Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks

Jie-Jing Shao, Haiyan Yin, Yueming Lyu, Xingrui Yu, Lan-Zhe Guo, Ivor Tsang, James Kwok, Yu-Feng Li
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:110314-110337, 2026.

Abstract

Foundation model-driven agents often struggle with long-horizon planning due to the transient nature of purely prompting-based reasoning. While existing skill induction methods mitigate this by distilling experience into state-blind parameterized scripts, they fail to capture the conditional logic required for robust execution in dynamic environments. In this paper, we propose Neuro-Symbolic Skill Induction (NSI), a framework that lifts interaction traces into modular, logic-grounded programs. By synthesizing explicit control flows and dynamic variable binding, NSI empowers agents to discover when and why to act. This paradigm enables the efficient generalization, allowing agents to induce skills from few-shot examples and flexibly adapt to unseen goals. Experiments on a series of agentic tasks demonstrate that NSI consistently outperforms state-of-the-art baselines, empowering agents to self-evolve into architects of logic-grounded skills. Project Page: https://sh-jj.github.io/NSI.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-shao26i, title = {Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks}, author = {Shao, Jie-Jing and Yin, Haiyan and Lyu, Yueming and Yu, Xingrui and Guo, Lan-Zhe and Tsang, Ivor and Kwok, James and Li, Yu-Feng}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {110314--110337}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/shao26i/shao26i.pdf}, url = {https://proceedings.mlr.press/v306/shao26i.html}, abstract = {Foundation model-driven agents often struggle with long-horizon planning due to the transient nature of purely prompting-based reasoning. While existing skill induction methods mitigate this by distilling experience into state-blind parameterized scripts, they fail to capture the conditional logic required for robust execution in dynamic environments. In this paper, we propose Neuro-Symbolic Skill Induction (NSI), a framework that lifts interaction traces into modular, logic-grounded programs. By synthesizing explicit control flows and dynamic variable binding, NSI empowers agents to discover when and why to act. This paradigm enables the efficient generalization, allowing agents to induce skills from few-shot examples and flexibly adapt to unseen goals. Experiments on a series of agentic tasks demonstrate that NSI consistently outperforms state-of-the-art baselines, empowering agents to self-evolve into architects of logic-grounded skills. Project Page: https://sh-jj.github.io/NSI.} }
Endnote
%0 Conference Paper %T Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks %A Jie-Jing Shao %A Haiyan Yin %A Yueming Lyu %A Xingrui Yu %A Lan-Zhe Guo %A Ivor Tsang %A James Kwok %A Yu-Feng Li %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-shao26i %I PMLR %P 110314--110337 %U https://proceedings.mlr.press/v306/shao26i.html %V 306 %X Foundation model-driven agents often struggle with long-horizon planning due to the transient nature of purely prompting-based reasoning. While existing skill induction methods mitigate this by distilling experience into state-blind parameterized scripts, they fail to capture the conditional logic required for robust execution in dynamic environments. In this paper, we propose Neuro-Symbolic Skill Induction (NSI), a framework that lifts interaction traces into modular, logic-grounded programs. By synthesizing explicit control flows and dynamic variable binding, NSI empowers agents to discover when and why to act. This paradigm enables the efficient generalization, allowing agents to induce skills from few-shot examples and flexibly adapt to unseen goals. Experiments on a series of agentic tasks demonstrate that NSI consistently outperforms state-of-the-art baselines, empowering agents to self-evolve into architects of logic-grounded skills. Project Page: https://sh-jj.github.io/NSI.
APA
Shao, J., Yin, H., Lyu, Y., Yu, X., Guo, L., Tsang, I., Kwok, J. & Li, Y.. (2026). Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:110314-110337 Available from https://proceedings.mlr.press/v306/shao26i.html.

Related Material