Voice-Based Control with Siamese Networks

  • Gautham Venkataraman, Siddharth Balaji Mahalingam and Subalalitha C.N

Abstract

In this digital-technical era, the growth of modern technology in our day to day lives makes everything considered fictional such as giving commands to a machine via vocal communication. However, with rapid technological advancements in the field of Natural Language Processing (NLP) and Artificial Intelligence (AI), in order to make things faster, easier and efficient these days we can find voice user interfaces (VUI) in electronic devices like smartphones, machines, etc. The creation of VUI is cumbersome, being plagued by accuracy problems and information processing. In this paper, we are proposing a design - ‘Voice-Based Control with Siamese Networks’ consisting of two identical neural networks made to measure and to rank the semantic similarity between audio inputs. Initial testing on simple video games reveals the performance of Siamese neural networks where we compare it to the existing traditional approaches using VUI as well as other models with strong results. The growth of applications on VUI in our day-to-day lives will lead to ease of access.

Published
2020-06-02
How to Cite
Gautham Venkataraman, Siddharth Balaji Mahalingam and Subalalitha C.N. (2020). Voice-Based Control with Siamese Networks. International Journal of Advanced Science and Technology, 29(9s), 6527 - 6533. Retrieved from https://sersc.org/journals/index.php/IJAST/article/view/20351