You get smart. Then you get uncomfortably smart. That rare, off-the-charts, stop-everyone-in-their-tracks type of ...

Implementation and reproducible experiments behind the research paper "Improving Sequential Recommendation with LLMs", which substantially extends our RecSys paper. The paper is uploaded to Arxiv.

Abstract: Combining the classical Kalman filter (KF) with a deep neural network (DNN) enables tracking in partially known state space (SS) models. A major limitation of current DNN-aided designs stems ...