Machine learning dialects | My summer placement at the Hartree Centre
One of our summer placement students, Daniel Whettam, shares his experiences at the Hartree Centre.
Summer placement student, Daniel Whettam, writes about his experiences developing an understanding of machine learning and AI technologies at the Science and Technology Facilities Council’s (STFC) Hartree Centre.
Over this summer I attended a work placement at the Hartree Centre, supervised by AI Technical Lead Andrew Gargett. I am a Computer Science graduate, starting my Masters in Data Science at the University of Edinburgh in September 2018, so a placement at the Hartree Centre was right up my street! I’m generally interested in machine learning/AI research, so working here gave me the opportunity to get some research experience before I started my masters and would provide a good base for any further study.
During my placement I worked on speech recognition for dialects as part of a wider project on assisted living technologies. My work involved using deep learning and transfer learning to create a system that can understand the Liverpudlian dialect. This linked in with another intern who was developing an emotionally aware chatbot, which would ultimately allow us to create an emotionally aware, Liverpudlian conversational system. This system would be used to improve interaction with assisted-living technologies, providing a much more natural experience than what is currently available.
We also did some work with the Pepper robot, which was very enjoyable. The robot was initially intended to be a part of the assisted living work, however early on into the placement some technical difficulties delayed our work with Pepper, and we shifted to a more general approach that could work on any device.
I also got to spend some time with the data science team and I attended some meetings with them. It was interesting to hear about the sorts of things they were working on. Additionally, I went along to the machine learning journal club with some of the data science team, as well as other Hartree Centre staff. The journal club was another opportunity for me to learn from people with a lot more experience than me, which I found to be very useful.
Over the three months I spent at the Hartree Centre, I had a very positive experience and learned a huge amount. I spent a lot of time reading and learning about deep learning, speech recognition and transfer learning, and as a result my understanding of those areas has skyrocketed, and I have gained some invaluable understanding of the field. I also really enjoyed working in a research environment, and I’m very keen to do more research in the future. I hope to be back at some point in the coming years!
Provide your details to receive regular updates from the STFC Hartree Centre.