Awesome-Offline-Model-Based-Optimization

📰 Must-Read Papers on Offline Model-Based Optimization 🔥

This repository collects important papers for our new survey: "Offline Model-Based Optimization: Comprehensive Review" (will release soon).

• 💻: Code
• 📖: bibtex

Latest Updates

• [2025/03/04] First Release of Awesome-Offline-Model-Based Optimization

🔍 Table of Contents

• 🌟 What is Offline Model-Based Optimization?
• 🔗 Benchmark
• 🎯 Surrogate Modeling
• 🤔 Generative Modeling

🌟 What is Offline Model-Based Optimization?

In offline optimization, the goal is to discover a new design, denoted by $\boldsymbol{x}^*$, that maximizes the objective(s) $\boldsymbol{f}(\boldsymbol{x})$. This is achieved using an offline dataset $\mathcal{D}$, which consists of $N$ designs paired with their property labels. In particular, the dataset is given by where each design vector $\boldsymbol{x}_i$ belongs to a design space $\mathcal{X} \subseteq \mathbb{R}^d$, and each property label $\boldsymbol{y}_i \in \mathbb{R}^m$ contains the corresponding $m$ objective values for that design. The function $\boldsymbol{f}: \mathcal{X} \rightarrow \mathbb{R}^m$ maps a design to its $m$-dimensional objective value vector.

In offline single-objective optimization (SOO), only one objective is considered (i.e., $m=1$). For instance, the design $\boldsymbol{x}$ might represent a neural network architecture, with $f(\boldsymbol{x})$ denoting the network's accuracy on a given dataset.

Offline multi-objective optimization (MOO) extends the framework to simultaneously address multiple objectives using the dataset $\mathcal{D}$. In this setting, the goal is to find solutions that balance competing objectives effectively. For instance, when designing a neural architecture, one might seek to achieve both high accuracy and high efficiency.

🔗 Benchmark

Task

Synthetic Function

• Virtual Library of Simulation Experiments: Test Functions and Datasets (Sonja Surjanovic & Derek Bingham, 2013) 💻 📖

• BayesO Benchmarks: Benchmark Functions for Bayesian Optimization (Jungtaek Kim, 2023) 💻 📖

• Pymoo: Multi-Objective Optimization in Python (Julian Blank & Kalyanmoy Ded, IEEE Access 2020) 💻 📖

• Offline Multi-Objective Optimization (Ke Xue & Rong-Xi Tan et al., ICML 2024) 💻 📖

Real-World System

• An Easy-to-Use Real-World Multi-Objective Optimization Problem Suite (Ryoji Tanabe et al., Applied Soft Computing 2020) 💻 📖

• GTOPX Space Mission Benchmarks (Martin Schlueter et al., SoftwareX 2021) 💻 📖

• SOO-Bench: Benchmarks for Evaluating the Stability of Offline Black-Box Optimization (Hong Qian et al., ICLR 2025) 💻 📖

• Offline Multi-Objective Optimization (Ke Xue & Rong-Xi Tan et al., ICML 2024) 💻 📖

• Design-Bench: Benchmarks for Data-Driven Offline Model-Based Optimization (Brandon Trabucco & Xinyang Geng et al., ICML 2022) 💻 📖

• Data-Driven Offline Optimization for Architecting Hardware Accelerators (Aviral Kumar & Amir Yazdanbakhsh et al., ICLR 2022) 💻 📖

Scientific Design

• ViennaRNA Package 2.0 (Ronny Lorenz et al., Algorithms for Molecular Biology 2011) 💻 📖

• Conservative Objective Models for Effective Offline Model-Based Optimization (Brandon Trabucco & Aviral Kumar et al., ICML 2021) 💻 📖

• Design-Bench: Benchmarks for Data-Driven Offline Model-Based Optimization (Brandon Trabucco & Xinyang Geng et al., ICML 2022) 💻 📖

• Proximal Exploration for Model-Guided Protein Sequence Design (Zhizhou Ren & Jiahan Li et al., ICML 2022) 💻 📖

• Antigen-Specific Antibody Design and Optimization with Diffusion-Based Generative Models for Protein Structures (Shitong Luo & Yufeng Su et al., NeurIPS 2022) 💻 📖

• Offline Multi-Objective Optimization (Ke Xue & Rong-Xi Tan et al., ICML 2024) 💻 📖

• Latent Conservative Objective Models for Offline Data-Driven Crystal Structure Prediction (Han Qi & Xinyang Geng et al., ICLR ML4Materials 2023) 📖

• Bayesian Optimization of Active Materials for Lithium-Ion Batteries (Homero Valladares et al., IEEE IECON 2021) 📖

Machine Learning Model

• Design-Bench: Benchmarks for Data-Driven Offline Model-Based Optimization (Brandon Trabucco & Xinyang Geng et al., ICML 2022) 💻 📖
• Offline Multi-Objective Optimization (Ke Xue & Rong-Xi Tan et al., ICML 2024) 💻 📖
• HPOBench: A Collection of Reproducible Multi-Fidelity Benchmark Problems for HPO (Katharina Eggensperger et al., NeurIPS Datasets and Benchmarks Track 2021) 💻 📖

