HYDRA 2026

Welcome to the 5th International Workshop on HYbrid models for coupling Deductive and inductive ReAsoning (HYDRA).
The workshop is co-located with the 18th International Conference on Logic Programming and Non-monotonic Reasoning (LPNMR 2026).

Scope of the workshop

Deductive reasoning has been widely used in various fields such as planning, scheduling, robotics, and other industrial applications due to its ability to derive specific conclusions based on valid evidence or facts. However, deductive reasoning requires consistent premises and a proper knowledge base to make accurate inferences. On the other hand, inductive reasoning, such as Machine and Deep Learning, extracts generalized conclusions from limited and specific observations. While these approaches are promising in recognizing meaningful patterns from large amounts of data, they lack means for interpreting the model's choices and incorporating prior knowledge. Therefore, combining and intertwining deductive and inductive methods can offer a more comprehensive approach to Artificial Intelligence, leveraging the strengths of both methods.

Motivations and Main Goals

The HYDRA workshop aims to bridge the gap between deductive and inductive reasoning, two powerful yet fundamentally distinct paradigms in Artificial Intelligence (AI). Deductive reasoning relies on explicit premises and logical inference rules to derive specific conclusions, whereas inductive reasoning generalizes from observations, often leveraging Machine Learning and Deep Learning techniques. The integration of these approaches offers the potential to develop more robust and flexible AI systems capable of reasoning effectively across diverse contexts. However, neither deductive nor inductive methods alone can be considered fully comprehensive solutions to AI challenges. Investigating how these paradigms can be effectively combined is therefore essential for developing novel solutions that exploit the strengths and address the limitations of both.

Within HYDRA, we invite submissions presenting original research on all aspects of hybrid deductive–inductive reasoning, including theoretical frameworks, practical applications, and experimental studies. We are particularly interested in contributions addressing key challenges such as the integration of logical and statistical models, the design of algorithms capable of reasoning under incomplete or uncertain knowledge, and the development of tools for explaining and interpreting hybrid models. We also encourage research exploring the ethical and societal implications of these technologies, including issues of fairness, accountability, and transparency.

The HYDRA workshop aims to bring together researchers from the scientific community and welcomes both theoretical and applied contributions on frameworks, methods, and applications for integrating deductive and inductive systems across a variety of scenarios. Submissions may include full papers, summaries of recently published work, as well as work-in-progress contributions. Building on the success of its previous editions, HYDRA welcomes new contributions and perspectives.

Keywords

• Logic programming

• Deductive Reasoning

• Inductive Reasoning

• Abductive Reasoning

• Hybrid reasoning models

• Deep Learning

• Machine Learning

• Neuro-Symbolic methods