This document discusses using machine learning for chemical reaction design and discovery at the Institute for Chemical Reaction Design and Discovery (ICReDD). ICReDD was established in 2018 at the Tokyo Institute of Technology to develop machine learning methods for predicting catalytic properties and discovering new reactions. The goal is to accelerate heterogeneous catalysis research and development. ICReDD researchers are using machine learning to predict properties like d-band centers and adsorption energies from catalyst surfaces and correlate these with experimentally measured catalytic activities for reactions like NOx reduction. This can help design new catalyst materials for reducing emissions of pollutants like CO, HC, and NOx from vehicles.