PulseAugur
EN
LIVE 08:05:54

Airbnb's SIFT Transformer boosts booking conversion with personalized filter ranking

By PulseAugur Editorial · [1 sources] ·

Researchers at Airbnb have developed SIFT (Search Intent-to-Filter Transformer), a novel transformer-based model designed to enhance personalized filter ranking in two-sided marketplaces. SIFT learns guest preferences directly from behavioral sequences, replacing traditional manual feature engineering with a unified guest representation that supports multiple prediction tasks. This approach significantly improves booking conversion and filter engagement, as demonstrated by offline metrics showing a +51.9% increase in booking PR-AUC and a +62.8% increase in amenity-engagement PR-AUC over the previous baseline. In production, SIFT has led to a +20.0% increase in engagement with recommended filters and has been successfully extended to new filter types, including a hotel-intent filter that boosted uncancelled hotel bookings by +3.8%. AI

IMPACT Enhances personalization in e-commerce platforms, potentially improving conversion rates and user experience.

RANK_REASON Academic paper detailing a new model and its performance metrics. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Airbnb's SIFT Transformer boosts booking conversion with personalized filter ranking

How we ranked this

Signal score
19 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Academic paper detailing a new model and its performance metrics. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
product, paper
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. arXiv cs.LG TIER_1 English(EN) · Shashank Dabriwal, Tanya Piplani, Hao Li, Yiwei Wang, Ashish Jain, Kedar Bellare, Stephanie Moyerman ·

    SIFT: Search Intent-to-Filter Transformer for Multi-Task Personalized Filter Ranking at Airbnb

    arXiv:2610.07810v1 Announce Type: new Abstract: Search filters help guests navigate vast catalogs in two-sided marketplaces like Airbnb, and recommending the right filters can meaningfully lift booking conversion. Many such production filter-ranking systems, however, represent th…

RELATED ENTITIES

RELATED TOPICS