Evaluation Metric

Usefulness

• Design-Bench: Benchmarks for Data-Driven Offline Model-Based Optimization (Brandon Trabucco & Xinyang Geng et al., ICML 2022) 💻 📖
• Biological Sequence Design with GFlowNets (Moksh Jain et al., ICML 2022) 💻 📖
• Offline Multi-Objective Optimization (Ke Xue & Rong-Xi Tan et al., ICML 2024) 💻 📖
• ParetoFlow: Guided Flows in Multi-Objective Optimization (Ye Yuan & Can Chen et al., ICLR 2025) 💻 📖

Diversity

• Biological Sequence Design with GFlowNets (Moksh Jain et al., ICML 2022) 💻 📖
• Bootstrapped Training of Score-Conditioned Generator for Offline Design of Biological Sequences (Minsu Kim et al., NeurIPS 2023) 💻 📖
• Improving Protein Optimization with Smoothed Fitness Landscapes (Andrew Kirjner & Jason Yim et al., ICLR 2024) 💻 📖
• Diversity By Design: Leveraging Distribution Matching for Offline Model-Based Optimization (Michael S. Yao et al., 2025) 💻 📖

Novelty

• Biological Sequence Design with GFlowNets (Moksh Jain et al., ICML 2022) 💻 📖
• Bootstrapped Training of Score-Conditioned Generator for Offline Design of Biological Sequences (Minsu Kim et al., NeurIPS 2023) 💻 📖
• Improving Protein Optimization with Smoothed Fitness Landscapes (Andrew Kirjner & Jason Yim et al., ICLR 2024) 💻 📖

Stability

• SOO-Bench: Benchmarks for Evaluating the Stability of Offline Black-Box Optimization (Hong Qian et al., ICLR 2025) 💻 📖

🎯 Surrogate Modeling

Auxiliary Loss

• Conservative Objective Models for Effective Offline Model-Based Optimization (Brandon Trabucco & Aviral Kumar et al., ICML 2021) 💻 📖
• RoMA: Robust Model Adaptation for Offline Model-Based Optimization (Sihyun Yu et al., NeurIPS 2021) 💻 📖
• Bidirectional Learning for Offline Infinite-Width Model-Based Optimization (Can Chen et al., NeurIPS 2022) 💻 📖
• Data-Driven Offline Decision-Making via Invariant Representation Learning (Han Qi & Yi Su & Aviral Kumar et al., NeurIPS 2022) 💻 📖
• Bidirectional Learning for Offline Model-Based Biological Sequence Design (Can Chen et al., ICML 2023) 💻 📖
• Parallel-Mentoring for Offline Model-Based Optimization (Can Chen et al., NeurIPS 2023) 💻 📖
• Learning Surrogates for Offline Black-Box Optimization via Gradient Matching (Minh Hoang et al., ICML 2024) 💻 📖
• Boosting Offline Optimizers with Surrogate Sensitivity (Manh Cuong Dao et al., ICML 2024) 💻 📖
• Incorporating Surrogate Gradient Norm to Improve Offline Optimization Techniques (Manh Cuong Dao et al., NeurIPS 2024) 💻 📖
• Offline Model-Based Optimization by Learning to Rank (Rong-Xi Tan et al., ICLR 2025) 💻 📖

Data-Drive Adaptation

• Autofocused Oracles for Model-Based Design (Clara Fannjiang et al., NeurIPS 2020) 💻 📖
• Conservative Objective Models for Effective Offline Model-Based Optimization (Brandon Trabucco & Aviral Kumar et al., ICML 2021) 💻 📖
• Bidirectional Learning for Offline Model-Based Biological Sequence Design (Can Chen et al., ICML 2023) 💻 📖
• Parallel-Mentoring for Offline Model-Based Optimization (Can Chen et al., NeurIPS 2023) 💻 📖
• Importance-Aware Co-Teaching for Offline Model-Based Optimization (Ye Yuan & Can Chen et al., NeurIPS 2023) 💻 📖
• Functional Graphical Models: Structure Enables Offline Data-Driven Optimization (Jakub Grudzien Kuba et al., AISTATS 2024) 📖

Collaborative Ensembling

• Offline Model-Based Optimization via Normalized Maximum Likelihood Estimation (Justin Fu et al., ICLR 2021) 💻 📖
• Conflict-Averse Gradient Optimization of Ensembles for Effective Offline Model-Based Optimization (Sathvik Kolli, 2023) 📖
• Parallel-Mentoring for Offline Model-Based Optimization (Can Chen et al., NeurIPS 2023) 💻 📖
• Importance-Aware Co-Teaching for Offline Model-Based Optimization (Ye Yuan & Can Chen et al., NeurIPS 2023) 💻 📖

