HomeHome

Searching for Meaning in RNNs using Deep Neural Inspection

Kevin Lin; Eugene Wu

Title:
Searching for Meaning in RNNs using Deep Neural Inspection
Author(s):
Lin, Kevin
Wu, Eugene
Date:
Type:
Reports
Department(s):
Computer Science
Persistent URL:
Series:
Columbia University Computer Science Technical Reports
Part Number:
003-17
Publisher:
Department of Computer Science, Columbia University
Publisher Location:
New York
Abstract:
Recent variants of Recurrent Neural Networks (RNNs)---in particular, Long Short-Term Memory (LSTM) networks---have established RNNs as a deep learning staple in modeling sequential data in a variety of machine learning tasks. However, RNNs are still often used as a black box with limited understanding of the hidden representation that they learn. Existing approaches such as visualization are limited by the manual effort to examine the visualizations and require considerable expertise, while neural attention models change, rather than interpret, the model. We propose a technique to search for neurons based on existing interpretable models, features, or programs.
Subject(s):
Neural networks (Computer science)
Computer science
Neurons
Item views
85
Metadata:
text | xml
Suggested Citation:
Kevin Lin, Eugene Wu, , Searching for Meaning in RNNs using Deep Neural Inspection, Columbia University Academic Commons, .

Columbia University Libraries | Policies | FAQ