🥇 Last week, I won a GOLD medal in a Kaggle competition of Leash Bio, placing 8th out of around 2000 teams! This was my second competition in the field of life sciences, where I earned a GOLD. And now I have become a Kaggle Master! The problem was to predict which drug-like small molecules (chemicals) will bind to three possible protein targets, which in turn may help develop a drug against blood pressure, diabetes progression, and cancer. There were 133 million molecules being tested in the wet lab that participants used for training and evaluation of the ML models. See the full description of the competition: https://lnkd.in/dR2eDpaW It's been a while since I've seen the lack of proper evaluation of ML models in Drug Discovery. This is one of the proofs, among a few others, that we had in our company showing that the similarity of molecules is an important factor that should be considered for preparing the benchmarks in at least the tasks of small molecule properties prediction. See my full solution, which is built with SMILES-based transformer models, here: https://lnkd.in/dz8Hw5EB. I also credit Simon Steshin for his public work on the critical assessment of ML methods for molecule property prediction tasks.
Congratulations!
Anton Chislov
1mo
Congrats!
Islam Midov, CFA
1mo
wow! Well done!
Olesya N.
1mo
Congratulations Vlad! Great achievement 🥳
Private Equity | Venture Capital | Mergers & Acquisitions
1moThis is so cool Debra Webster