Summary of Predicting From Strings: Language Model Embeddings For Bayesian Optimization, by Tung Nguyen et al.
Predicting from Strings: Language Model Embeddings for Bayesian Optimization
by Tung Nguyen, Qiuyi Zhang, Bangding Yang, Chansoo Lee, Jorg Bornschein, Yingjie Miao, Sagi Perel, Yutian Chen, Xingyou Song
First submitted to arxiv on: 14 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Bayesian Optimization is widely used in experimental design and blackbox optimization to improve search efficiency, but has traditionally been restricted to regression models that only work with fixed search spaces and tabular input features. The proposed Embed-then-Regress paradigm uses string embedding capabilities of pre-trained language models to apply in-context regression over string inputs. This allows for general-purpose regression in Bayesian Optimization across various domains, including synthetic, combinatorial, and hyperparameter optimization, achieving comparable results to state-of-the-art Gaussian Process-based algorithms. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Bayesian Optimization is a way to improve searching efficiency, but it’s only been used with specific types of data. The new approach, Embed-then-Regress, lets you use Bayesian Optimization with any type of data that can be turned into words or text. This means you can use it for lots of different problems, like finding the best combination of ingredients in a recipe or optimizing the performance of a computer program. |
Keywords
» Artificial intelligence » Embedding » Hyperparameter » Optimization » Regression
