The JKU Participates in the World's Largest AI Conference

JKU PhD candidates Johannes Schimunek and Philipp Seidl took part in a conference in Rwanda.

F.l.: Philipp Seidl and Johannes Schimunek , photo credit:personal
F.l.: Philipp Seidl and Johannes Schimunek , photo credit:personal

Johannes Schimunek and Philipp Seidl, both PhD candidates at the LIT AI Lab & Institute for Machine Learning at the JKU, held a presentation about their research titled: "Context-Enriched Molecule Representations Improve Few-Shot Drug Discovery" at the prestigious ICLR , opens an external URL in a new window(International Conference on Learning Representation) AI conference in Kigali, the capital of Rwanda.

Their research involves considering molecular properties within the context of many other molecules before classifying these molecules. The project was conducted in collaboration with researchers from Merck, a German pharmaceutical company, and also ties into currently booming language models as they, too, use context to make predictions. The AI researchers remarked: "Before predicting a word, the word is always considered in the context of other words. For example, the word ‘it’ on its own does not mean much. You first have to look for a word, or context, that it refers to, such as ‘the cat’."