Search Results for author:
Found 11 papers, 0 papers with code
Addressing Ambiguity of Emotion Labels Through Meta-Learning
We therefore propose a dynamic label correction and sample contribution weight estimation model.
Towards Universal End-to-End Affect Recognition from Multilingual Speech by ConvNets
It outperforms a similar CNN using spectrograms as input by 12. 8% for emotion and 6. 3% for personality, based on F-scores.
GlobalTrait: Personality Alignment of Multilingual Word Embeddings
We propose a multilingual model to recognize Big Five Personality traits from text data in four different languages: English, Spanish, Dutch and Italian.
Investigating Audio, Video, and Text Fusion Methods for End-to-End Automatic Personality Prediction
Furthermore, we can see the prediction relevance of each modality for each trait.
Investigating Audio, Visual, and Text Fusion Methods for End-to-End Automatic Personality Prediction
Furthermore, we can see the prediction relevance of each modality for each trait.
Zara: A Virtual Interactive Dialogue System Incorporating Emotion, Sentiment and Personality Recognition
Zara, or {`}Zara the Supergirl{'} is a virtual robot, that can exhibit empathy while interacting with an user, with the aid of its built in facial and emotion recognition, sentiment analysis, and speech module.
Towards Empathetic Human-Robot Interactions
Although research on empathetic robots is still in the early stage, we described our approach using signal processing techniques, sentiment analysis and machine learning algorithms to make robots that can "understand" human emotion.
Deep Learning of Audio and Language Features for Humor Prediction
Our work is a starting point to developing more effective machine learning and neural network models on the humor prediction task, as well as developing machines capable in understanding humor in general.
