arXiv subjects

Sheila A. McIlraith

Publications and source records attributed to Sheila A. McIlraith.

33 records · Page 2Linked to original sources

arXiv2022

Learning to Follow Instructions in Text-Based Games

Text-based games present a unique class of sequential decision making problem in which agents interact with a partially observable, simulated environment via actions and observations conveyed through natural language. Such observations typically include instructions that, in a reinforcement learning (RL) setting, can directly or indirectly guide a player towards completing reward-worthy tasks. In this work, we study the ability of RL agents to follow such instructions. We conduct experiments that show that the performance of state-of-the-art text-based game agents is largely unaffected by the presence or absence of such instructions, and that these agents are typically unable to execute tasks to completion. To further study and address the task of instruction following, we equip RL agents with an internal structured representation of natural language instructions in the form of Linear Temporal Logic (LTL), a formal language that is increasingly used for temporally extended reward specification in RL. Our framework both supports and highlights the benefit of understanding the temporal semantics of instructions and in measuring progress towards achievement of such a temporally extended behaviour. Experiments with 500+ games in TextWorld demonstrate the superior performance of our approach.

cs.LG↗

arXiv2022

Challenges to Solving Combinatorially Hard Long-Horizon Deep RL Tasks

Deep reinforcement learning has shown promise in discrete domains requiring complex reasoning, including games such as Chess, Go, and Hanabi. However, this type of reasoning is less often observed in long-horizon, continuous domains with high-dimensional observations, where instead RL research has predominantly focused on problems with simple high-level structure (e.g. opening a drawer or moving a robot as fast as possible). Inspired by combinatorially hard optimization problems, we propose a set of robotics tasks which admit many distinct solutions at the high-level, but require reasoning about states and rewards thousands of steps into the future for the best performance. Critically, while RL has traditionally suffered on complex, long-horizon tasks due to sparse rewards, our tasks are carefully designed to be solvable without specialized exploration. Nevertheless, our investigation finds that standard RL methods often neglect long-term effects due to discounting, while general-purpose hierarchical RL approaches struggle unless additional abstract domain knowledge can be exploited.

cs.LG↗

arXiv2022

Reward Machines: Exploiting Reward Function Structure in Reinforcement Learning

Reinforcement learning (RL) methods usually treat reward functions as black boxes. As such, these methods must extensively interact with the environment in order to discover rewards and optimal policies. In most RL applications, however, users have to program the reward function and, hence, there is the opportunity to make the reward function visible -- to show the reward function's code to the RL agent so it can exploit the function's internal structure to learn optimal policies in a more sample efficient manner. In this paper, we show how to accomplish this idea in two steps. First, we propose reward machines, a type of finite state machine that supports the specification of reward functions while exposing reward function structure. We then describe different methodologies to exploit this structure to support learning, including automated reward shaping, task decomposition, and counterfactual reasoning with off-policy learning. Experiments on tabular and continuous domains, across different tasks and RL agents, show the benefits of exploiting reward structure with respect to sample efficiency and the quality of resultant policies. Finally, by virtue of being a form of finite state machine, reward machines have the expressive power of a regular language and as such support loops, sequences and conditionals, as well as the expression of temporally extended properties typical of linear temporal logic and non-Markovian reward specification.

cs.LG↗

arXiv2021

Learning Reward Machines: A Study in Partially Observable Reinforcement Learning

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.

cs.LG↗

arXiv2021

AppBuddy: Learning to Accomplish Tasks in Mobile Apps via Reinforcement Learning

Human beings, even small children, quickly become adept at figuring out how to use applications on their mobile devices. Learning to use a new app is often achieved via trial-and-error, accelerated by transfer of knowledge from past experiences with like apps. The prospect of building a smarter smartphone - one that can learn how to achieve tasks using mobile apps - is tantalizing. In this paper we explore the use of Reinforcement Learning (RL) with the goal of advancing this aspiration. We introduce an RL-based framework for learning to accomplish tasks in mobile apps. RL agents are provided with states derived from the underlying representation of on-screen elements, and rewards that are based on progress made in the task. Agents can interact with screen elements by tapping or typing. Our experimental results, over a number of mobile apps, show that RL agents can learn to accomplish multi-step tasks, as well as achieve modest generalization across different apps. More generally, we develop a platform which addresses several engineering challenges to enable an effective RL training environment. Our AppBuddy platform is compatible with OpenAI Gym and includes a suite of mobile apps and benchmark tasks that supports a diversity of RL research in the mobile app setting.

