Personality Recognition from Mouse and Keyboard Dynamics under Label Scarcity
Description: Personality is stable, which means no matter how much data we collect from a person, he or she always has only one corresponding score for a personality trait that will be the ground-truth label in recognition. This may lead to the harsh requirement for the dataset to cover a wide range of personality scores, increasing the difficulty of data collection and personality research.
Representation learning has been shown to be an effective approach to capturing the semantic pattern of the input that can generalise to different scenarios, datasets and downstream tasks. It has been widely investigated in NLP and CV. For example, word2vec uses a vector to represent each word.
This project will predict personality from a user's mouse and keyboard dynamics in a small dataset, using the representations learnt from larger mouse and keyboard datasets (without personality labels) in an unsupervised way.
Supervisor: Guanhua Zhang
Distribution: 20% Literature, 20% Data Preparation, 40% Implementation, 20% Analysis and Evaluation
Requirements: Strong programming skills, practical experience in data processing and deep learning
Literature: Juan Abdon Miranda-Correa, Motjaba Khomami Abadi, Nicu Sebe, and Ioannis Patras. 2018. AMIGOS: A dataset for affect, personality and mood research on individuals and groups. IEEE Transactions on Affective Computing, 12(2), p.479-493.
Ashish Vaswani et al. 2017. Attention is all you need. Advances in neural information processing systems 30.
Lei Han et al. 2020. Modelling user behavior dynamics with embeddings. Proceedings of the 29th ACM International Conference on Information and Knowledge Management.
António Góis and André FT Martins. 2019. Translator2vec: Understanding and representing human post-editors. arXiv:1907.10362.