Leveraging Image Aesthetics and Photo Sorting Algorithms to Enhance Airbnb Listings
Airbnb’s new computer‑vision pipeline trains a deep‑learning aesthetic model with an EMD loss to rank photos, automatically sorts new‑listing images by design and room type, and scales real‑time similarity search via HNSW‑based ANN on AWS OpenSearch, boosting click‑through, bookings, and enabling unsupervised visual recommendations.
Airbnb utilizes computer vision to extract aesthetic and design information from property photos, which text or fixed facility lists cannot convey. The company developed a deep learning-based image aesthetic evaluation model using EMD loss function to predict photographer ratings, improving ad CTR and booking rates.
The article details an automatic photo sorting algorithm for new listings, prioritizing photos based on home design aesthetics and room classification to maximize booking probability. This system was deployed in 2021, significantly boosting new listing success rates.
For scalability, Airbnb implemented approximate nearest neighbor (ANN) search using HNSW indexing for real-time visual similarity searches. This architecture supports millions of listings through distributed indexing on AWS OpenSearch, enabling efficient real-time image search without exhaustive pairwise comparisons.
Additionally, the research explores unsupervised contrastive learning for image embeddings, reducing reliance on manual annotations. Applications include ad quality optimization, similar listing recommendations, and expanding Airbnb's specialty categories through visual analysis.
Airbnb Technology Team
Official account of the Airbnb Technology Team, sharing Airbnb's tech innovations and real-world implementations, building a world where home is everywhere through technology.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.