#recsys

Four simple models that use only past sales (window count, time decay, velocity, z-score), run on the same fake data so you can watch them disagree.

Most recommenders treat your history as an unordered bag. Model the sequence instead and predictions start to feel like mind-reading.

Meta GEM 推薦基石模型深剖：大廠限定的工程地獄與落地取捨 Meta 最近揭露了其廣告推薦中央大腦 GEM（Generative Ads Recommendation...

When I wanted to build an item-item recommendation system with just implicit ratings (e.g. viewed,...

In the last week or so, I've been building an app to generate real-time movie recommendations based...