Hand gestures are widely used in daily life for expressing emotions, yet gesture input is not part of existing emotion tracking systems. To seek a practical and effortless way of using gestures to inform emotions, we explore the relationships between gestural features and commonly experienced emotions by focusing on single-hand gestures that are easy to perform and capture. First, we collected 756 gestures (in photo and video pairs) from 63 participants who expressed different emotions in a survey, and then interviewed 11 of them to understand their gesture-forming rationales. We found that the valence and arousal level of the expressed emotions significantly correlated with participants’ finger-pointing direction and their gesture strength, and synthesized four channels through which participants externalized their expressions with gestures. Reflecting on the findings, we discuss how emotions can be characterized and contextualized with gestural cues and implications for designing multimodal emotion tracking systems and beyond.

Artificial intelligence (AI) requires a large amount of data to train high-quality machine learning (ML) models. However, due to privacy issues, individuals or organizations are not willing to share data with others, which results in “data islands”. This motivates the emergence of Federated Learning (FL), a novel ML framework allowing clients to exchange model parameters rather than the raw data. Unfortunately, the private data may be reconstructed by malicious participants by exploiting the context of model parameters in FL. This poses further challenges to privacy protection. To address this issue, we propose to integrate Wasserstein Generative Adversarial Network (WGAN) and differential privacy (DP) to protect the model parameters. WGAN is used to generate controllable random noise, which is then injected into model parameters. The new mechanism satisfies DP requirements while the data utility is highly improved. We experimentally demonstrate superior performances from aspects of convergence, accuracy, and data utility.