Publications

NSP: A Neuro-Symbolic Natural Language Navigational Planner

Published in ICMLA, 2024

NSP uses a feedback loop from the symbolic execution environment to the neural generation process to self-correct syntax errors and satisfy execution time constraints. We evaluate our neuro-symbolic approach using a benchmark suite with 1500 path-planning problems. The experimental evaluation shows that our neuro-symbolic approach produces 90.1% valid paths that are on average 19-77% shorter than state-of-the-art neural approaches.

Recommended citation: William English, Dominic Simon, Sumit Jha, and Rickard Ewetz, “NSP: A Neuro-Symbolic Natural Language Navigational Planner”, International Conference on Machine Learning and Applications (ICMLA), 2024. https://arxiv.org/abs/2409.06859

Neuro-Symbolic Program Synthesis for Multi-Hop Natural Language Navigation

Published in ICAA, 2024

The NeSy Program Synthesis approach is evaluated using 600 multi- hop navigation tasks with 1 to 10 hops. Compared with neural approaches, the our approach improves the success rate and path efficiency by an average of 64.3% and 19.4% across all tasks, respectively.

Recommended citation: W. English, D. Simon, M. R. Ahmed, S. K. Jha, and R. Ewetz, “Neuro-Symbolic Program Synthesis for Multi-Hop Natural Language Navigation”, International Conference on Assured Autonomy (ICAA), 2024.