# BOIDS: High-dimensional Bayesian Optimization via Incumbent-guided Direction Lines and Subspace Embeddings

arXiv: https://arxiv.org/abs/2412.12918 · HTML: https://mlsanity.com/paper/2412.12918/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 51 references that are in the map.
- First posted on arXiv 17 Dec 2024.

## Builds on (most cited first, 3 of 3)

- [OpenAI Gym](https://mlsanity.com/paper/1606.01540/) (2016)
- [High Dimensional Bayesian Optimization Using Dropout](https://mlsanity.com/paper/1802.05400/) (2018)
- [Sustainable Concrete via Bayesian Optimization](https://mlsanity.com/paper/2310.18288/) (2023)

## Built on by (most cited first, 4 of 4)

- [MOBO-OSD: Batch Multi-Objective Bayesian Optimization via Orthogonal Search Directions](https://mlsanity.com/paper/2510.20872/) (2025)
- [MOCA-HESP: Meta High-dimensional Bayesian Optimization for Combinatorial and Mixed Spaces via Hyper-ellipsoid Partitioning](https://mlsanity.com/paper/2508.06847/) (2025)
- [Sample-Efficient Optimisation over the Outputs of Generative Models](https://mlsanity.com/paper/2509.23800/) (2025)
- [Trajectory-Aware Best-Arm Identification for Local Search Allocation in Bayesian Optimization and Beyond](https://mlsanity.com/paper/2605.31050/) (2026)

## Abstract (arXiv)

When it comes to expensive black-box optimization problems, Bayesian Optimization (BO) is a well-known and powerful solution. Many real-world applications involve a large number of dimensions, hence scaling BO to high dimension is of much interest. However, state-of-the-art high-dimensional BO methods still suffer from the curse of dimensionality, highlighting the need for further improvements. In this work, we introduce BOIDS, a novel high-dimensional BO algorithm that guides optimization by a sequence of one-dimensional direction lines using a novel tailored line-based optimization procedure. To improve the efficiency, we also propose an adaptive selection technique to identify most optimal lines for each round of line-based optimization. Additionally, we incorporate a subspace embedding technique for better scaling to high-dimensional spaces. We further provide theoretical analysis of our proposed method to analyze its convergence property. Our extensive experimental results show that BOIDS outperforms state-of-the-art baselines on various synthetic and real-world benchmark problems.