Transfer learning for question answering on SQuAD

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F18%3A00326870" target="_blank" >RIV/68407700:21230/18:00326870 - isvavai.cz</a>

Alternative codes found

RIV/68407700:21730/18:00326870

Result on the web

—

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

Transfer learning for question answering on SQuAD

Original language description

This paper focuses on the benefits of transfer learning on question answering tasks. We show how transfer learning improves the results of state of the art question answering system on evaluation set focused on a specific domain not contained in the training data. We perform several experiments on the Stanford Question Answering Dataset (SQuAD). We use transfer learning approach to retrain model trained with a large amount of generic data on smaller topic specific datasets. We evaluate the change in performance of the system with regards to the size of the new training set and type of data used in training.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

S - Specificky vyzkum na vysokych skolach

Publication year

2018

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Proceedings of the International Student Scientific Conference Poster – 22/2018

ISBN

978-80-01-06428-3

ISSN

—

e-ISSN

—

Number of pages

5

Pages from-to

—

Publisher name

Czech Technical University in Prague

Place of publication

Praha

Event location

Praha

Event date

May 10, 2018

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—