224n Project: Natural Language Learning Supports Reinforcement Learning
نویسنده
چکیده
Deep learning systems have recently achieved near human level performance on game playing. However, they require a much larger amount of data to achieve this than humans do, and often (though not always) still perform worse than humans. One factor that may account for this is the richer feedback humans receive, in particular we often receive a great deal of natural language instruction when learning tasks. In this paper, we explore providing auxiliary natural language instruction to a network in the simple game context of tic-tac-toe, and show that it accelerates learning.
منابع مشابه
CS 224N/229: Joint Final Project: Large-Vocabulary Continuous Speech Recognition with Linguistic Features for Deep Learning
Until this day, automated speech recognition (ASR) still remains one of the most challenging tasks in both machine learning and natural language processing. ASR research faces data with high variability, which requires highly expressive models be built. Recently, deep neural networks (DNN) have been successfully applied to various fields, including speech recognition. In this course project, We...
متن کاملDialogue manager domain adaptation using Gaussian process reinforcement learning
Spoken dialogue systems allow humans to interact with machines using natural speech. As such, they have many benefits. By using speech as the primary communication medium, a computer interface can facilitate swift, human-like acquisition of information. In recent years, speech interfaces have become ever more popular, as is evident from the rise of personal assistants such as Siri, Google Now, ...
متن کاملThe Origin of the Speeches: Language Evolution through Collaborative Reinforcement Learning
This project proposes that language evolve through reinforcement learning where agents communicate with each other and provide rewards if communication is successful. The fundamental difference between the learning mechanisms that humans use to communicate with one another and how machines learn to communicate is that the system used by humans presupposes that the adult already knows the meanin...
متن کاملCS229 Final Project: Language Grounding in Minecraft with Gated-Attention Networks
A key question in language understanding is the problem of language grounding – how do symbols such as words get their meaning? We examine this question in the context of task oriented language grounding in gameplay. In order to perform tasks and challenges specified by natural language instructions, agents need to extract semantically meaningful representations of language and map it to the vi...
متن کاملChallenges in Policy Learning for British Sign Language Users in an Example of Navigation-Based Task
In the light of natural human interactions, this paper identies the lack of multimodality in current policy learning approaches for human-robot/human-computer interactions. Larger part of policies, trained for the navigation-based tasks are trained on using either textual or natural language inputs, which are unimodal. is paper suggests and formalises British Sign Language as an input to the ...
متن کامل