Digitizing our Chemical Sense - Building the RGB of Olfaction with AI

Harvard John A. Paulson School Of Engineering And Applied Sciences, 150 Western Avenue, Room 3.301, Boston, 02134

This event highlights the innovative ways AI is being used to tackle complex research questions within industry settings. Benjamin will delve into the interdisciplinary approach of combining AI with chemistry to digitize and decode the sense of smell, showcasing the vast potential for AI applications beyond traditional boundaries.

Apr 24, 3:45 – 5:15 PM



Key Themes

Machine LearningTensorFlow / Keras

About this event

The properties of the physical world come to life through our human senses. Digitizing these senses allows us to catalog, search and design perceptions. We have made notable progress in the domain of vision and hearing, but what about the rest? Olfaction is a chemical sense, and in this talk I will present our work recently published in Science (DOI: 10.1126/science.ade4401) on how we build a digital representation of smell for individual compounds, which we call the Primary Odor Map (POM) and which can be used in a variety of olfactory tasks. The talk will address topics of deep learning, molecular representations, selecting and purchasing molecules, training humans to reliably rate odors, as well as various applications of our digital representation.


  • Benjamin Sanchez-Lengeling

    Google Deepmind

    Research Scientist


  • Luis Henrique Simplicio Ribeiro

    GDSC Lead

  • Alison Yao

    Technical Lead

  • Emilia Mazzolenis

    Outreach Lead

  • Aniedi Udo-Obong

    Regional Leader

Contact Us