Search papers, labs, and topics across Lattice.
This paper presents a novel two-stage architecture for real-time click-through rate (CTR) prediction that effectively utilizes long interaction histories while adhering to stringent serving constraints in online advertising. By decoupling history encoding from real-time inference, an offline transformer encodes user interactions into a compact representation, which is then combined with recent events during ad serving. The approach achieves 72-80% of the performance of a full-history model while significantly improving ranking metrics and revenue in production A/B tests without increasing latency.
A two-stage architecture for CTR prediction boosts ad ranking metrics by over 2.7% while maintaining real-time serving speeds.
Long interaction histories are among the most informative inputs for click-through rate (CTR) prediction, yet in online advertising they collide with a hard serving constraint: ads must be scored within a few hundred milliseconds to enter the auction, which rules out running a large sequence encoder at request time. We describe how a production advertising system resolves this conflict by decoupling history encoding from real-time inference. A high-capacity offline transformer asynchronously encodes the user's full cross-surface interaction history into a compact representation cached in a feature store, while a lightweight runtime model combines this cached representation with the user's most recent events and the request context at serving time. The offline encoder is pre-trained autoregressively on large-scale interaction logs with a dual objective - feedback prediction and next-item prediction - and the two-stage architecture is then fine-tuned for CTR prediction on the target advertising surface. Offline, the split design recovers 72-80% of the quality of a full-history runtime transformer that would be too expensive to deploy, and the cached representation is robust enough to staleness to permit inexpensive refresh policies. In production A/B experiments, the system improves the primary ranking metric by +2.77% in search advertising and +2.1% on the Yandex Advertising Network, with revenue gains of +2.26% and +0.43% respectively - without increasing serving latency.
