Goal-Oriented Sequential Bayesian Experimental Design for Causal Learning
Not on the map. Why?
Abstract (arXiv)
We present GO-CBED, a goal-oriented Bayesian framework for sequential causal experimental design. Unlike conventional approaches that select interventions aimed at inferring the full causal model, GO-CBED directly maximizes the expected information gain (EIG) on user-specified causal quantities of interest, enabling more targeted and efficient experimentation. The framework is both non-myopic, optimizing over entire intervention sequences, and goal-oriented, targeting only model aspects relevant to the causal query. To address the intractability of exact EIG computation, we introduce a variational lower bound estimator, optimized jointly through a transformer-based policy network and normalizing flow-based variational posteriors. The resulting policy enables real-time decision-making via an amortized network. We demonstrate that GO-CBED consistently outperforms existing baselines across various causal reasoning and discovery tasks-including synthetic structural causal models and semi-synthetic gene regulatory networks-particularly in settings with limited experimental budgets and complex causal mechanisms. Our results highlight the benefits of aligning experimental design objectives with specific research goals and of forward-looking sequential planning.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 41 of its 67 references that are in the map. First posted on arXiv 10 Jul 2025.
Builds on
41 of 67 references.
References · 41
- 2016Density estimation using Real NVPcited by 2,042
- 2019On Variational Bounds of Mutual Informationcited by 559
- 2015Causal inference using invariant prediction: identification and confidence intervalscited by 536
- 2016Stein Variational Gradient Descent: A General Purpose Bayesian Inference Algorithmcited by 507
- 2018Pyro: Deep Universal Probabilistic Programmingcited by 380
- 2020Differentiable Causal Discovery from Interventional Datacited by 218
- 2021D'ya like DAGs? A Survey on Structure Learning and Causal Discoverycited by 211
- 2023Modern Bayesian Experimental Designcited by 125
- 2021Deep Adaptive Design: Amortizing Sequential Bayesian Experimental Designcited by 106
- 2021DiBS: Differentiable Bayesian Structure Learningcited by 106
- 2022Amortized Inference for Causal Structure Learningcited by 90
- 2017Causal Structure Learningcited by 84
29 more
- 2021Efficient Neural Causal Discovery without Acyclicity Constraintscited by 82
- 2021BCD Nets: Scalable Variational Approaches for Bayesian Causal Discoverycited by 68
- 2023Prediction-Oriented Bayesian Active Learningcited by 62
- 2021Implicit Deep Adaptive Design: Policy-Based Experimental Design without Likelihoodscited by 61
- 2022Causal Discovery in Heterogeneous Environments Under the Sparse Mechanism Shift Hypothesiscited by 55
- 2019ABCD-Strategy: Budgeted Experimental Design for Targeted Causal Structure Discoverycited by 52
- 2022Interventions, Where and How? Experimental Design for Causal Models at Scalecited by 49
- 2022Optimizing Sequential Experimental Design with Deep Reinforcement Learningcited by 46
- 2024Optimal experimental design: Formulations and computationscited by 45
- 2017Cost-Optimal Learning of Causal Graphscited by 44
- 2021Variational Causal Networks: Approximate Bayesian Inference over Causal Structurescited by 43
- 2017Budgeted Experiment Design for Causal Structure Learningcited by 38
- 2020Active Invariant Causal Prediction: Experiment Selection through Stabilitycited by 37
- 2022Active Bayesian Causal Inferencecited by 35
- 2013Data Analysis with Bayesian Networks: A Bootstrap Approachcited by 32
- 2021Bayesian Sequential Optimal Experimental Design for Nonlinear Models Using Policy Gradient Reinforcement Learningcited by 24
- 2023Causal machine learning for single-cell genomicscited by 23
- 2018Goal-Oriented Optimal Design of Experiments for Large-Scale Bayesian Linear Inverse Problemscited by 19
- 2019Optimal experimental design via Bayesian optimization: active causal structure learning for Gaussian process networkscited by 18
- 2021Near-Optimal Multi-Perturbation Experimental Design for Causal Structure Learningcited by 17
- 2023Differentiable Multi-Target Causal Bayesian Experimental Designcited by 15
- 2024Variational Bayesian Optimal Experimental Design with Normalizing Flowscited by 13
- 2024Amortized Active Causal Induction with Deep Reinforcement Learningcited by 10
- 2024Goal-Oriented Bayesian Optimal Experimental Design for Nonlinear Models using Markov Chain Monte Carlocited by 9
- 2023Variational Sequential Optimal Experimental Design using Reinforcement Learningcited by 8
- 2022A Unified Experiment Design Approach for Cyclic and Acyclic Causal Modelscited by 7
- 2021An efficient method for goal-oriented linear Bayesian optimal experimental design: Application to optimal sensor placemencited by 5
- 2022Trust Your nabla: Gradient-based Intervention Targeting for Causal Discoverycited by 5
- 2024A Likelihood-Free Approach to Goal-Oriented Bayesian Optimal Experimental Designcited by 2
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2026Why LLMs Fail at Causal Discovery and How Interventional Agents Escape
- 2026Active Causal Experimentalist (ACE): Learning Intervention Strategies via Direct Preference Optimization
- 2022Causal Explanations for Sequential Decision Making Under Uncertainty
- 2022Generalizing Goal-Conditioned Reinforcement Learning with Variational Causal Reasoning
- 2026CausalArena: Benchmarking Causal Discovery in the Foundation Model Era
- 2026Amortising Bayesian Experimental Design for Sequential Information Gathering in LLMs
- 2025Competition is the key: A Game Theoretic Causal Discovery Approach
- 2026TabCausal: Pretraining Across Causal Environments for Tabular Causal Discovery