Using reinforcement learning to learn how to play text-based games (2018-01-06T00:00:00.000000Z)

TL;DR

This work presents a general text game playing agent, testing its generalisation and transfer learning performance and showing its ability to play multiple games at once, and presents pyfiction, an open-source library for universal access to different text games that could serve as a baseline for future research.

Abstract

The ability to learn optimal control policies in systems where action space is defined by sentences in natural language would allow many interesting real-world applications such as automatic optimisation of dialogue systems. Text-based games with multiple endings and rewards are a promising platform for this task, since their feedback allows us to employ reinforcement learning techniques to jointly learn text representations and control policies. We present a general text game playing agent, testing its generalisation and transfer learning performance and showing its ability to play multiple games at once. We also present pyfiction, an open-source library for universal access to different text games that could, together with our agent that implements its interface, serve as a baseline for future research.

Authors

Mikuláš Zelinka

2 papers

TL;DR

Abstract

Authors

References43 items

Hybrid Code Networks: practical and efficient end-to-end dialog control with supervised and reinforcement learning

Embracing data abundance: BookTest Dataset for Reading Comprehension

Deep Reinforcement Learning with an Unbounded Action Space

Language Understanding for Text-based Games using Deep Reinforcement Learning

Human-level control through deep reinforcement learning

Deep Sentence Embedding Using Long Short-Term Memory Networks: Analysis and Application to Information Retrieval

Adam: A Method for Stochastic Optimization

Playing Atari with Deep Reinforcement Learning

Efficient Estimation of Word Representations in Vector Space

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

Intrinsically Motivated Reinforcement Learning: An Evolutionary Perspective

Intrinsically Motivated Reinforcement Learning

Long Short-Term Memory

Temporal Difference Learning and TD-Gammon

Learning long-term dependencies with gradient descent is difficult

Signature Verification Using A "Siamese" Time Delay Neural Network

Approximation capabilities of multilayer feedforward networks

Backpropagation Through Time: What It Does and How to Do It

The perceptron: a probabilistic model for information storage and organization in the brain.

Deep Learning

Keras

TensorFlow: Large-scale machine learning on heterogeneous systems

Prioritized Experience Replay. CoRR

Understanding LSTM networks

Dynamic programming. Courier Corporation

LSTM Neural Networks for Language Modeling

Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude. COURSERA: Neural networks for machine learning

Similarity Measures for Text Document Clustering

Technical Note: Q-Learning

Twisty Little Passages: An Approach to Interactive Fiction

Book Reviews: Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition

Reinforcement Learning -An Introduction. Adaptive computation and machine learning

Interaction between the agent and the environment (Sutton and Barto

Untersuchungen zu dynamischen neuronalen netzen. Diploma

Cited on page

State of the Art—A Survey of Partially Observable Markov Decision Processes: Theory, Models, and Algorithms

Beyond Regression : "New Tools for Prediction and Analysis in the Behavioral Sciences

Generalisation on paraphrased actions in Machine of Death

Harm de Vries, Junyoung Chung, and Yoshua Bengio. RM-SProp and equilibrated adaptive learning rates for non-convex optimization

Game properties and agent performance summary

28 4.2.1 Performance of the SSAQN agent in individual games

64 List of Tables 4.1 Summary of game statistics

A simple RNN (Olah [2015])

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names