Affect Estimation in 3D Space Using Multi-Task Active Learning for Regression

Abstract

Acquisition of labeled training samples for affective computing is usually costly and time-consuming, as affects are intrinsically subjective, subtle and uncertain, and hence multiple human assessors are needed to evaluate each affective sample. Particularly, for affect estimation in the 3D space of valence, arousal and dominance, each assessor has to perform the evaluations in three dimensions, which makes the labeling problem even more challenging. Many sophisticated machine learning approaches have been proposed to reduce the data labeling requirement in various other domains, but so far few have considered affective computing. This paper proposes two multi-task active learning for regression approaches, which select the most beneficial samples to label, by considering the three affect primitives simultaneously. Experimental results on the VAM corpus demonstrated that our optimal sample selection approaches can result in better estimation performance than random selection and several traditional single-task active learning approaches. Thus, they can help alleviate the data labeling problem in affective computing, i.e., better estimation performance can be obtained from fewer labeling queries.

Patch Learning

“…So, the next step is to build a patch model for x ∈ [1. 5,3]. Using only the training examples within this patch, we obtain the first patch model f 1 (x) = 1.65 + 9.81x − 2.01x 2 .…”

“…3(d), we can see that the fitting errors are large when x ∈ [4,5]. So, the next step is to build a patch model for x ∈ [4,5]. Using only the training examples within this patch, we obtain the second patch model f 2 (x) = 19.29 − 8.03x + 1.96x 2 .…”

“…Machine learning has been widely used in our everyday life, e.g., face recognition [1], natural language processing [2], recommender systems [3], affective computing [4], [5], braincomputer interfaces [6], [7], etc. However, training a highperformance machine learning model is usually a challenging and iterative process, relying on experience and trial-and-error: a simple model is first designed; if its performance is not satisfactory, then some remedies are taken to enhance it.…”

“…We could develop such an ANFIS algorithm, but it is beyond the scope of this paper. Additionally, this demonstrates one advantage of PL: it does not add weights to the training examples, so any existing machine learning algorithms can be readily used as the patch model 5. The original SSE in [0, 2.11] was 1230.8, and there were 212 points; hence, the original RMSE was 1230.8/212 = 2.41.…”

Active Learning for Black-Box Adversarial Attacks in EEG-Based Brain-Computer Interfaces

“…There are two main scenarios of active learning in the literature: query synthesis [29]- [32] and sampling. The latter can be further divided into stream-based sampling [33], [34] and pool-based sampling [35]- [38]. Sampling-based active learning selects real unlabeled instances from a pool or steam for labeling.…”