Combining Textual and Speech Features in the NLI Task Using State-of-the-Art Machine Learning Techniques

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F17%3A43932939" target="_blank" >RIV/49777513:23520/17:43932939 - isvavai.cz</a>

Výsledek na webu

<a href="http://www.aclweb.org/anthology/W/W17/W17-5021.pdf" target="_blank" >http://www.aclweb.org/anthology/W/W17/W17-5021.pdf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.18653/v1/W17-5021" target="_blank" >10.18653/v1/W17-5021</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Combining Textual and Speech Features in the NLI Task Using State-of-the-Art Machine Learning Techniques

Popis výsledku v původním jazyce

We summarize the involvement of our CEMI team in the Native Language Identification shared task, NLI Shared Task~2017, which deals with both textual and speech input data. We submitted the results achieved by using three different system architectures; each of them combines multiple supervised learning models trained on various feature sets. As expected, better results are achieved with the systems that use both the textual data and the spoken responses. Combining the input data of two different modalities led to a rather dramatic improvement in classification performance. Our best performing method is based on a set of feed-forward neural networks whose hidden-layer outputs are combined together using a softmax layer. We achieved a macro-averaged F1 score of 0.9257 on the evaluation (unseen) test set and our team placed first in the main task together with other three teams.

Název v anglickém jazyce

Combining Textual and Speech Features in the NLI Task Using State-of-the-Art Machine Learning Techniques

Popis výsledku anglicky

We summarize the involvement of our CEMI team in the Native Language Identification shared task, NLI Shared Task~2017, which deals with both textual and speech input data. We submitted the results achieved by using three different system architectures; each of them combines multiple supervised learning models trained on various feature sets. As expected, better results are achieved with the systems that use both the textual data and the spoken responses. Combining the input data of two different modalities led to a rather dramatic improvement in classification performance. Our best performing method is based on a set of feed-forward neural networks whose hidden-layer outputs are combined together using a softmax layer. We achieved a macro-averaged F1 score of 0.9257 on the evaluation (unseen) test set and our team placed first in the main task together with other three teams.

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

20205 - Automation and control systems

Projekt

<a href="/cs/project/GBP103%2F12%2FG084" target="_blank" >GBP103/12/G084: Centrum pro multi-modální interpretaci dat velkého rozsahu</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2017

Kód důvěrnosti údajů

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