This project contains experimental code for classying opinion and persuasiveness from speech using vanilla long short-term memory (LSTMs) recurrent neural nets implementation from Keras.
Please use the following citation:
@inproceedings{santos2016,
author = {Pedro Bispo Santos and Lisa Beinborn and Iryna Gurevych},
title = {A domain-agnostic approach for opinion prediction on speech},
year = 2016,
booktitle = {Proceedings of the Workshop on Computational Modeling of People's Opinions, Personality, and Emotions in Social Media held in conjunction with COLING 2016},
pages = {163-172},
location = {Osaka,Japan}
}
Abstract: We explore a domain-agnostic approach for analyzing speech with the goal of opinion prediction. We represent the speech signal by mel-frequency cepstral coefficients and apply long short-term memory neural networks to automatically learn temporal regularities in speech. In contrast to previous work, our approach does not require complex feature engineering and works without textual transcripts. As a consequence, it can easily be applied on various speech analysis tasks for different languages and the results show that it can nevertheless be competitive to the state- of-the-art in opinion prediction. In a detailed error analysis for opinion mining we find that our approach performs well in identifying speaker-specific characteristics, but should be combined with additional information if subtle differences in the linguistic content need to be identified.
Contact person: Pedro Santos, https://www.ukp.tu-darmstadt.de/people/doctoral-researchers/pedro-santos/
https://www.ukp.tu-darmstadt.de/
intelligence_squared-- Persuasiveness prediction experiments on the Intelligence Squared dataset (http://www.intelligencesquaredus.org/).moud_dataset-- Opinion mining experiments on the Multimodal Opinion Mining Dataset (https://web.eecs.umich.edu/~mihalcea/downloads.html#MOUD)