Cost-aware Bayesian optimization via Pandora's Box Gittins Index, a pre-LLM human-written paper
avt_im · x · 2026-10-17
Author avtim highlights an arXiv paper, "Cost-aware Bayesian Optimization via the Pandora's Box Gittins Index" (Qian Xie, Raul Astudillo, Peter I. Frazier, Ziv Scully, Alexander Terenin), stressing that its core ideas and results predate the current LLM paper wave and that not a word was LLM-written.
- The paper connects the economics decision problem "Pandora's Box" with cost-aware Bayesian optimization, using an acquisition function derived from the Gittins index to explicitly incorporate evaluation costs.
- Empirically it performs well, especially in medium-high dimensions, and the gains carry over to classical Bayesian optimization without explicit costs.
- Presented as a first step toward integrating Gittins index theory into Bayesian optimization.
More from Research
- HF Transformers adds activation checkpoint offloading to CPU to free GPU memory — StasBekman · 2026-10-17
- S4VY: Segment Anything in 4D, with consistent identity across space and time — realChrisChoy · 2026-10-17
- TF-MoDISco optimization merged: 5-15x speedup, 2-3x memory savings, identical output — anshulkundaje · 2026-10-17
- Apple releases RSI-Inference Bench: best agent solves just 26.7% of vLLM/SGLang tasks — SumitGup · 2026-10-17
- Satoshi's distrust of NIST seeds revisited: Claude's P256 discovery shows it wasn't needed — StefanoGogioso · 2026-10-17
- TMS Paper Is Under-Appreciated: Polysemanticity Lives in Weak Activations — livgorton · 2026-10-17