We are excited to present our newest work, combining the power of a symbolic approach and Large Language Models (LLMs). Our Symbolic API presents a novel approach to bridge the gap between classical programming (Software 1.0) and differentiable programming (Software 2.0).

Conceptually, our framework uses machine learning – and specifically LLMs – at its core, and curates’ operations based on dedicated `zero` or `few-shot` learning prompt designs.

Our main philosophy is to divide and conquer a complex problem into more manageable, smaller problems. Therefore, each operation solves an atomic task, however, by chaining these operations back together we can solve more complex problems.

In this turn, we also demonstrate how to combine the strengths of both neural networks and symbolic reasoning to create AI systems that can solve a wide range of hard cognitive tasks.

This includes fact-based generation of text, flow control of a generative process towards a desired outcome, and interpretability within generative processes. Read more at our GitHub Repository or try it out directly our PyPI package.