A framework to boost efficiency of Bayesian inference in probabilistic programs is introduced by embedding a sampler inside a variational posterior approximation, which we call the refined variational approximation. Its strength lies both in ease of implementation and in automatically tuning the sampler parameters to speed up mixing time. Several strategies to approximate the \emph{evidence lower bound} (ELBO) computation are introduced, including a rewriting of the ELBO objective. A specialization towards state-space models is proposed. Experimental evidence of its efficient performance is shown by solving an influence diagram in a high-dimensional space using a conditional variational autoencoder (cVAE) as a deep Bayes classifier; an unconditional VAE on density estimation tasks; and state-space models for time-series data.

We propose an actor-critic, model-free, and online Reinforcement Learning (RL) framework for continuous-state continuous-action Markov Decision Processes (MDPs) when the reward is highly sparse but encompasses a high-level temporal structure. We represent this temporal structure by a finite-state machine and construct an on-the-fly synchronised product with the MDP and the finite machine. The temporal structure acts as a guide for the RL agent within the product, where a modular Deep Deterministic Policy Gradient (DDPG) architecture is proposed to generate a low-level control policy. We evaluate our framework in a Mars rover experiment and we present the success rate of the synthesised policy.

Factored stochastic constraint programming (FSCP) is a formalism to represent multi-stage decision making problems under uncertainty. FSCP models support factorized probabilistic models and involve constraints over decision and random variables. These models have many applications in real-world problems. However, solving these problems requires evaluating the best course of action for each possible outcome of the random variables and hence is computationally challenging. FSCP problems often involve repeated subproblems which ideally should be solved once. In this paper we show how identifying and exploiting these identical subproblems can simplify solving them and leads to a compact representation of the solution. We compile an And-Or search tree to a compact decision diagram. Preliminary experiments show that our proposed method significantly improves the search efficiency by reducing the size of the problem and outperforms the existing methods.

Driven by an increasing need for model interpretability, interpretable models have become strong competitors for black-box models in many real applications. In this paper, we propose a novel type of model where interpretable models compete and collaborate with black-box models. We present the Model-Agnostic Linear Competitors (MALC) for partially interpretable classification. MALC is a hybrid model that uses linear models to locally substitute any black-box model, capturing subspaces that are most likely to be in a class while leaving the rest of the data to the black-box. MALC brings together the interpretable power of linear models and good predictive performance of a black-box model. We formulate the training of a MALC model as a convex optimization. The predictive accuracy and transparency (defined as the percentage of data captured by the linear models) balance through a carefully designed objective function and the optimization problem is solved with the accelerated proximal gradient method. Experiments show that MALC can effectively trade prediction accuracy for transparency and provide an efficient frontier that spans the entire spectrum of transparency.

The ability to interpret the mental state of another agent based on its behavior, also called Theory of Mind (ToM), is crucial for humans in any kind of social interaction. Artificial systems, such as intelligent assistants, would also greatly benefit from such mentalizing capabilities. However, humans and systems alike are bound by limitations in their available computational resources. This raises the need for satisficing mentalizing, reconciling accuracy and efficiency in mental state inference that is good enough for a given situation. In this paper, we present different Bayesian models of ToM reasoning and evaluate them based on actual human behavior data that were generated under different kinds of uncertainties. We propose a Switching approach that combines specialized models, embodying simplifying presumptions, in order to achieve a more statisficing mentalizing compared to a Full Bayesian ToM model.

Emotional Intelligence in Human-Computer Interaction has attracted increasing attention from researchers in multidisciplinary research fields including psychology, computer vision, neuroscience, artificial intelligence, and related disciplines. Human prone to naturally interact with computers face-to-face. Human Expressions is an important key to better link human and computers. Thus, designing interfaces able to understand human expressions and emotions can improve Human-Computer Interaction (HCI) for better communication. In this paper, we investigate HCI via a deep multi-facial patches aggregation network for Face Expression Recognition (FER). Deep features are extracted from facial parts and aggregated for expression classification. Several problems may affect the performance of the proposed framework like the small size of FER datasets and the high number of parameters to learn. For That, two data augmentation techniques are proposed for facial expression generation to expand the labeled training. The proposed framework is evaluated on the extended Cohn-Konade dataset (CK+) and promising results are achieved.

