SymbolicAI: A framework for logic-based approaches combining generative models and solvers

Marius-Constantin Dinu, Claudiu Leoveanu-Condrei, Markus Holzleitner, Werner Zellinger, Sepp Hochreiter
Proceedings of The 3rd Conference on Lifelong Learning Agents, PMLR 274:869-914, 2025.

Abstract

We introduce _SymbolicAI_, a versatile and modular framework employing a logic-based approach to concept learning and flow management in generative processes. SymbolicAI enables the seamless integration of generative models with a diverse range of solvers by treating large language models (LLMs) as semantic parsers that execute tasks based on both natural and formal language instructions, thus bridging the gap between symbolic reasoning and generative AI. We leverage probabilistic programming principles to tackle complex tasks, and utilize differentiable and classical programming paradigms with their respective strengths. The framework introduces a set of polymorphic, compositional, and self-referential operations for multi-modal data that connects multi-step generative processes and aligns their outputs with user objectives in complex workflows. As a result, we can transition between the capabilities of various foundation models with in-context learning capabilities and specialized, fine-tuned models or solvers proficient in addressing specific problems. Through these operations based on in-context learning our framework enables the creation and evaluation of explainable computational graphs. Finally, we introduce a quality measure and its empirical score for evaluating these computational graphs, and propose a benchmark that compares various state-of-the-art LLMs across a set of complex workflows. We refer to the empirical score as the “Vector Embedding for Relational Trajectory Evaluation through Cross-similarity”, or _VERTEX_ score for short. See [framework codebase](https://github.com/ExtensityAI/symbolicai) and [benchmark](https://github.com/ExtensityAI/benchmark) for more information.

Cite this Paper

BibTeX
@InProceedings{pmlr-v274-dinu25a, title = {SymbolicAI: A framework for logic-based approaches combining generative models and solvers}, author = {Dinu, Marius-Constantin and Leoveanu-Condrei, Claudiu and Holzleitner, Markus and Zellinger, Werner and Hochreiter, Sepp}, booktitle = {Proceedings of The 3rd Conference on Lifelong Learning Agents}, pages = {869--914}, year = {2025}, editor = {Lomonaco, Vincenzo and Melacci, Stefano and Tuytelaars, Tinne and Chandar, Sarath and Pascanu, Razvan}, volume = {274}, series = {Proceedings of Machine Learning Research}, month = {29 Jul--01 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v274/main/assets/dinu25a/dinu25a.pdf}, url = {https://proceedings.mlr.press/v274/dinu25a.html}, abstract = {We introduce _SymbolicAI_, a versatile and modular framework employing a logic-based approach to concept learning and flow management in generative processes. SymbolicAI enables the seamless integration of generative models with a diverse range of solvers by treating large language models (LLMs) as semantic parsers that execute tasks based on both natural and formal language instructions, thus bridging the gap between symbolic reasoning and generative AI. We leverage probabilistic programming principles to tackle complex tasks, and utilize differentiable and classical programming paradigms with their respective strengths. The framework introduces a set of polymorphic, compositional, and self-referential operations for multi-modal data that connects multi-step generative processes and aligns their outputs with user objectives in complex workflows. As a result, we can transition between the capabilities of various foundation models with in-context learning capabilities and specialized, fine-tuned models or solvers proficient in addressing specific problems. Through these operations based on in-context learning our framework enables the creation and evaluation of explainable computational graphs. Finally, we introduce a quality measure and its empirical score for evaluating these computational graphs, and propose a benchmark that compares various state-of-the-art LLMs across a set of complex workflows. We refer to the empirical score as the “Vector Embedding for Relational Trajectory Evaluation through Cross-similarity”, or _VERTEX_ score for short. See [framework codebase](https://github.com/ExtensityAI/symbolicai) and [benchmark](https://github.com/ExtensityAI/benchmark) for more information.} }
Endnote
%0 Conference Paper %T SymbolicAI: A framework for logic-based approaches combining generative models and solvers %A Marius-Constantin Dinu %A Claudiu Leoveanu-Condrei %A Markus Holzleitner %A Werner Zellinger %A Sepp Hochreiter %B Proceedings of The 3rd Conference on Lifelong Learning Agents %C Proceedings of Machine Learning Research %D 2025 %E Vincenzo Lomonaco %E Stefano Melacci %E Tinne Tuytelaars %E Sarath Chandar %E Razvan Pascanu %F pmlr-v274-dinu25a %I PMLR %P 869--914 %U https://proceedings.mlr.press/v274/dinu25a.html %V 274 %X We introduce _SymbolicAI_, a versatile and modular framework employing a logic-based approach to concept learning and flow management in generative processes. SymbolicAI enables the seamless integration of generative models with a diverse range of solvers by treating large language models (LLMs) as semantic parsers that execute tasks based on both natural and formal language instructions, thus bridging the gap between symbolic reasoning and generative AI. We leverage probabilistic programming principles to tackle complex tasks, and utilize differentiable and classical programming paradigms with their respective strengths. The framework introduces a set of polymorphic, compositional, and self-referential operations for multi-modal data that connects multi-step generative processes and aligns their outputs with user objectives in complex workflows. As a result, we can transition between the capabilities of various foundation models with in-context learning capabilities and specialized, fine-tuned models or solvers proficient in addressing specific problems. Through these operations based on in-context learning our framework enables the creation and evaluation of explainable computational graphs. Finally, we introduce a quality measure and its empirical score for evaluating these computational graphs, and propose a benchmark that compares various state-of-the-art LLMs across a set of complex workflows. We refer to the empirical score as the “Vector Embedding for Relational Trajectory Evaluation through Cross-similarity”, or _VERTEX_ score for short. See [framework codebase](https://github.com/ExtensityAI/symbolicai) and [benchmark](https://github.com/ExtensityAI/benchmark) for more information.
APA
Dinu, M., Leoveanu-Condrei, C., Holzleitner, M., Zellinger, W. & Hochreiter, S.. (2025). SymbolicAI: A framework for logic-based approaches combining generative models and solvers. Proceedings of The 3rd Conference on Lifelong Learning Agents, in Proceedings of Machine Learning Research 274:869-914 Available from https://proceedings.mlr.press/v274/dinu25a.html.

Related Material