A conference on Machine Learning for Molecular Sciences (ML4SCIENCE 2023), was held at Germanus Springs Resort at Kodaikanal from 9 – 12 March. The event was organized by IIIT Hyderabad, TIFR Hyderabad and Madurai Kamaraj University, and was the first meeting in the proposed annual series after the initial pre-pandemic meeting on ML4Science at IIIT Hyderabad in November 2019. It focused on the applications of Modern Machine Learning methods in the general area of molecular sciences like Molecule/Drug/Material design, Interatomic potentials, MD simulations, etc. The event included various talks and poster presentations that facilitated learning and knowledge sharing amongst the participants that included faculty and students from various institutes across the country. Apart from these, a hackathon was also organised for the students. There were a total of 23 participants, including 6 from IIITH, who were randomly grouped into 11 teams of 2 or 3 participants each team. The students were given SMILES representation of molecules and 4 corresponding unknown properties. The task was to develop and train a ML model to make accurate predictions of these 4 properties given the SMILES string. On the final day they were given 1000 new SMILES strings as a test set for which they were supposed to predict all 4 properties and submit the prediction csv file.
Fourth-year students, Sriram Devata and Kanakala Ganesh Chandan, won the first and second prizes respectively. Rohit Modee, a Ph.D scholar was also awarded for methodology for innovative usage of ML irrespective of performance of the framework used. Sandeep Nagar, a Ph.D student, was also awarded for the best team name.
More details at: https://sites.google.com/view/ml4sci/home