Generative Model Integration

• Autofocused Oracles for Model-Based Design (Clara Fannjiang et al., NeurIPS 2020) 💻 📖
• Data-Driven Offline Decision-Making via Invariant Representation Learning (Han Qi & Yi Su & Aviral Kumar et al., NeurIPS 2022) 💻 📖
• Robust Guided Diffusion for Offline Black-Box Optimization (Can Chen et al., TMLR 2024) 💻 📖

🤔 Generative Modeling

Variational Autoencoder (VAE)

• Automatic Chemical Design using a Data-Driven Continuous Representation of Molecules (Rafael Gómez-Bombarelli & Jennifer N. Wei & David Duvenaud & José Miguel Hernández-Lobato et al., ACS central science 2018) 💻 📖
• Conditioning by Adaptive Sampling for Robust Design (David H. Brookes et al., ICML 2019) 💻 📖
• RoMA: Robust Model Adaptation for Offline Model-Based Optimization (Sihyun Yu et al., NeurIPS 2021) 💻 📖
• Latent Bayesian Optimization via Autoregressive Normalizing Flows (Seunghun Lee et al., ICLR 2025) 💻 📖

Generative Adversarial Network (GAN)

• Model Inversion Networks for Model-Based Optimization (Aviral Kumar et al., NeurIPS 2019) 💻 📖
• Data-Driven Offline Decision-Making via Invariant Representation Learning (Han Qi & Yi Su & Aviral Kumar et al., NeurIPS 2022) 💻 📖
• Generative Adversarial Model-Based Optimization via Source Critic Regularization (Michael S. Yao et al., NeurIPS 2024) 💻 📖

Autoregressive Model

• Plug and Play Language Models: A Simple Approach to Controlled Text Generation (Sumanth Dathathri et al., ICLR 2020) 💻 📖
• Model-Based Reinforcement Learning for Biological Sequence Design (Christof Angermueller et al., ICLR 2020) 📖
• Generative Pretraining for Black-Box Optimization (Satvik Mashkaria & Siddarth Krishnamoorthy et al., ICML 2022) 💻 📖
• Bootstrapped Training of Score-Conditioned Generator for Offline Design of Biological Sequences (Minsu Kim et al., NeurIPS 2023) 💻 📖
• ExPT: Synthetic Pretraining for Few-Shot Experimental Design (Tung Nguyen et al., NeurIPS 2023) 💻 📖

Diffusion Model

• Diffusion Models for Black-Box Optimization (Siddarth Krishnamoorthy et al., ICML 2023) 💻 📖
• Exploring Chemical Space with Score-Based Out-of-Distribution Generation (Seul Lee et al., ICML 2023) 💻 📖
• Robust Guided Diffusion for Offline Black-Box Optimization (Can Chen et al., TMLR 2024) 💻 📖
• Guided Trajectory Generation with Diffusion Models for Offline Model-Based Optimization (Taeyoung Yun et al., NeurIPS 2024) 💻 📖
• Design Editing for Offline Model-Based Optimization (Ye Yuan & Youyuan Zhang et al., 2024) 📖
• Low To High-Value Designs: Offline Optimization via Generalized Diffusion (Manh Cuong Dao et al., 2025) 📖

Flow Matching

• Dirichlet Flow Matching with Applications to DNA Sequence Design (Hannes Stark & Bowen Jing et al., ICML 2024) 💻 📖

• ParetoFlow: Guided Flows in Multi-Objective Optimization (Ye Yuan & Can Chen et al., ICLR 2025) 💻 📖

• Flow Q-Learning (Seohong Park et al., 2025) 💻 📖

• AffinityFlow: Guided Flows for Antibody Affinity Maturation (Can Chen et al., 2025) 📖

Energy-Based Model

• Conservative Objective Models Are a Special Kind of Contrastive Divergence-Based Energy Model (Christopher Beckham et al., 2023) 💻 📖

• Protein Discovery with Discrete Walk-Jump Sampling (Nathan C. Frey & Daniel Berenberg et al., ICLR 2024) 💻 📖

• Latent Energy-Based Odyssey: Black-Box Optimization via Expanded Exploration in the Energy-Based Latent Space (Peiyu Yu & Dinghuai Zhang et al., 2024) 📖

Control by Generative Flow Network (GFlowNet)

• Biological Sequence Design with GFlowNets (Moksh Jain et al., ICML 2022) 💻 📖
• Multi-Objective GFlowNets (Moksh Jain et al., ICML 2023) 💻 📖
• Generative Flow Networks Assisted Biological Sequence Editing (Pouya M. Ghari et al., NeurIPS GenBio 2023) 📖
• Improved Off-Policy Reinforcement Learning in Biological Sequence Design (Hyeonah Kim et al., NeurIPS AI for New Drug Modalities 2024) 💻 📖
• Learning to Scale Logits for Temperature-Conditional GFlowNets (Minsu Kim & Joohwan Ko et al., ICML 2024) 💻 📖
• Posterior Inference with Diffusion Models for High-Dimensional Black-box Optimization (Taeyoung Yun et al., 2025) 💻 📖