MapPFN: Learning Causal Perturbation Maps in Context

arXiv:2601.21092v2 Announce Type: replace Abstract: Planning effective interventions in biological systems requires treatment-effect models that adapt to unseen biological contexts by identifying their specific underlying mechanisms. Yet single-cell perturbation datasets span only a handful of biological contexts, and existing methods cannot leverage new interventional evidence at inference time to adapt beyond their training data. To meta-learn a perturbation effect estimator, we present MapPFN, a prior-data fitted network (PFN) pre-trained on synthetic data generated from a prior over causal perturbations. Given a set of experiments, MapPFN uses in-context learning to predict post-perturbation distributions. Pre-trained on in silico gene knockouts alone, MapPFN identifies differentially expressed genes on par with models trained on real single-cell data. Fine-tuned, it consistently outperforms baselines across downstream datasets. Our code, model and data are available at https://marvinsxtr.github.io/MapPFN.

Related

• Transformers for dynamical systems learn transfer operators in-context
• Transformers Learn Latent Mixture Models In-Context via Mirror Descent
• A robust and adaptive MPC formulation for Gaussian process models
• Learning to Plan, Planning to Learn: Adaptive Hierarchical RL-MPC for Sample-Efficient Decision Making
• Dissipative Latent Residual Physics-Informed Neural Networks for Modeling and Identification of Electromechanical Systems

Source: arXiv cs.LG | 2026-04-22