Embeddings are a way of representing hobbies as points in a space where similar activities cluster together, allowing recommendation systems to find activities closer to your preferences even if you've never tried them. This approach captures the nuance that someone drawn to painting might also enjoy pottery—not because both are art, but because both require a particular kind of patient focus and material intuition.
Activity embeddings are high-dimensional numerical representations of hobbies and recreational activities that encode their underlying characteristics such as physical intensity, social dynamics, creativity level, and required equipment into a format that AI models can compare and cluster. By mapping both activities and user preferences into the same embedding space, an AI can identify which hobbies a person is most likely to enjoy even if they have never tried them.
This technique powers next-generation hobby discovery tools that go far beyond simple category matching. Instead of recommending rock climbing just because a user likes hiking, an AI using activity embeddings identifies the deeper preference patterns such as a love of problem-solving under physical pressure and surfaces activities that satisfy those patterns in unexpected and highly personalized ways.
Peri can explain this concept, give practical examples, help you decide whether it applies to your situation, or recommend a journey if appropriate.
Explore related journeys or tell Peri what you're working through.