Saritha Kinkiri
Associate Lecturer
About
I have been delivering lectures, tutorials and lab demonstrations at the Faculty of Engineering. In addition, I have prepared reading the material for the lectures, and experimental procedures labs. I have also graded reports and course work for students.
I have been teaching in several settings ranging from one-to-one basis to large lectures face to face and online. I am also supervising final year projects and an MSc student from the School of Engineering and Sciences. I am working on several research projects. Most of the projects are on human computer interaction technologies and Machine Learning algorithms.
I contribute to several outreach activities, such as applicant days and clearing events. Also, I have helped in the mathematics masterclasses, which encourage school kids to study STEM-related topics
I work on the following undergraduate modules: Electrical Circuits, Digital Electronic Systems, Analogue Electronic Systems, Digital Electronic System, Electronic System Development, Engineering Math I, Practical and Experimental Skills Advanced Analogue Electronics and VHDL.
I am involved in these postgraduate modules: Design of Advanced Electronics Systems, Machines Intelligence, Embedded Electronics and Communications, Mathematics and Statistical, and Drivers Labs.
I work on these lab demonstrations: Introduction to programming with MATLAB, DC circuits, Filter circuits, Data Communications, Introduction to Proteus Software and Digital simulation and Computer Components and Analysis.
Research
- Human computer Interaction
- Speech/voice Recognition
- NLP
PhD: Current or potential PhD supervision interest, PhD summary (bullet format or 200 words max)
The underlying approach of the research is to observe and analyse language-independent speaker identification based on the fundamental characteristics of human voices. The aim is to explore speech parameters in both controlled and uncontrolled tasks, such as free speech or reading a script respectively. For example, this research will explore how many people out of a sample population of 100 participants can be excluded through using a combination of simple voice characteristics, such as dominant frequencies and pauses. The collection of data shows that a number of principal voice characteristics are independent of the language being spoken. The training and testing of the recogniser play a major role in identifying a speaker, but in this thesis, the work will concentrate on the elimination of a speaker from a pool of potential candidates using the fastest possible means for the least amount of data training. There is other research available on recognising individuals from lots of data training, but
This research is focused on trying to make the lightest weight system possible and to explore how much a security system can be enhanced for very little data training.
Key Publications
- Saritha Kinkiri, Wim J.C Melis: ‘Reducing Data Storage Requirements for Machine Learning Algorithms Using Principle Component analysis’; 1st International Conference on Applied System Innovation (ICASI) , on 22 to 25th of May 2016, Okinawa, Japan and Published on IEEE ( DOI: 1109/ICASI.2016.7539804) – Best Paper Award.
- Saritha Kinkiri, Wim J.C Melis and Simeon Keates: ‘Creating Patterns for Machine Learning Using Multiple Alignment Making’; 1st International Conference of Human Brain Project (HBP), on 6 to 8th of February 2017, Vienna, Austria (DOI: 3389/978-2-88945-421) – Best Poster Presentation.
- Saritha Kinkiri, Wim J.C Melis and Simeon Keates: ‘Machine Learning for Voice Recognition’; Second Medway Engineering Conference on Systems on 6th June 2017, London, United Kingdom.
- Saritha Kinkiri and Simeon Keates: ‘Identification of a Speaker from Familiar and Unfamiliar voices’; 5th International Conference on robotics and Artificial Intelligence, on 22 to 24th of November 2019, Singapore (ACM, DOI:10.1145/3373724.3373742).
- Saritha Kinkiri and Simeon Keates: ‘Characteristics of a Human Voice’; 2nd International Conference on Advance in Signal Processing and Artificial Intelligence on 1 to 3rd April 2020, Berlin, Germany.
- Saritha Kinkiri and Simeon Keates: ‘Phonemes: An Explanatory study Applied to Identify a speaker’; 2nd International Conference on Machine Learning, Image processing, Network Security and Data Science on 18th – 19th June 2020, Silchar, India (SCOPUS, DOI: 10.1007/978-981-15-6318-8-6).
- Saritha Kinkiri and Simeon Keates: ‘Applications of Speaker Identification for Universal Access’; 22nd International Conference on Human Computer Interaction on 19 -24 July Copenhagen, Denmark (Springer, DOI: 10.1007/978-3-030-49108-6_40).
- Saritha and Simeon Keates: ‘Speaker Identification: Variations of a Human Voice; 6th International Conference on Advances in Computing & Communication Engineering on 22 to 25th July 2020, Las Vegas, USA. (IEEE, DOI: 10.1109/ICACCE49060.2020.9154998)