Steven James

Steven James
University of the Witwatersrand | wits · School of Computer Science

PhD

About

40
Publications
6,715
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
92
Citations
Introduction

Publications

Publications (40)
Preprint
We propose world value functions (WVFs), a type of goal-oriented general value function that represents how to solve not just a given task, but any other goal-reaching task in an agent's environment. This is achieved by equipping an agent with an internal goal space defined as all the world states where it experiences a terminal transition. The age...
Preprint
Full-text available
A major challenge in reinforcement learning is specifying tasks in a manner that is both interpretable and verifiable. One common approach is to specify tasks through reward machines -- finite state machines that encode the task to be solved. We introduce skill machines, a representation that can be learned directly from these reward machines that...
Preprint
Full-text available
An open problem in artificial intelligence is how to learn and represent knowledge that is sufficient for a general agent that needs to solve multiple tasks in a given world. In this work we propose world value functions (WVFs), which are a type of general value function with mastery of the world - they represent not only how to solve a given task,...
Preprint
Full-text available
In this work, we investigate the properties of data that cause popular representation learning approaches to fail. In particular, we find that in environments where states do not significantly overlap, variational autoencoders (VAEs) fail to learn useful features. We demonstrate this failure in a simple gridworld domain, and then provide a solution...
Preprint
Full-text available
We are concerned with the question of how an agent can acquire its own representations from sensory data. We restrict our focus to learning representations for long-term planning, a class of problems that state-of-the-art learning methods are unable to solve. We propose a framework for autonomously learning state abstractions of an agent's environm...
Conference Paper
Full-text available
We propose a method for autonomously learning an object-centric representation of a continuous and high-dimensional environment that is suitable for planning. Such representations can immediately be transferred between tasks that share the same types of objects, resulting in agents that require fewer samples to learn a model of a new task. We first...
Conference Paper
Full-text available
We leverage logical composition in reinforcement learning to create a framework that enables an agent to autonomously determine whether a new task can be immediately solved using its existing abilities, or whether a task-specific skill should be learned. In the latter case, the proposed algorithm also enables the agent to learn the new task faster...
Preprint
Full-text available
Using function approximation to represent a value function is necessary for continuous and high-dimensional state spaces. Linear function approximation has desirable theoretical guarantees and often requires less compute and samples than neural networks, but most approaches suffer from an exponential growth in the number of functions as the dimensi...
Preprint
Full-text available
We present a framework that, given a set of skills a robot can perform, abstracts sensor data into symbols that we use to automatically encode the robot's capabilities in Linear Temporal Logic. We specify reactive high-level tasks based on these capabilities, for which a strategy is automatically synthesized and executed on the robot, if the task i...
Preprint
Full-text available
Procedurally generated video game content has the potential to drastically reduce the content creation budget of game developers and large studios. However, adoption is hindered by limitations such as slow generation, as well as low quality and diversity of content. We introduce an evolutionary search-based approach for evolving level generators us...
Article
Full-text available
Graph neural networks (GNNs) build on the success of deep learning models by extending them for use in graph spaces. Transfer learning has proven extremely successful for traditional deep learning problems, resulting in faster training and improved performance. Despite the increasing interest in GNNs and their use cases, there is little research on...
Preprint
Full-text available
Learning disentangled representations with variational autoencoders (VAEs) is often attributed to the regularisation component of the loss. In this work, we highlight the interaction between data and the reconstruction term of the loss as the main contributor to disentanglement in VAEs. We note that standardised benchmark datasets are constructed i...
Preprint
Full-text available
Graph neural networks (GNNs) build on the success of deep learning models by extending them for use in graph spaces. Transfer learning has proven extremely successful for traditional deep learning problems: resulting in faster training and improved performance. Despite the increasing interest in GNNs and their use cases, there is little research on...
Preprint
Full-text available
With increasing interest in procedural content generation by academia and game developers alike, it is vital that different approaches can be compared fairly. However, evaluating procedurally generated video game levels is often difficult, due to the lack of standardised, game-independent metrics. In this paper, we introduce two simulation-based ev...
Preprint
Full-text available
We propose a framework that learns to execute natural language instructions in an environment consisting of goal-reaching tasks that share components of their task descriptions. Our approach leverages the compositionality of both value functions and language, with the aim of reducing the sample complexity of learning novel tasks. First, we train a...
Conference Paper
Full-text available
The ability to compose learned skills to solve new tasks is an important property of lifelong-learning agents. In this work, we formalise the logical composition of tasks as a Boolean algebra. This allows us to formulate new tasks in terms of the negation, disjunction and conjunction of a set of base tasks. We then show that by learning goal-orient...
Preprint
Full-text available
The ability to produce novel behaviours from existing skills is an important property of lifelong-learning agents. We build on recent work which formalises a Boolean algebra over the space of tasks and value functions, and show how this can be leveraged to tackle the lifelong learning problem. We propose an algorithm that determines whether a new t...