cs.AI↗

arXiv2021

Be Considerate: Objectives, Side Effects, and Deciding How to Act

Recent work in AI safety has highlighted that in sequential decision making, objectives are often underspecified or incomplete. This gives discretion to the acting agent to realize the stated objective in ways that may result in undesirable outcomes. We contend that to learn to act safely, a reinforcement learning (RL) agent should include contemplation of the impact of its actions on the wellbeing and agency of others in the environment, including other acting agents and reactive processes. We endow RL agents with the ability to contemplate such impact by augmenting their reward based on expectation of future return by others in the environment, providing different criteria for characterizing impact. We further endow these agents with the ability to differentially factor this impact into their decision making, manifesting behavior that ranges from self-centred to self-less, as demonstrated by experiments in gridworld environments.

cs.AI↗

arXiv2020

Planning from Pixels using Inverse Dynamics Models

Learning task-agnostic dynamics models in high-dimensional observation spaces can be challenging for model-based RL agents. We propose a novel way to learn latent world models by learning to predict sequences of future actions conditioned on task completion. These task-conditioned models adaptively focus modeling capacity on task-relevant dynamics, while simultaneously serving as an effective heuristic for planning with sparse rewards. We evaluate our method on challenging visual goal completion tasks and show a substantial increase in performance compared to prior model-free approaches.

cs.LG↗

arXiv2020

Interpretable Sequence Classification via Discrete Optimization

Sequence classification is the task of predicting a class label given a sequence of observations. In many applications such as healthcare monitoring or intrusion detection, early classification is crucial to prompt intervention. In this work, we learn sequence classifiers that favour early classification from an evolving observation trace. While many state-of-the-art sequence classifiers are neural networks, and in particular LSTMs, our classifiers take the form of finite state automata and are learned via discrete optimization. Our automata-based classifiers are interpretable---supporting explanation, counterfactual reasoning, and human-in-the-loop modification---and have strong empirical performance. Experiments over a suite of goal recognition and behaviour classification datasets show our learned automata-based classifiers to have comparable test performance to LSTM-based classifiers, with the added advantage of being interpretable.

cs.LG↗

arXiv2020

The act of remembering: a study in partially observable reinforcement learning

Reinforcement Learning (RL) agents typically learn memoryless policies---policies that only consider the last observation when selecting actions. Learning memoryless policies is efficient and optimal in fully observable environments. However, some form of memory is necessary when RL agents are faced with partial observability. In this paper, we study a lightweight approach to tackle partial observability in RL. We provide the agent with an external memory and additional actions to control what, if anything, is written to the memory. At every step, the current memory state is part of the agent's observation, and the agent selects a tuple of actions: one action that modifies the environment and another that modifies the memory. When the external memory is sufficiently expressive, optimal memoryless policies yield globally optimal solutions. Unfortunately, previous attempts to use external memory in the form of binary memory have produced poor results in practice. Here, we investigate alternative forms of memory in support of learning effective memoryless policies. Our novel forms of memory outperform binary and LSTM-based memory in well-established partially observable domains.

cs.LG↗

arXiv2020

Towards the Role of Theory of Mind in Explanation

Theory of Mind is commonly defined as the ability to attribute mental states (e.g., beliefs, goals) to oneself, and to others. A large body of previous work - from the social sciences to artificial intelligence - has observed that Theory of Mind capabilities are central to providing an explanation to another agent or when explaining that agent's behaviour. In this paper, we build and expand upon previous work by providing an account of explanation in terms of the beliefs of agents and the mechanism by which agents revise their beliefs given possible explanations. We further identify a set of desiderata for explanations that utilize Theory of Mind. These desiderata inform our belief-based account of explanation.

