Jonny O’Dwyer

Jonny O'Dwyer, Niall Murray, Ronan Flynn

Speech has been a widely used modality in the field of affective computing. Recently however, there has been a growing interest in the use of multi-modal affective computing systems. These multi-modal systems incorporate both verbal and non-verbal features for affective computing tasks. Such multi-modal affective computing systems are advantageous for emotion assessment of individuals in audio-video communication environments such as teleconferencing, healthcare, and education. From a review of the literature, the use of eye gaze features extracted from video is a modality that has remained largely unexploited for continuous affect prediction. This work presents a review of the literature within the emotion classification and continuous affect prediction sub-fields of affective computing for both speech and eye gaze modalities. Additionally, continuous affect prediction experiments using speech and eye gaze modalities are presented. A baseline system is proposed using open source software, the performance of which is assessed on a publicly available audio-visual corpus. Further system performance is assessed in a cross-corpus and cross-lingual experiment. The experimental results suggest that eye gaze is an effective supportive modality for speech when used in a bimodal continuous affect prediction system. The addition of eye gaze to speech in a simple feature fusion framework yields a prediction improvement of 6.13% for valence and 1.62% for arousal.

Jonny O'Dwyer, Malika Bendechache, Louise McCormack et al.

Present implementations of artificial intelligence (AI) ethics do not adequately take feelings, or affect, into account. If AI should be aligned with human ethics, it seems reasonable to thoroughly investigate the possibility of AI behaviour that mirrors virtuous human ethical conduct, where feelings play a role in the actions, judgements or statements one makes. Furthermore, while prominent theories of normative ethics are often discussed in terms of their differences and shortcomings, Virtue, Consequentialist, and Kantian Deontological ethics all share a common feature of considering human feeling to some degree while the popular descriptive ethics theory, Moral Foundations Theory, positions feelings as central to many of its foundations. Therefore, in the present paper, a data set of moral valence is proposed, consisting of 500 annotations by six human participants for both action/judgement and consequence moral valence, ranging from -1 to 1 for text-presented scenarios from the Commonsense Norm Bank data set. The resulting valence features share significant relationships with multi-class (immoral/discretionary/moral) and binary immoral/moral categories while additionally providing a noteworthy test set Matthew's correlation coefficient of 0.764 using regularised logistic regression for binary classification. This provides early evidence of the usefulness of valence features for morality estimation of text, indicating that valenced consequences of responses for others can be considered toward more human morally-aligned AI. In the interest of promoting further affective-moral computing research, this study's annotations will be made available for research on request.

Jonny O'Dwyer

Continuous affect prediction involves the discrete time-continuous regression of affect dimensions. Dimensions to be predicted often include arousal and valence. Continuous affect prediction researchers are now embracing multimodal model input. This provides motivation for researchers to investigate previously unexplored affective cues. Speech-based cues have traditionally received the most attention for affect prediction, however, non-verbal inputs have significant potential to increase the performance of affective computing systems and in addition, allow affect modelling in the absence of speech. However, non-verbal inputs that have received little attention for continuous affect prediction include eye and head-based cues. The eyes are involved in emotion displays and perception while head-based cues have been shown to contribute to emotion conveyance and perception. Additionally, these cues can be estimated non-invasively from video, using modern computer vision tools. This work exploits this gap by comprehensively investigating head and eye-based features and their combination with speech for continuous affect prediction. Hand-crafted, automatically generated and CNN-learned features from these modalities will be investigated for continuous affect prediction. The highest performing feature sets and feature set combinations will answer how effective these features are for the prediction of an individual's affective state.

Jonny O'Dwyer, Niall Murray, Ronan Flynn

Eye-based information channels include the pupils, gaze, saccades, fixational movements, and numerous forms of eye opening and closure. Pupil size variation indicates cognitive load and emotion, while a person's gaze direction is said to be congruent with the motivation to approach or avoid stimuli. The eyelids are involved in facial expressions that can encode basic emotions. Additionally, eye-based cues can have implications for human annotators of emotions or feelings. Despite these facts, the use of eye-based cues in affective computing is in its infancy, however, and this work is intended to start to address this. Eye-based feature sets, incorporating data from all of the aforementioned information channels, that can be estimated from video are proposed. Feature set refinement is provided by way of continuous arousal and valence learning and prediction experiments on the RECOLA validation set. The eye-based features are then combined with a speech feature set to provide confirmation of their usefulness and assess affect prediction performance compared with group-of-humans-level performance on the RECOLA test set. The core contribution of this paper, a refined eye-based feature set, is shown to provide benefits for affect prediction. It is hoped that this work stimulates further research into eye-based affective computing.

Jonny O'Dwyer, Ronan Flynn, Niall Murray

In recent times, there has been significant interest in the machine recognition of human emotions, due to the suite of applications to which this knowledge can be applied. A number of different modalities, such as speech or facial expression, individually and with eye gaze, have been investigated by the affective computing research community to either classify the emotion (e.g. sad, happy, angry) or predict the continuous values of affective dimensions (e.g. valence, arousal, dominance) at each moment in time. Surprisingly after an extensive literature review, eye gaze as a unimodal input to a continuous affect prediction system has not been considered. In this context, this paper evaluates the use of eye gaze as a unimodal input to a continuous affect prediction system. The performance of continuous prediction of arousal and valence using eye gaze is compared with the performance of a speech system using the AVEC 2014 speech feature set. The experimental evaluation when using eye gaze as the single modality in a continuous affect prediction system produced a correlation result for valence prediction that is better than the correlation result obtained with the AVEC 2014 speech feature set. Furthermore, the eye gaze feature set proposed in this paper contains 98% fewer features compared to the number of features in the AVEC 2014 feature set.