QD-Rec evolves diverse recommendations with quality-diversity search
Original title: QD-Rec: Open-Ended Evolution of Recommendations via Quality-Diversity
A contextual bandit selects items and a foundation model writes hypotheses kept as archive stepping stones.
QD-Rec formalizes open-ended recommendation as a diverse set of items paired with natural-language hypotheses for a decision maker. New recommendations enter an archive if novel or better than their nearest neighbor. Across M&A, customer targeting, and paper and book recommendation, the paper says QD-Rec had the highest mean quality and the largest archive in most settings.
- Each recommendation pairs an item with a reason it fits the decision maker.
- The method is aimed at fixed choice sets with open reasons, not a single ranked list.
- Archive admission requires novelty or improvement over the nearest neighbor.
From the original article
1 min read at huggingface.co
Open-ended search combined with foundation models has produced advances such as faster algorithms, new results on open mathematical problems, and self-improving agents. We extend this paradigm to recommendation for decision-making, focusing on decisions where the set of possible choices is fixed but the reasons for choosing them are not. In high-stakes decisions, e.
By Takashi Ishida, Taro Makino, Issa Sugiura · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.