Computer Science > Computation and Language

Title:Learning to Compose Neural Networks for Question Answering

Authors:Jacob Andreas, Marcus Rohrbach, Trevor Darrell, Dan Klein

(Submitted on 7 Jan 2016 (v1), last revised 7 Jun 2016 (this version, v4))

Abstract: We describe a question answering model that applies to both images and structured knowledge bases. The model uses natural language strings to automatically assemble neural networks from a collection of composable modules. Parameters for these modules are learned jointly with network-assembly parameters via reinforcement learning, with only (world, question, answer) triples as supervision. Our approach, which we term a dynamic neural model network, achieves state-of-the-art results on benchmark datasets in both visual and structured domains.

Subjects:

Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Neural and Evolutionary Computing (cs.NE)

Cite as:

arXiv:1601.01705 [cs.CL]

(or arXiv:1601.01705v4 [cs.CL] for this version)

Try the Bibliographic Explorer
(can be disabled at any time)

Enable Don't show again

Bibliographic data

[Enable Bibex(What is Bibex?)]

Submission history

From: Jacob Andreas [view email]
[v1] Thu, 7 Jan 2016 21:21:59 UTC (4,810 KB)
[v2] Wed, 1 Jun 2016 18:20:37 UTC (5,947 KB)
[v3] Mon, 6 Jun 2016 01:44:25 UTC (7,152 KB)
[v4] Tue, 7 Jun 2016 23:25:51 UTC (7,293 KB)

Which authors of this paper are endorsers? | Disable MathJax) (What is MathJax?)

Original url: Access
Created at: 2018-11-02 11:54:29
Category: default
Tags: none

未标明原创文章均为采集，版权归作者所有，转载无需和我联系，请注明原出处，南摩阿彌陀佛，知识，不只知道，要得到