Ishan studies computer science at the University of Florida.
He has collaborated with the causal AI team at Microsoft and the ADT Baramati team to develop casual machine learning algorithms using the DoWhy library.
The solution will be open sourced on the farmvibes platform and uses causal inference to predict for crop yield of farm. The solution provides accurate adjustments for optimising crop yield and displaying crop parameters for ‘what-if’ scenarios i.e effect on yield if fertilizer increased by 5kg/acre etc.
Ishan has also applied causal machine learning to detect underlying belief systems for LLMs(large language models).
Given a certain text, the problem is to detect what belief system, i.e what fundamental propositions are the basis of reasoning for that text. Done by performing LDA for topic modelling and then detecting cause-effect pairs in text (i.e propositions).