Discriminating Cognitive Disequilibrium and Flow in Problem Solving: A Semi-Supervised Approach Using Involuntary Dynamic Behavioral Signals

Abstract

Problem solving is one of the most important 21st century skills. However, effectively coaching young students in problem solving is challenging because teachers must continuously monitor their cognitive and affective states, and make real-time pedagogical interventions to maximize their learning outcomes. It is an even more challenging task in social environments with limited human coaching resources. To lessen the cognitive load on a teacher and enable affect-sensitive intelligent tutoring, many researchers have investigated automated cognitive and affective detection methods. However, most of the studies use culturally-sensitive indices of affect that are prone to social editing such as facial expressions, and only few studies have explored involuntary dynamic behavioral signals such as gross body movements. In addition, most current methods rely on expensive labelled data from trained annotators for supervised learning. In this paper, we explore a semi-supervised learning framework that can learn low-dimensional representations of involuntary dynamic behavioral signals (mainly gross-body movements) from a modest number of short time series segments. Experiments on a real-world dataset reveal a significant advantage of these representations in discriminating cognitive disequilibrium and flow, as compared to traditional complexity measures from dynamical systems literature, and demonstrate their potential in transferring learned models to previously unseen subjects.

Weakly Supervised Classification of Vital Sign Alerts as Real or Artifact

“…Samples not covered by any LF were filtered out, and the remaining probabilistic labels produced by the label model were translated into crisp binary training labels, which were then used to train a Random Forest (RF) model 33 to classify VS alerts as real versus artifact. RFs have been widely used in literature to learn complex decision boundaries for various classification problems 34,35 and have also been shown to be effective for learning discriminative models of real versus artifact VS alert classification 2 . We trained RF models with 1000 decision trees having a maximum depth of 5 implemented using scikit-learn 36 .…”

Exploring Semi-Supervised Learning for Predicting Listener Backchannels

“…Several AI techniques based on Semi-Supervision [14,21] and Weak-Supervision [11,26] have been utilized to decrease the annotation costs in different application domains. Here we explore self-training based semi-supervised learning paradigm [37] for identifying listener backchannel instances and the signals associated with them, using only a small subset of the manually annotated data.…”