Méthodes infographiques pour l'apprentissage des agents autonomes virtuels

There are two primary approaches to behavioural animation of an Autonomous Virtual Agent (AVA). The first one, or behavioural model, defines how AVA reacts to the current state of its environment. In the second one, or cognitive model, this AVA uses a thought process allowing it to deliberate over its possible actions. Despite the success of these approaches in several domains, there are two notable limitations which we address in this thesis. First, cognitive models are traditionally very slow to execute, as a tree search, in the form of mapping: states → actions, must be performed. On the one hand, an AVA can only make sub-optimal decisions and, on the other hand, the number of AVAs that can be used simultaneously in real-time is limited. These constraints restrict their applications to a small set of candidate actions. Second, cognitive and behavioural models can act unexpectedly, producing undesirable behaviour in certain regions of the state space. This is because it may be impossible to exhaustively test them for the entire state space, especially if the state space is continuous. This can be worrisome for end-user applications involving AVAs, such as training simulators for cars and aeronautics. Our contributions include the design of novel learning methods for approximating behavioural and cognitive models. They address the problem of input selection helped by a novel architecture ALifeE including virtual sensors and perception, regardless of the machine learning technique utilized. The input dimensionality must be kept as small as possible, this is due to the "curse of dimensionality", well known in machine learning. Thus, ALifeE simplifies and speeds up the process for the designer.

Autonomous virtual agents learning a cognitive model and evolving

Toni Conde, Daniel Thalmann

In this paper, we propose a new integration approach to simulate an autonomous virtual agent's cognitive learning of a task for interactive virtual environment applications. Our research focuses on the behavioural animation of virtual humans capable of acting independently. Our contribution is important because we present a solution for fast learning with evolution. We propose the concept of a learning unit architecture that functions as a control unit of the autonomous virtual agent's brain. Although our technique has, proved to be effective in our case study, there is no guarantee that it will work for every imaginable autonomous virtual agent and virtual environment. The results are illustrated in a domain that requires effective coordination of behaviours, such as driving a car inside a virtual city

2005

Tractable Approaches to Learning and Planning in High Dimensions

Leonidas Lefakis

The continuous increase, witnessed in the last decade, of both the amount of available data and the areas of application of machine learning, has lead to a demand for both learning and planning algorithms that are capable of handling large-scale problems. Thus scalability has become an important characteristic of modern machine learning algorithms. In this thesis we concentrate on tractable approaches to both learning and planning, capable of solving problems in high-dimensional spaces. We investigate multiple forms of high-dimensionality, where dimensionality can refer to the number of samples to be processed, the number of samples available for training, the dimensionality of the feature space, or even the dimensionality of the state-space. We present work on classifier cascades, developing a novel framework for interpreting, training, and employing such structures of classifiers. A Boosting algorithm developed in this framework highlights its ability to jointly train the classifiers at the cascade nodes. The work presented on goal-planning also results in a novel framework, this time for automatic macro-action discovery in imitation learning, specifically addressing problems with large state-spaces as well as with observation signals of high-dimensionality. The proposed DPBoost algorithm is capable of solving very complex goal-planning tasks where previous state-of-the-art fails. Addressing computer hardware limitations, and in particular memory limitations, we propose a new learning framework, which we call reservoir learning, which attempts to directly account for this limited memory during the training process. In particular we propose to consider learning in the presence of a reservoir in which the learner may store samples. Based on Boosting we develop a novel strategy for populating this reservoir which avoids redundancy by taking into account the joint behavior of samples. Finally high-dimensionality is also addressed in its most typical setting, that is when the feature space is large. We propose novel feature selection algorithms in the context of classification. Based on information theoretic tools, these algorithms attempt to maximize the joint mutual information between the selected, continuous, variables and the, discrete, class label variable.

EPFL2014

Model-based reinforcement learning and navigation in animals and machines

Dane Sterling Corneil

For decades, neuroscientists and psychologists have observed that animal performance on spatial navigation tasks suggests an internal learned map of the environment. More recently, map-based (or model-based) reinforcement learning has become a highly active research area in machine learning. With a learned model of their environment, both animals and artificial agents can generalize between tasks and learn rapidly. In this thesis, I present approaches for developing efficient model--based behaviour in machines and explaining model--based behaviour in animals. From a neuroscience perspective, I focus on the hippocampus, believed to be a major substrate of model-based behaviour in the brain. I consider how hippocampal connectivity enable path--finding between different locations in an environment. The model describes how environments with boundaries and barriers can be represented in recurrent neural networks (i.e. attractor networks), and how the transient activity in these networks, after being stimulated with a goal location, could be used for determining a path to the goal. I also propose how the connectivity of these map--like networks can be learned from the spatial firing patterns observed in the input pathway to the hippocampus (i.e. grid cells and border cells). From a machine learning perspective, I describe a reinforcement learning model that integrates model-based methods and "episodic control", an approach to reinforcement learning based on episodic memory. According to episodic control, the agent learns how to act in the environment by storing snapshot-like memories of its observations, then comparing its current observations to similar snapshot memories where it took an action that resulted in high reward. In our approach, the agent augments these real-world memories with episodes simulated offline using a learned model of the environment. These ``simulated memories'' allow the agent to adapt faster when the reward locations change. Next, I describe Variational State Tabulation (VaST), a model--based method for learning quickly with continuous and high-dimensional observations (like those found in 3D navigation tasks). The VaST agent learns to map its observations to a limited number of discrete abstract states, and build a transition model over those abstract states. The long--term values of different actions in each state are updated continuously and efficiently in the background as the agent explores the environment. I show how the VaST agent can learn faster than other state-of-the-art algorithms, even changing its policy after a single new experience, and how it can respond quickly to changing rewards in complex 3D environments. The models I present allow the agent to rapidly adapt to changing goals and rewards, a key component of intelligence. They use a combination of features attributed to model-based and episodic controllers, suggesting that the division between the two fields is not strict. I therefore also consider the consequences of these findings on theories of model-based learning, episodic control and hippocampal function.