Informationen zu Lehrveranstaltungen in →Opal / Information about lectures can be found in →Opal

Skip to content.


search  |  internal  |  deutsch
Personal tools
TU Dresden » Faculty of Mechanical Science and Engineering » Institute for Materials Science » Chair of Materials Science and Nanotechnology

Thursday, 10 June 2021
(at 16:00 in room Zoom meeting)
Add to your Calendar

Learning to Smell: Using Deep Learning to Predict the Olfactory Properties of Molecules

Dr. Alexander B. Wiltschko

Google Brain

Predicting the relationship between a molecule's structure and its odor remains a difficult, decades-old task. This problem, termed quantitative structure-odor relationship (QSOR) modeling, is an important challenge in chemistry, impacting human nutrition, manufacture of synthetic fragrance, the environment, and sensory neuroscience. We propose the use of graph neural networks for QSOR, and show they significantly out-perform prior methods on a novel data set labeled by olfactory experts. Additional analysis shows that the learned embeddings from graph neural networks capture a meaningful odor space representation of the underlying relationship between structure and odor, as demonstrated by strong performance on two challenging transfer- learning tasks. Machine learning has already had a large impact on the senses of sight and sound. Based on these early results with graph neural networks for molecular properties, we hope machine learning can eventually do for olfaction what it has already done for vision and hearing.

Announcement (pdf)

Invited by Perceptronics

Within the nanoSeminar

last modified: 2021.06.10 Thu
author: webadmin