Amer El-Samman
Probabilistic AI models on Chemistry
AI Modelling Chemistry
From the 118 elements in the periodic table, there is a whole universe of molecular possibilities. More unknown molecules than there are stars in the whole universe. Some of those molecules may cure life-threatening diseases, help us process information faster, or store electrical energy that power up metropolitan cities. But what are these molecules and how can we leverage AI technology to help us speed up the search towards the future of material and technology?
The Hidden Language of Graph Neural Networks
A very cool discovery has been made! Decision-making frameworks of graph neural network models conceal a map of reaction formulae. You heard that right! They understand the world as a chemist does. Analogous to how natural language models understand words through math: “King” - “Man” + “Woman” = “Queen". Here, it is the language of graphs which models reactions: “Hydrogen” + “Oxygen” = “Water.”
The Brain’s Brain: The Model Inside Graph Neural Networks
Graph Neural Networks are the some of the most sophisticated probabilistic AI models in chemistry, but what pilots their decision-making? What drives inner operations? Dive in to see how we tap into the deeply hidden decision-making framework of complex AI models.
