Go to content
UR Home

Classifying user information needs in cooking dialogues – an empirical performance evaluation of transformer networks

URN to cite this document:
urn:nbn:de:bvb:355-epub-463085
DOI to cite this document:
10.5283/epub.46308
Schwabl, Patrick
[img]
Preview
License: Creative Commons Attribution Non-commercial 4.0
PDF
Master Thesis Text
(7MB)
[img]License: Creative Commons Attribution Non-commercial 4.0
ZIP Archive - Supplemental Material
Code, Data and other supplemental material
(16MB)
Date of publication of this fulltext: 14 Jul 2021 06:25


Abstract (German)

In this master’s thesis, I carry out 3720 machine learning experiments. I want to test how transformer networks perform in a dialogue processing task. Transformer networks are deep neural networks that have first been proposed in 2017 and have since rapidly set new state of the art results on many tasks. To evaluate their performance in dialogue classification, I use two tasks from two datasets. ...

plus

Translation of the abstract (German)

In dieser Masterarbeit führe ich 3720 Experimente zum maschinellen Lernen durch. Ich möchte testen, wie Transformer-Netze bei einer Dialogverarbeitungsaufgabe abschneiden. Transformator-Netzwerke sind tiefe neuronale Netze, die erstmals 2017 vorgeschlagen wurden und seitdem bei vielen Aufgaben schnell neue State-of-the-Art-Ergebnisse erzielt haben. Um ihre Leistung bei der Dialogklassifizierung ...

plus


Owner only: item control page
  1. Homepage UR

University Library

Publication Server

Contact:

Publishing: oa@ur.de
0941 943 -4239 or -69394

Dissertations: dissertationen@ur.de
0941 943 -3904

Research data: daten@ur.de
0941 943 -5707

Contact persons