Learning to Taste: A Multimodal Wine Dataset
This work addresses the challenge of modeling complex human sensory perception like taste for applications in food and beverage AI, but it is incremental as it builds on existing multimodal and embedding methods.
The authors tackled the problem of understanding wine flavor by creating WineSensed, a large multimodal dataset with images, reviews, and fine-grained flavor annotations from human experiments, and demonstrated that their concept embedding algorithm improves coarse flavor classification and aligns with human perception.
We present WineSensed, a large multimodal wine dataset for studying the relations between visual perception, language, and flavor. The dataset encompasses 897k images of wine labels and 824k reviews of wines curated from the Vivino platform. It has over 350k unique bottlings, annotated with year, region, rating, alcohol percentage, price, and grape composition. We obtained fine-grained flavor annotations on a subset by conducting a wine-tasting experiment with 256 participants who were asked to rank wines based on their similarity in flavor, resulting in more than 5k pairwise flavor distances. We propose a low-dimensional concept embedding algorithm that combines human experience with automatic machine similarity kernels. We demonstrate that this shared concept embedding space improves upon separate embedding spaces for coarse flavor classification (alcohol percentage, country, grape, price, rating) and aligns with the intricate human perception of flavor.