# Towards Data-and Knowledge-Driven Artificial Intelligence: A Survey on Neuro-Symbolic Computing

**Authors:** Wenguan Wang, Yi Yang, Fei Wu  
**arXiv:** [2210.15889](https://arxiv.org/abs/2210.15889)  
**Date:** 2022-10-28  
**Summary page:** https://www.summarizepaper.com/en/paper/2210.15889v5/

## Key points

- Neural-symbolic computing (NeSy) bridges the gap between symbolic reasoning and statistical learning in Artificial Intelligence.
- NeSy combines interpretability of symbolic representation with robust learning capabilities of neural networks.
- The paper "Towards Data-and Knowledge-Driven Artificial Intelligence: A Survey on Neuro-Symbolic Computing" by Wang, Yang, and Wu provides a comprehensive overview of recent developments in NeSy research.
- Key driving factors behind the development of NeSy are discussed, shedding light on researchers' motivations in this area.
- Recent landmark approaches in NeSy research are categorized based on neural-symbolic integration, knowledge representation, knowledge embedding, and functionality.
- Successful applications of modern NeSy approaches across various domains are highlighted to showcase practical impact.
- Benchmarking results for several NeSy methods on three representative application tasks demonstrate their effectiveness in real-world scenarios.
- Open challenges and potential future research directions in Neuro-Symbolic Computing are identified to guide new researchers entering the field.

## Summary

, , , , 

Neural-symbolic computing (NeSy) is a rapidly growing field within Artificial Intelligence that aims to bridge the gap between symbolic reasoning and statistical learning. This integration offers the potential to combine the interpretability of symbolic representation with the robust learning capabilities of neural networks, paving the way for advancements in AI. In their paper titled "Towards Data-and Knowledge-Driven Artificial Intelligence: A Survey on Neuro-Symbolic Computing," authors Wenguan Wang, Yi Yang, and Fei Wu provide a comprehensive overview of recent developments in NeSy research. The survey begins by tracing the history of NeSy, highlighting early work and foundational concepts that have shaped the field. The authors delve into key driving factors behind the development of NeSy, shedding light on the motivations driving researchers in this area. They categorize recent landmark approaches based on several main characteristics central to NeSy research, including neural-symbolic integration, knowledge representation, knowledge embedding, and functionality. Furthermore, the authors discuss successful applications of modern NeSy approaches across various domains, showcasing the practical impact of this research paradigm. They also present benchmarking results for several NeSy methods on three representative application tasks, demonstrating their effectiveness in real-world scenarios. Finally, the survey identifies open challenges and outlines potential future research directions to guide new researchers entering this rapidly evolving field. Overall, this detailed summary provides valuable insights into the state-of-the-art in Neuro-Symbolic Computing and underscores its role in advancing data-and knowledge-driven AI. By synthesizing recent developments and important contributions in NeSy research, Wang et al. 's survey serves as a roadmap for researchers looking to contribute to this exciting area at the intersection of symbolic reasoning and statistical learning.