cs.AI↗

arXiv2019

Towards Neural-Guided Program Synthesis for Linear Temporal Logic Specifications

Synthesizing a program that realizes a logical specification is a classical problem in computer science. We examine a particular type of program synthesis, where the objective is to synthesize a strategy that reacts to a potentially adversarial environment while ensuring that all executions satisfy a Linear Temporal Logic (LTL) specification. Unfortunately, exact methods to solve so-called LTL synthesis via logical inference do not scale. In this work, we cast LTL synthesis as an optimization problem. We employ a neural network to learn a Q-function that is then used to guide search, and to construct programs that are subsequently verified for correctness. Our method is unique in combining search with deep learning to realize LTL synthesis. In our experiments the learned Q-function provides effective guidance for synthesis problems with relatively small specifications.

cs.AI↗

arXiv2019

Towards Empathetic Planning

Critical to successful human interaction is a capacity for empathy - the ability to understand and share the thoughts and feelings of another. As Artificial Intelligence (AI) systems are increasingly required to interact with humans in a myriad of settings, it is important to enable AI to wield empathy as a tool to benefit those it interacts with. In this paper, we work towards this goal by bringing together a number of important concepts: empathy, AI planning, and reasoning in the presence of knowledge and belief. We formalize the notion of Empathetic Planning which is informed by the beliefs and affective state of the empathizee. We appeal to an epistemic logic framework to represent the beliefs of the empathizee and propose AI planning-based computational approaches to compute empathetic solutions. We illustrate the potential benefits of our approach by conducting a study where we evaluate participants' perceptions of the agent's empathetic abilities and assistive capabilities.

cs.AI↗

arXiv2018

Finite LTL Synthesis with Environment Assumptions and Quality Measures

In this paper, we investigate the problem of synthesizing strategies for linear temporal logic (LTL) specifications that are interpreted over finite traces -- a problem that is central to the automated construction of controllers, robot programs, and business processes. We study a natural variant of the finite LTL synthesis problem in which strategy guarantees are predicated on specified environment behavior. We further explore a quantitative extension of LTL that supports specification of quality measures, utilizing it to synthesize high-quality strategies. We propose new notions of optimality and associated algorithms that yield strategies that best satisfy specified quality measures. Our algorithms utilize an automata-game approach, positioning them well for future implementation via existing state-of-the-art techniques.

cs.LO↗

arXiv2016

Finite LTL Synthesis is EXPTIME-complete

LTL synthesis -- the construction of a function to satisfy a logical specification formulated in Linear Temporal Logic -- is a 2EXPTIME-complete problem with relevant applications in controller synthesis and a myriad of artificial intelligence applications. In this research note we consider De Giacomo and Vardi's variant of the synthesis problem for LTL formulas interpreted over finite rather than infinite traces. Rather surprisingly, given the existing claims on complexity, we establish that LTL synthesis is EXPTIME-complete for the finite interpretation, and not 2EXPTIME-complete as previously reported. Our result coincides nicely with the planning perspective where non-deterministic planning with full observability is EXPTIME-complete and partial observability increases the complexity to 2EXPTIME-complete; a recent related result for LTL synthesis shows that in the finite case with partial observability, the problem is 2EXPTIME-complete.

cs.LO↗

arXiv2009

On Planning with Preferences in HTN

In this paper, we address the problem of generating preferred plans by combining the procedural control knowledge specified by Hierarchical Task Networks (HTNs) with rich qualitative user preferences. The outcome of our work is a language for specifyin user preferences, tailored to HTN planning, together with a provably optimal preference-based planner, HTNPLAN, that is implemented as an extension of SHOP2. To compute preferred plans, we propose an approach based on forward-chaining heuristic search. Our heuristic uses an admissible evaluation function measuring the satisfaction of preferences over partial plans. Our empirical evaluation demonstrates the effectiveness of our HTNPLAN heuristics. We prove our approach sound and optimal with respect to the plans it generates by appealing to a situation calculus semantics of our preference language and of HTN planning. While our implementation builds on SHOP2, the language and techniques proposed here are relevant to a broad range of HTN planners.

cs.AI↗