Clubhouse logo

Armish Sonkar



Language models are as conscious as the user using it Otherwise, I convert business problems into data problems & build solutions: risk models, pricing, cyber risk quantification, forecasting, optimisation, data science / engineering solutions & machine learning models for Financial services, Insurance, Telcos and consulting firms || Professional problem solver “Gentlemen, this is a war room, you cannot fight in here” - Dr. Strangelove or how I began to love the bomb, loved this film Move-78 would be the name of my resistance movement if machine takes over (Ref: Lee Sedol Vs. AlphaGo, 4th game). Self supervised learning looks to be the future BTW, Deep Neural Nets comes close to be Artificial Intelligence (A.I) as it tries to mimic natural Intelligence & memory using basic calculus, rest are single layer pattern recognition techniques either statistical / econometric / probabilistic / tree based traditional machine learning systems || Reinforcement learning is constraint learning system with boundary conditions which btw is a supervision : Also, Programmed/automated behaviour is different than learning behaviour, software programs (no matter how elaborate & detailed) are not even learning systems & definitely not A.I