Reproduce “RecMem: Recurrent Memory Compression for Long-Sequence Generative Recommendation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Haoyi Hu, Gaoyang Guo, Yunjia Xi, Wen Chen, Weiwen Liu, Jian Wu, Yuning Jiang, Weinan Zhang
RecMem compresses archived user interactions into a fixed recurrent memory while keeping the latest 200 interactions directly available to its predictor. With 16 slots, it uses a 216-token context. On MerRec, RecMem-Stream reaches 42.36% Recall@10 and 72.20% Recall@200, versus HSTU-1200 at 48.18% and 74.44% using 1,200 tokens. The results show retained predictive signal, not equal accuracy or universal performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.