A-NeSI: A Scalable Approximate Method for Probabilistic Neurosymbolic Inference

Emile van Krieken,Thiviyan Thanapalasingam,Jakub M. Tomczak,Frank van Harmelen,Annette ten Teije

2023-09-23

Abstract:We study the problem of combining neural networks with symbolic reasoning. Recently introduced frameworks for Probabilistic Neurosymbolic Learning (PNL), such as DeepProbLog, perform exponential-time exact inference, limiting the scalability of PNL solutions. We introduce Approximate Neurosymbolic Inference (A-NeSI): a new framework for PNL that uses neural networks for scalable approximate inference. A-NeSI 1) performs approximate inference in polynomial time without changing the semantics of probabilistic logics; 2) is trained using data generated by the background knowledge; 3) can generate symbolic explanations of predictions; and 4) can guarantee the satisfaction of logical constraints at test time, which is vital in safety-critical applications. Our experiments show that A-NeSI is the first end-to-end method to solve three neurosymbolic tasks with exponential combinatorial scaling. Finally, our experiments show that A-NeSI achieves explainability and safety without a penalty in performance.

Machine Learning,Artificial Intelligence,Logic in Computer Science

What problem does this paper attempt to address?

The paper aims to address the scalability issue of Probabilistic Neurosymbolic Inference (PNL) in Neurosymbolic Learning. Specifically, existing PNL frameworks like DeepProbLog require exponential time complexity for exact inference, which greatly limits their application scope. Therefore, the authors propose a new method—Approximate Neurosymbolic Inference (A-N ESI)—to achieve approximate inference with polynomial time complexity, enabling the handling of tasks with exponential combinatorial complexity.

The main contributions of A-N ESI include:
1. Achieving approximate inference without altering the probabilistic logic semantics;
2. Training using data generated from background knowledge;
3. Generating symbolic explanations to elucidate prediction results;
4. Ensuring the satisfaction of logical constraints during testing, which is crucial for applications with high safety requirements.

Experimental results show that A-N ESI is the first to enable end-to-end learning on three tasks with exponential combinatorial complexity in neurosymbolic tasks, without sacrificing interpretability and safety in terms of performance.