mlsanityData to 7 Oct 2026

Unlocking Multimodal Protein Language Models at Inference Time

arXiv 2608.25855 ↗First posted 26 Aug 2026cs.CE cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Multimodal protein language models (pLMs) learn joint protein sequence-structure distributions, and their generation performance should also depend critically on inference-time sampling strategies. Yet prior work has focused more on model training than on how inference-time strategies behave. In this paper, we establish a three-stage investigation framework to empirically study the inference design space of multimodal pLMs across three representative pLMs and four fundamental tasks. We evaluate vanilla sampling, task-specific classifier-free guidance, and reward-guided beam search on multimodal pLMs, corresponding to controls over sampling distributions, per-step logits, and parallel trajectories. Throughout the complementary advancements centered on exploration-exploitation trade-off, we (1) reveal the suboptimality of default inference protocols and identify task-oriented sampling preferences; (2) observe substantial quantitative gains across tasks, consistently boosting the upper bound performance of multimodal pLMs without updating model parameters; (3) derive conclusions about base models that differ from prior consensus.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 19 of its 36 references that are in the map. First posted on arXiv 26 Aug 2026.

Builds on

19 of 36 references.

References · 19

  1. 2017Neural Discrete Representation Learningcited by 4,650
  2. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameterscited by 1,928
  3. 2024Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Designcited by 294
  4. 2024Guiding a Diffusion Model with a Bad Version of Itselfcited by 286
  5. 2024Scaling up Masked Diffusion Models on Textcited by 195
  6. 2024Simple Guidance Mechanisms for Discrete Diffusion Modelscited by 148
  7. 2024Derivative-Free Guidance in Continuous and Discrete Diffusion Models with Soft Value-Based Decodingcited by 127
  8. 2024Improved motif-scaffolding with SE(3) flow matchingcited by 43
  9. 2025La-Proteina: Atomistic Protein Generation via Partially Latent Flow Matchingcited by 26
  10. 2025Unifying Tree Search Algorithm and Reward Design for LLM Reasoning: A Surveycited by 8
  11. 2025RFG: Test-Time Scaling for Diffusion Large Language Model Reasoning with Reward-Free Guidancecited by 7
  12. 2025Computational Protein Science in the Era of Large Language Models (LLMs)cited by 6
7 more
  1. 2025ProtInvTree: Deliberate Protein Inverse Folding with Reward-guided Tree Searchcited by 5
  2. 2025GLProtein: Global-and-Local Structure Aware Protein Representation Learningcited by 5
  3. 2025HD-Prot: A Protein Language Model for Joint Sequence-Structure Modeling with Continuous Structure Tokenscited by 4
  4. 2026Controlling Repetition in Protein Language Modelscited by 4
  5. 2026Enhancing Molecular Property Predictions by Learning from Bond Modelling and Interactionscited by 3
  6. 2026Geometric Flow Matching for Molecular Conformation Generation via Manifold Decompositioncited by 3
  7. 2025MotifBench: A standardized protein design benchmark for motif-scaffolding problemscited by 1

Built on by

Not cited yet.