Session: The art of formulating a business desire to an AI Research Framework
To implement and add a software/ML feature to an existing product requires development skills. We all can manage to learn that skill over time. However, problem solving - formulating a problem to its exact implementable components- in its broader sense is not just a skill but an art! It becomes even more nuanced in the space of Artificial Intelligence and Machine Learning. Having worked in the world of AI/ML for 6 years, defining unsolved problems of the industry for several startups and drilling down a new AI feature into implementable research components , I have developed an approach to handle them for data/AI related problems. In this talk I will provide a glimpse of that with a walk down of an extremely complex AI task that I researched, solved and even contributed to the novelty of the state-of-the-art Deep Learning algorithms.
For curious ones, here is the project I will outline and use as an example: An AI Framework for Multimodal Understanding of Video Emotions.
Bio
"I am Divya Choudhary (AI/ML Researcher|Yogini|AI/ML Consultant, Data Science Mentor), Sr Research Scientist (AI/ML) with Samsung Research America. In my current role, I have had the opportunity to work on exciting Deep Learning research areas focused on but not limited to the space of Video Models, Video Understanding, Scene Understanding, Multi-modal video emotions recognition, Adversarial Training, on-device ML optimizations etc. A few of my research papers on these projects are mentioned in my resume. As a part of my role, I also lead a team of 5 (mix of RS and Engineers) with research direction for many collaboration projects like Activity Recognition, Video Content Understanding, Text Classification etc. I have worked in the industry for 6 years now in the space of AI/ML. Before my Masters, I worked with 4 diverse high growth startups in Bengaluru, India solving unsolved problems with the application of ML/DL through projects spanning Address Restructuring, Food ETA prediction, Feature Bank, Language Modeling, Clustering etc. During my Masters in Computer Science from USC, I was fortunate to continue my AI/ML research on DARPA research projects at the Information Science Institute, USC. In general, apart from the implementation, I really enjoy the process of problem solving and strategic planning of projects.
Apart from being an AI/ML researcher, I am also a certified yoga instructor who believes in yogic philosophy, mindfulness and leading a healthy life!"