Excited to share QD-Rec @sakanaai.bsky.social, our new work on open-ended recommendation for decision-making with quality-diversity!!
Blog: pub.sakana.ai/qd-rec/
Code: github.com/SakanaAI/QD-...
Paper: arxiv.org/abs/2610.23470
Thread below 🧵.
Excited to share QD-Rec @sakanaai.bsky.social, our new work on open-ended recommendation for decision-making with quality-diversity!! Blog: pub.sakana.ai/qd-rec/ Code: github.com/SakanaAI/QD-... Paper: arxiv.org/abs/2610.23470 Thread below 🧵.
0 views
Published on 22/10/2026 00:09
1:04
Excited to share QD-Rec @sakanaai.bsky.social, our new work on open-ended recommendation for decision-making with quality-diversity!!
Blog: pub.sakana.ai/qd-rec/
Code: github.com/SakanaAI/QD-...
Paper: arxiv.org/abs/2610.23470
Thread below 🧵.
1 comment
?
Many decisions involve choosing among existing options, e.g., companies to acquire or customers to target. Even when these options already exist, the reasons for choosing them can keep evolving. We thought open-ended search could be a natural fit for exploring both what to recommend and why.
Inspired by recent quality–diversity (QD) research, QD-Rec uses a foundation model to develop hypotheses about why they fit a decision maker’s needs, while using a contextual bandit to select the next items to consider.