Conference Paper
Full-text available
We present a framework for autonomously learning a portable representation that describes a collection of low-level continuous environments. We show that these abstract representations can be learned in a task-independent egocentric space specific to the agent that, when grounded with problem-specific information, are provably sufficient for planni...
Article
Recent work in signal propagation theory has shown that dropout limits the depth to which information can propagate through a neural network. In this paper, we investigate the effect of initialisation on training speed and generalisation for ReLU networks within this depth limit. We ask the following research question: given that critical initialis...
Preprint
Full-text available
We propose a framework for defining a Boolean algebra over the space of tasks. This allows us to formulate new tasks in terms of the negation, disjunction and conjunction of a set of base tasks. We then show that by learning goal-oriented value functions and restricting the transition dynamics of the tasks, an agent can solve these new tasks with n...
Preprint
Full-text available
We present a method for learning options from segmented demonstration trajectories. The trajectories are first segmented into skills using nonparametric Bayesian clustering and a reward function for each segment is then learned using inverse reinforcement learning. From this, a set of inferred trajectories for the demonstration are generated. Optio...
Preprint
Full-text available
Deep neural networks are typically too computationally expensive to run in real-time on consumer-grade hardware and low-powered devices. In this paper, we investigate reducing the computational and memory requirements of neural networks through network pruning and quantisation. We examine their efficacy on large networks like AlexNet compared to re...
Preprint
We propose a framework for defining a Boolean algebra over the space of tasks. This allows us to formulate new tasks in terms of the negation, disjunction and conjunction of a set of base tasks. We then show that by learning goal-oriented value functions and restricting the transition dynamics of the tasks, an agent can solve these new tasks with n...
Preprint
Full-text available
Pre-training a deep neural network on the ImageNet dataset is a common practice for training deep learning models, and generally yields improved performance and faster training times. The technique of pre-training on one task and then retraining on a new one is called transfer learning. In this paper we analyse the effectiveness of using deep trans...
Preprint
Full-text available
Recent work in signal propagation theory has shown that dropout limits the depth to which information can propagate through a neural network. In this paper, we investigate the effect of initialisation on training speed and generalisation for ReLU networks within this depth limit. We ask the following research question: given that critical initialis...
Conference Paper
Full-text available
An important property for lifelong-learning agents is the ability to combine existing skills to solve new unseen tasks. In general, however, it is unclear how to compose existing skills in a principled manner. Under the assumption of deterministic dynamics, we prove that optimal value function composition can be achieved in entropy-regularised rein...
Preprint
Full-text available
We present a framework for autonomously learning a portable representation that describes a collection of low-level continuous environments. We show that these abstract representations can be learned in a task-independent egocentric space specific to the agent that, when grounded with problem-specific information, are provably sufficient for planni...
Preprint
Full-text available
Parameterised actions in reinforcement learning are composed of discrete actions with continuous action-parameters. This provides a framework for solving complex domains that require combining high-level actions with flexible control. The recent P-DQN algorithm extends deep Q-networks to learn over such action spaces. However, it treats all action-...
Conference Paper
Full-text available
We present a framework for autonomously learning a portable symbolic representation that describes a collection of low-level continuous environments. We show that abstract representations can be learned in a task-independent space specific to the agent that, when combined with problem-specific information, can be used for planning. We demonstrate k...
Preprint
Full-text available
An important property for lifelong-learning agents is the ability to combine existing skills to solve unseen tasks. In general, however, it is unclear how to compose skills in a principled way. We provide a "recipe" for optimal value function composition in entropy-regularised reinforcement learning (RL) and then extend this to the standard RL sett...
Conference Paper
Full-text available
An open question in artificial intelligence is how to learn useful representations of the real world. One approach is to learn symbols, which represent the world and its contents, as well as models describing the effects on these symbols when interacting with the world. To date, however, research has investigated learning such representations for a...
Article
Monte Carlo Tree Search (MCTS) is a family of directed search algorithms that has gained widespread attention in recent years. Despite the vast amount of research into MCTS, the effect of modifications on the algorithm, as well as the manner in which it performs in various domains, is still not yet fully known. In particular, the effect of using kn...
Conference Paper
Full-text available
Monte Carlo Tree Search (MCTS) is a family of directed search algorithms that has gained widespread attention in recent years. Despite the vast amount of research into MCTS, the effect of modifications on the algorithm, as well as the manner in which it performs in various domains, is still not yet fully known. In particular, the effect of using kn...

Network

Cited By

Projects

Projects (3)
Project
How can we learn abstractions that are transferable between tasks?
Project
Given a set of skills learned previously, how can we combine them to create new skills without further learning?
Archived project
Investigating the effects on biased rollout policies in MCTS