WAI › Presentations › Natalie van der Wal (11 June 2012)

Description

Title Humans versus Machine: Are humans better or worse in detecting emotions than a machine learning algorithm?
Abstract In this presentation I will be showing results from the experiment I have conducted (in the Intertainlab) in which speech was recorded from 80 participants while induced with a neutral/positive/negative emotion. A machine learning algorithm (Random Forest model) has been built to detect the changes in (16 different) emotions. Results are promising!! Moreover, a validation experiment has been performed with 15 participants who have done the same task as the machine learning algorithm. In this way I can evaluate who is better in detecting the changes in (16 different) emotions: humans or the machine?????? And the winner is..... the machine learning algorithm! Come and enjoy my talk this Monday to find out the details :-)

Other presentations by Natalie van der Wal

DateTitle
05 January 2009 Modelling Emotion Contagion: Multi-Agent Models for Mutual Absorption of Emotions and Emotion Contagion Spirals
11 October 2010
24 October 2011 A Telephone that can measure your mood.. (...results of study in which most of you participated, ALL COME!!)
11 June 2012 Humans versus Machine: Are humans better or worse in detecting emotions than a machine learning algorithm?
25 March 2013 Reward Priming: testing the effect of reward on visual search & possibilities for more applied research.
30 March 2015 THE SCIENCE OF LAUGHTER, CROWDS AND TRAFFIC ACCIDENTS!!