Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks

Abstract: 

Natural language processing research has made major advances with the concept of representing words, sentences, paragraphs, and even documents by embedded vector representations. We apply this idea to the problem of relating foods, as expressed in natural language meal descriptions, to corresponding database entries. We generate fixed-length embeddings for U.S. Department of Agriculture (USDA) food database entries, as well as vector-based representations of natural language meal descriptions, through a convolutional neural network (CNN) architecture that predicts whether or not a USDA food item is present in the meal description. We compute dot products between each token in a meal description and a USDA food entry. By ranking the network's predicted average dot product between each possible database food entry and a meal description, we show it is possible to directly predict the USDA foods mentioned in a meal without requiring intermediate steps that would be used in a conventional database access application. We report the performance of this model on a binary verification task of over 48k meal descriptions, and show that this approach, when integrated with a Markov model, substantially outperforms our previous best multi-stage approach involving a conditional random field tagger, probabilistic segmentation, and database lookup.

icassp_17.pdf

PDF icon icassp_17.pdf (128 downloads)
up
0 users have voted:

Paper Details

Authors:
Zachary Collins, Jim Glass
Submitted On:
3 March 2017 - 12:34pm
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Mandy Korpusik
Paper Code:
2627

Document Files

icassp_17.pdf

(128 downloads)

Subscribe

[1] Zachary Collins, Jim Glass, "Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/1616. Accessed: Jan. 20, 2018.
@article{1616-17,
url = {http://sigport.org/1616},
author = {Zachary Collins; Jim Glass },
publisher = {IEEE SigPort},
title = {Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks},
year = {2017} }
TY - EJOUR
T1 - Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks
AU - Zachary Collins; Jim Glass
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/1616
ER -
Zachary Collins, Jim Glass. (2017). Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks. IEEE SigPort. http://sigport.org/1616
Zachary Collins, Jim Glass, 2017. Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks. Available at: http://sigport.org/1616.
Zachary Collins, Jim Glass. (2017). "Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks." Web.
1. Zachary Collins, Jim Glass. Semantic Mapping of Natural Language Input to Database Entries via Convolutional Neural Networks [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/1616