Synopsis

Anthropomorphism is the attribution of human traits, emotions, or intentions to non-human entities. The traditional IVR agents are intelligent bots and have limitations in human-like conversation style, tone, and affection. The key objective of this research is to develop an affective anthropomorphic intelligent agent (AAIA) using generative deep learning and transformer-based NLP. The agent learns an invertible mapping of data to a latent space that can be manipulated and generates a Mel-spectrogram frame for voice synthesis and style transfer. The contextual affect of the conversion will be maintained through affective computing. You can train the AIA to be a digital clone of your loving idol or companion.

Relevance of the Topic

The topic is relevant to computer science, computer engineering, and data science students who are enthusiasts in generative deep learning, natural language processing, and audio signal processing.

Future Research/Scope

You can train the AAIA to be a digital clone of your loving idol or companion.

Skills Learned

After completion of this research, students will learn affective computing methods for sentences and human voices and will learn the utilization of generative deep learning methods in IVR design.

Relevant courses to the topic

•  Artificial Intelligence, Neural Networks, Data Science, Machine Learning, HCI, and Natural Language Processing

Reading List

• J. Li, M. Galley, C. Brockett, G. P. Spithourakis, J. Gao, and B. Dolan, “A persona-based neural conversation model,” arXiv:1603.06155 [cs], 6 2016. arXiv: 1603.06155.
• B. Schuller and A. Batliner, Computational paralinguistics: emotion, affect and personality in speech and language processing. Hoboken, N.J: Wiley, first edition ed., 2014.
• M. El Ayadi, M. S. Kamel, and F. Karray, “Survey on speech emotion recognition: Features, classification schemes, and databases,” Pattern Recognition, vol. 44, pp. 572–587, 3 2011.
• Y. Lee, A. Rabiee, and S.-Y. Lee, “Emotional end-to-end neural speech synthesizer,” arXiv:1711.05447 [cs, eess], 11 2017. arXiv: 1711.05447.
• Y. Gao, R. Singh, and B. Raj, “Voice impersonation using generative adversarial networks,” arXiv:1802.06840 [cs, eess], 2 2018. arXiv: 1802.06840.