We consider the problem of computing optimal policies in average-r eward Markov decision processes. This classical problem can be formulated as a linear program dire ctly amenable to saddle-point optimization methods, albeit with a number of variables that is linear in the n umber of states. T o address this issue, recent work has considered a linearly relaxed version of the res ulting saddle-point problem. Our work aims at achieving a better understanding of this relaxed optimization pro blem by characterizing the conditions necessary for convergence to the optimal policy, and designing a n optimization algorithm enjoying fast convergence rates that are independent of the size of the state s pace.

Directed acyclic graphs, or DAGs, are commonly used to repre sent causal relationships in complex biological systems. For example, in gene regulatory ne tworks, directed edges represent regulatory interactions among genes, which are represente d as nodes of the graph. While causal effects in biological networks can be accurately inferred fro m perturbation experiments [33]-- including single or double gene knockouts [30, 42]--these ar e costly to run. Estimating DAGs from observational data is thus an important exploratory ta sk for generating causal hypotheses [10, 15], and designing more efficient experiments. Since the number of possible directed graphs grows super-ex ponentially in the number of nodes, estimation of DAGs is an NPhard problem [6]. Methods of estimating DAGs from observational data can be broadly categorized into three cl asses. The first class, score-based methods, search over the space of all possible graphs, and at tempt to maximize a goodness-of-fit score, generally using a greedy algorithm.

In this work, we introduce a new approach for the efficient solution of autonomous decision and planning problems, with a special focus on decision making under uncertainty and belief space planning (BSP) in high-dimensional state spaces. Usually, to solve the decision problem, we identify the optimal action, according to some objective function. Instead, we claim that we can sometimes generate and solve an analogous yet simplified decision problem, which can be solved more efficiently. Furthermore, a wise simplification method can lead to the same action selection, or one for which the maximal loss can be guaranteed. This simplification is separated from the state inference, and does not compromise its accuracy, as the selected action would finally be applied on the original state. At first, we develop the concept for general decision problems, and provide a theoretical framework of definitions to allow a coherent discussion. We then practically apply these ideas to BSP problems, in which the problem is simplified by considering a sparse approximation of the initial belief. The scalable sparsification algorithm we provide is able to yield solutions which are guaranteed to be consistent with the original problem. We demonstrate the benefits of the approach in the solution of a highly realistic active-SLAM problem, and manage to significantly reduce computation time, with practically no loss in the quality of solution. This rigorous and fundamental work is conceptually novel, and holds numerous possible extensions.

This paper introduces the modulated Hebbian plus Q network architecture (MOHQA) for solving challenging partially observable Markov decision processes (POMDPs) deep reinforcement learning problems with sparse rewards and confounding observations. The proposed architecture combines a deep Q-network (DQN), and a modulated Hebbian network with neural eligibility traces (MOHN). Bio-inspired neural traces are used to bridge temporal delays between actions and rewards. The purpose is to discover distal cause-effect relationships where confounding observations and sparse rewards cause standard RL algorithms to fail. Each of the two modules of the network (DQN and MOHN) is responsible for different aspects of learning. DQN learns low level features and control, while MOHN contributes to the high-level decisions by bridging rewards with past actions. The strength of the approach is to support a DQN standard framework when temporal difference errors are difficult to compute due to non-observable states. The system is tested on a set of generalized decision making problems encoded as decision tree graphs that deliver delayed rewards after key decision points and confounding observations. The simulations show that the proposed approach helps solve problems that are currently challenging for state-of-the-art deep reinforcement learning algorithms.