Scentience Sigma

The Scentience Sigma app is a multimodal data collection tool designed to aid in the construction of world models for robotics and AI, with a particular emphasis on olfactory and chemical data. Sigma fuses six different data modalities together into a single time-aligned temporal data stream: olfaction, vision, audio, language, inertial, and environmental. Sigma was built to support an open community initiative aimed at advancing machine olfaction and multimodal AI through shared datasets and reproducible benchmarks. All data captured through the app follows standardized protocols and can be contributed - at the user’s discretion - to a growing open research dataset governed by clear licensing and attribution requirements.

Onboard the app are several machine learning models trained to identify the world around you and infer associations between aroma/chemical data and visual, audio, inertial, and language data. The intent of Sigma is to help interested researchers construct and export data to better understand the olfactory patterns around us.

Sigma is currently for RESEARCH PURPOSES ONLY. It is intended for academic researchers interested in building training data to develop olfaction and world models for AI and robotics. A Scentience olfaction device is required for use.