Abstract: To learn a new task as we humans need not always start afresh but rather apply previous learned knowledge to perform the new task. In the same way, "transfer learning" allows a machine learning model to port the knowledge it has acquired during previous training of a task to a new task. Transfer learning has lately shown much promise and is very active area of research. In this tutorial session we will discuss the basic theory of transfer learning and few applications with hand on python coding session with tensorflow 2.0. We will briefly discuss some latest applications of transfer learning like privacy preserving ML.
Module1 :Theory: What is Transfer Learning? Types of Transfer Learning, Challenges
Module2: TL in image classification (hands-on), image to text (hands-on), style transfer
Module3: TL in NLP: text representation, classification (hands-on)
Module4: TL in audio - classification (hands-on), voice style transfer (hands-on)
Module5: TL in handwriting style transfer
* Notebooks will be shared for all.
Basic DL/ML knowledge. Familiarity tensorflow 2.0
Bio: Tamoghna is a AI Solution Architect in Client Computing Group at Intel, working on building next generation AI solutions for edge computing. Prior to this role he has worked as a data scientist at Intel working on various domains like supply chain - inventory optimization, anomaly detection and failure prediction of various IT infrastructure across Intel, building advanced search tools for bug sightings, to name a few. After his Masters in Computer Science from Indian Statistical Institute and a Masters in Mathematics form Calcutta University, he has worked as a research assistant in Microsoft Research India for 3+ years and then moved to other product companies to start his journey in the ML and AI space. He has been teaching AI courses at Intel and trained 250+ employees. Also, he was a core member of the internal AI training academy and AI content development which is a 3-level course in AI for Intel employees. He mentors many folks for their AI projects. He has 4 US patents filed on various innovative AI applications and products and also published few papers related to the work at Intel. He published a book on hands-on transfer learning with Python in 2018 from Packt (packtpub.com) and is working on another book to be published this year from bpb publications.