SeqFold: Timestamp-Addressable Compressed States for User Action Sequence Backbones
Existing sequential user modeling methods mainly focus on improving the backbone encoder or learning task-specific user representations for a single prediction objective. We instead study how to export the multi-layer user states of a frozen long-history backbone as compact, timestamp-addressable representations for downstream training backfill and serving. In this setting, the user state at a given timestamp summarizes the user's actions before that time, so downstream systems need only the state aligned to each target timestamp rather than the full hidden-state timeline. To this end, we present \texttt{SeqFold}, a compressed-state interface that combines learned multi-channel fusion, channel-specific temporal decay, and Lookup-Free Quantization (LFQ) to produce compact discrete codes from frozen backbone states. Task-specific decoders then map these shared codes into flexible embedding spaces for downstream models. To make timestamp-addressable export practical at scale, we further develop a fused streaming implementation that updates the fusion recurrence after every action while materializing states only at target timestamps. On large-scale proprietary datasets, SeqFold…
Compressed timestamp-addressable states for user action sequence backbones
- arXiv
- 2611.13704
- Authors
- Mianwei Zhou, Duo Zhang, Henry Tsai, Yuqi Zheng, Jian Li and 5 more
- Categories
- cs.AI · cs.SI