarily effective in leveraging background knowledge. In this paper, we focus on generation-based methods and propose a model, namely Context-aware Knowledge Pre-selection (CaKe), which introduces a pre-selection process that uses dynamic bi-directional attention to improve knowledge selection by using the utterance history context as prior information to select the most relevant background material. Experimental results show that our model is superior to current state-of-the-art baselines, indicating that it benefits from the pre-selection process, thus improving in-formativeness and fluency.

We consider the problem of two-player zero-sum game. In this setting, there are two agents working against each other. Both the agents observe the same state and the objective of the agents is to compute a strategy profile that maximizes their rewards. However, the reward of the second agent is negative of reward obtained by the first agent. Therefore, the objective of the second agent is to minimize the total reward obtained by the first agent. This problem is formulated as a min-max Markov game in the literature. The solution of this game, which is the max-min reward (of first player), starting from a given state is called the equilibrium value of the state. In this work, we compute the solution of the two-player zero-sum game utilizing the technique of successive relaxation. Successive relaxation has been successfully applied in the literature to compute a faster value iteration algorithm in the context of Markov Decision Processes. We extend the concept of successive relaxation to the two-player zero-sum games. We prove that, under a special structure, this technique computes the optimal solution faster than the techniques in the literature. We then derive a generalized minimax Q-learning algorithm that computes the optimal policy when the model information is not known. Finally, we prove the convergence of the proposed generalized minimax Q-learning algorithm.

How can we estimate the importance of nodes in a knowledge graph (KG)? A KG is a multi-relational graph that has proven valuable for many tasks including question answering and semantic search. In this paper, we present GENI, a method for tackling the problem of estimating node importance in KGs, which enables several downstream applications such as item recommendation and resource allocation. While a number of approaches have been developed to address this problem for general graphs, they do not fully utilize information available in KGs, or lack flexibility needed to model complex relationship between entities and their importance. To address these limitations, we explore supervised machine learning algorithms. In particular, building upon recent advancement of graph neural networks (GNNs), we develop GENI, a GNN-based method designed to deal with distinctive challenges involved with predicting node importance in KGs. Our method performs an aggregation of importance scores instead of aggregating node embeddings via predicate-aware attention mechanism and flexible centrality adjustment. In our evaluation of GENI and existing methods on predicting node importance in real-world KGs with different characteristics, GENI achieves 5-17% higher NDCG@100 than the state of the art.

One important problem in statistics and machine learning is to learn a finite mixture of distributions [18, 24]. In the parametric setting where the functional form of the distribution is known, this problem is to estimate parameters (e.g., mean and covariance) that specify the distribution of each mixture component. The parameter estimation problem for mixture models is inherently nonconvex, posing challenges for both computation and analysis. While many algorithms have been proposed, rigorous performance guarantees are often elusive. One exception is the Gaussian Mixture Model (GMM), for which much theoretical progress has been made in recent years. The goal of this paper is to study algorithmic guarantees for a much broader class of mixture models, namely log-concave distributions.

Machine learning develops rapidly, which has made many theoretical breakthroughs and is widely applied in various fields. Optimization, as an important part of machine learning, has attracted much attention of researchers. With the exponential growth of data amount and the increase of model complexity, optimization methods in machine learning face more and more challenges. A lot of work on solving optimization problems or improving optimization methods in machine learning has been proposed successively. The systematic retrospect and summary of the optimization methods from the perspective of machine learning are of great significance, which can offer guidance for both developments of optimization and machine learning research. In this paper, we first describe the optimization problems in machine learning. Then, we introduce the principles and progresses of commonly used optimization methods. Next, we summarize the applications and developments of optimization methods in some popular machine learning fields. Finally, we explore and give some challenges and open problems for the optimization in machine learning.

Despite the phenomenal success of deep neural networks in a broad range of learning tasks, there is a lack of theory to understand the way they work. In particular, Convolutional Neural Networks (CNNs) are known to perform much better than Fully-Connected Networks (FCNs) on spatially structured data: the architectural structure of CNNs benefits from prior knowledge on the features of the data, for instance their translation invariance. The aim of this work is to understand this fact through the lens of dynamics in the loss landscape. We introduce a method that maps a CNN to its equivalent FCN (denoted as eFCN). Such an embedding enables the comparison of CNN and FCN training dynamics directly in the FCN space. We use this method to test a new training protocol, which consists in training a CNN, embedding it to FCN space at a certain 'switch time' $t_w$, then resuming the training in FCN space. We observe that for all switch times, the deviation from the CNN subspace is small, and the final performance reached by the eFCN is higher than that reachable by the standard FCN. More surprisingly, for some intermediate switch times, the eFCN even outperforms the CNN it stemmed from. The practical interest of our protocol is limited by the very large size of the highly sparse eFCN. However, it offers an interesting insight into the persistence of the architectural bias under the stochastic gradient dynamics even in the presence of a huge number of additional degrees of freedom. It shows the existence of some rare basins in the FCN space associated with very good generalization. These can be accessed thanks to the CNN prior, and are otherwise missed.