Satchuthanan R. Branavan

Reinforcement Learning for Mapping Instructions to Actions

In this paper, we present a reinforcement learning approach for mapping natural language instructions to sequences of executable actions. We assume access to a reward function that defines the quality of the executed actions. During training, the learner repeatedly constructs action sequences for a set of documents, executes those actions, and observes the resulting reward. We use a policy gradient algorithm to estimate the parameters of a log-linear model for action selection. We apply our method to interpret instructions in two domains --- Windows troubleshooting guides and game tutorials. Our results demonstrate that this method can rival supervised learning techniques while requiring few or no annotated training examples.

Learning to Win by Reading Manuals in a Monte-Carlo Framework

This paper presents a novel approach for leveraging automatically extracted textual knowledge to improve the performance of control applications such as games. Our ultimate goal is to enrich a stochastic player with high-level guidance expressed in text. Our model jointly learns to identify text that is relevant to a given game state in addition to learning game strategies guided by the selected text. Our method operates in the Monte-Carlo search framework, and learns both text analysis and game strategies based only on environment feedback. We apply our approach to the complex strategy game Civilization II using the official game manual as the text guide. Our results show that a linguistically-informed game-playing agent significantly outperforms its language-unaware counterpart, yielding a 27% absolute improvement and winning over 78% of games when playing against the built-in AI of Civilization II.

Global Models of Document Structure using Latent Permutations

We present a novel Bayesian topic model for learning discourse-level document structure. Our model leverages insights from discourse theory to constrain latent topic assignments in a way that reflects the underlying organization of document topics. We propose a global model in which both topic selection and ordering are biased to be similar across a collection of related documents. We show that this space of orderings can be elegantly represented using a distribution over permutations called the generalized Mallows model. Our structure-aware approach substantially outperforms alternative approaches for cross-document comparison and single-document segmentation.

Content modeling using latent permutations

We present a novel Bayesian topic model for learning discourse-level document structure. Our model leverages insights from discourse theory to constrain latent topic assignments in a way that reflects the underlying organization of document topics. We propose a global model in which both topic selection and ordering are biased to be similar across a collection of related documents. We show that this space of orderings can be effectively represented using a distribution over permutations called the Generalized Mallows Model. We apply our method to three complementary discourse-level tasks: cross-document alignment, document segmentation, and information ordering. Our experiments show that incorporating our permutation-based model in these applications yields substantial improvements in performance over previously proposed methods.

Non-linear Monte-Carlo search in civilization II

This paper presents a new Monte-Carlo search algorithm for very large sequential decision-making problems. We apply non-linear regression within Monte-Carlo search, online, to estimate a state-action value function from the outcomes of random roll-outs. This value function generalizes between related states and actions, and can therefore provide more accurate evaluations after fewer rollouts. A further significant advantage of this approach is its ability to automatically extract and leverage domain knowledge from external sources such as game manuals. We apply our algorithm to the game of Civilization II, a challenging multiagent strategy game with an enormous state space and around 1021 joint actions. We approximate the value function by a neural network, augmented by linguistic knowledge that is extracted automatically from the official game manual. We show that this non-linear value function is significantly more efficient than a linear value function, which is itself more efficient than Monte-Carlo tree search. Our non-linear Monte-Carlo search wins over 78% of games against the built-in AI of Civilization II.