How I Became a Perfumer Podcast №8 – AI-Powered Perfumery with Milan Pavlovic
Mar 10, 2024
Milan Pavlovic, a software engineer and fragrance aficionado, founded Scentalytics to revolutionize the perfume industry with AI. He discusses how artificial intelligence enhances rather than replaces traditional perfumery by analyzing scent preferences. Milan highlights the technical challenges of creating an AI perfume app that compares formulas, and underscores the vital role of data in personalizing fragrances. The conversation delves into innovative strategies for connecting brands with perfumers, showcasing how data can redefine scent creation.
AI Snips
Chapters
Transcript
Episode notes
AI Enhances Perfumer Creativity
- Milan Pavlovic uses AI models to predict fragrance characteristics before creation based on desired notes.
- His goal is to augment perfumers' creativity, not replace them, speeding up their creative process.
Mathematical Fragrance Similarity
- Milan converts fragrance notes into numerical vectors to calculate similarity scores between perfumes.
- High similarity indicates closely related scents in perception, though not identical formulas.
Challenges Accessing Perfume Data
- Perfume formulas are usually proprietary and rarely shared, making exact data hard to obtain.
- Milan relies on publicly available product description data for his AI models due to this limitation.