Abhishek Thakur

Norway Norway

Abhishek is the World's First Kaggle Quadruple Grandmaster. He has ranked in the top 5 of all four categories on Kaggle. Abhishek's research interests include automated machine learning, hyperparameter optimization, and so on. He is also the author of the applied machine learning book: "Approaching (Almost) Any Machine Learning Problem". In his free time, he likes to make applied machine learning tutorials on his YouTube channel and participate in machine learning competitions. He currently works at Hugging Face where he is building AutoTrain: A no-code tool to train, evaluate and deploy state-of-the-art Machine Learning models automatically. Kaggle: www.kaggle.com/abhishek LinkedIn: www.linkedin.com/in/abhi1thakur Twitter: www.twitter.com/abhi1thakur YouTube: https://www.youtube.com/AbhishekThakurAbhi

Community Contributions

Transfer learning for time series

Transfer learning is a research problem in ML that focuses on storing knowledge gained while solving one problem and applying it to a different but related problem. IT has enabled progress in areas with limited data availability in both CV and NLP domains. Several modern applications of machine learning are built around TL, so it's only natural that people would start thinking about using this approach to time series: while less obvious to formulate (what does it mean to learn time series features across domains), the idea of transfer learning for time series has huge potential. In this notebook we explore the idea in some more detail.
Video/Podcast / 10-12-2022