Affective Computing

Affective computing studies how machines can infer and work with emotion, mood, engagement, and related states from speech, facial and bodily behavior, text, and physiological signals. Common approaches include supervised classifiers, multimodal fusion, temporal models that track changing affect, and appraisal-based or language-model methods for interpreting context. Applications span healthcare and wearables, education, human-robot interaction, and emotionally responsive digital tools, with evaluation increasingly extending from controlled datasets to naturalistic conversations and daily-life settings. A central challenge is that emotional cues are ambiguous and vary across people and cultures, so accuracy must be balanced with uncertainty, fairness, privacy, and respect for the limits of what a system can infer about an individual.