mlsanityData to 7 Oct 2026

How to Spend Your Oracle Budget: Practical Guidance for Protein Structure Prediction Models

arXiv 2608.12192 ↗First posted 12 Aug 2026cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Foundation models for protein structure prediction remain unreliable on certain targets. External oracles can flag and correct these failures, but biological oracles are expensive, making oracle budget a critical constraint. Existing guidance methods, such as FK-steering, DPO, and Best K-of-N sampling, differ in how they spend this budget, yet no systematic comparison exists to guide method selection. To bridge this gap, we benchmark these methods alongside the recently proposed Optimisation Over Outputs (O3), which applies off-the-shelf optimisers within a generative model's latent subspace. We extend the usage of O3 to protein structure prediction models. Overall, our work provides the first practical reference for oracle budget-aware guidance. Our evaluation on two protein targets, calmodulin (1CLL) and E. coli aspartate transcarbamoylase (9EEH), reveals that no single method consistently dominates across all budgets and oracles. Specifically, O3 proves most effective at low oracle budgets, while FK-steering and DPO demonstrate improved performance as the budget increases. We distil these findings into actionable recommendations for practitioners operating under real-world oracle-budget constraints.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 13 of its 33 references that are in the map. First posted on arXiv 12 Aug 2026.

Builds on

13 of 33 references.

References · 13

  1. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  2. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  4. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 2,994
  5. 2023Diffusion Model Alignment Using Direct Preference Optimizationcited by 767
  6. 2016Automatic chemical design using a data-driven continuous representation of moleculescited by 762
  7. 2025A General Framework for Inference-time Scaling and Steering of Diffusion Modelscited by 201
  8. 2023Unexpected Improvements to Expected Improvement for Bayesian Optimizationcited by 144
  9. 2020Sample-Efficient Optimization in the Latent Space of Deep Generative Models via Weighted Retrainingcited by 121
  10. 2024Understanding the performance gap between online and offline alignment algorithmscited by 110
  11. 2022Local Latent Space Bayesian Optimization over Structured Inputscited by 78
  12. 2025Return of the Latent Space COWBOYS: Re-thinking the use of VAEs for Bayesian Optimisation of Structured Spacescited by 13
1 more
  1. 2024Linear combinations of latents in generative models: subspaces and beyondcited by 8

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Protein Structure Tokenization: Benchmarking and New Recipe
  2. 2025Protein-SE(3): Benchmarking SE(3)-based Generative Models for Protein Structure Design
  3. 2026Scaling Atomistic Protein Binder Design with Generative Pretraining and Test-Time Compute
  4. 2025ProteinGuide: On-the-fly property guidance for protein sequence generative models
  5. 2025Can Molecular Foundation Models Know What They Don't Know? A Simple Remedy with Preference Optimization
  6. 2026Generative Modeling in Protein Design: Neural Representations, Conditional Generation, and Evaluation Standards
  7. 2025SPID: Distilled Protein Backbone Generation
  8. 2025ConfRover: Simultaneous Modeling of Protein Conformation and Dynamics via Autoregression