Search papers, labs, and topics across Lattice.
RecPFN introduces a novel prior-fitted network that leverages in-context learning for sequential recommendations, pretrained on synthetic clickstream data derived from a structural causal prior. This approach allows for efficient Bayesian-style inference from minimal support sets, enabling a lightweight transformer to generate next-item predictions in a single forward pass without weight updates. The model achieves state-of-the-art zero-shot performance across eight benchmarks, demonstrating robustness to domain shifts and competitive results with supervised methods in low-compute scenarios.
Achieving state-of-the-art zero-shot recommendation performance, RecPFN redefines efficiency in sequential recommendation systems by leveraging synthetic data and causal priors.
We introduce RecPFN, a prior-fitted network that brings in-context learning to sequential recommendation. RecPFN is pretrained entirely on synthetic clickstream environments sampled from a broad structural causal prior, enabling it to amortize Bayesian-style inference from a small support set. At inference, a lightweight decoder-only transformer conditions on a handful of domain sequences and produces next-item predictions for queries in a single forward pass, without any weight updates. Across eight public benchmarks, RecPFN achives state-of-the-art zero-shot performance while remaining strongly competitive with supervised methods in low-compute and low-data regimes. It is deployment-efficient and robust to domain shift, outperforming strong zero-shot baselines that rely on large real-interaction corpora. RecPFN provides a practical path toward generalizable, data-efficient recommenders and opens avenues for richer priors, longer-context ICL, and multimodal extensions. Code for training and evaluation will be made publicly available by the conference date.
