Introduction to Seminars

To keep up with the latest research and to learn something new every week, the ML team at Skit, runs seminars. The idea of a seminar is similar to a reading group specific to a topic/domain with focused time-bound goals (like implementation or a certain level of understanding of a topic). The people involved in a seminar document and present their learnings along the way.

These exist to complement the world of paper readings. While Paper Reading sessions are about reading more papers and sharing what we like, Reading Seminars are about learning something specific. These are much more structured and pointed towards a a goal. The idea is to have deeper discussions, over longer periods of time, about topics that might interest you. Either directly or indirectly, this will lead to a better output (from the speaker) and experience (for the audience) in the Paper Reading sessions that follow.

Some examples of past seminars that we have had :

• Computational Paralinguistics
• Multi-Style TTS
• Speech Representation Learning
• Knowledge Distillation
• End2End ASR
• State of the Art in ML
• Dialog State Trackers
• Code-mixing in Speech and Language

Our seminars are open for people outside the company as well, so incase you want to join a seminar, please write to us @machine-learning@skit.ai.