A groundbreaking study by Prof. Edward Szczerbicki and his team | Faculty of Management and Economics at the Gdańsk University of Technology

The renowned "International Journal of Approximate Reasoning" is a peer-reviewed scientific journal dedicated to artificial intelligence and methods for handling uncertainty and imprecision in data. It is highly rated on the list of the Ministry of Science and Higher Education (140 points).

The article is of great scientific significance: it introduces a novel and groundbreaking approach in the field of knowledge management and engineering, enabling Large Language Models (LLMs) to achieve "machine intelligence." LLMs already demonstrate remarkable proficiency in semantic understanding; however, a key barrier on the path to their excellence remains achieving true machine intelligence: counterfactual reasoning, that is, the ability to imagine "what if" scenarios.

As Prof. Edward Szczerbicki explains:

"In the article, we propose and rigorously verify a neurosymbolic framework called SACR (Structure-Aware Causal Reasoning), which integrates d-separation logic and an extension of Markov boundaries in the interpretation of causal graphs modeling LLM-based reasoning. SACR significantly outperforms state-of-the-art models in entangled logical scenarios, achieving higher accuracy in multi-path settings, including even 'what if' type scenarios."

We warmly congratulate Prof. Edward Szczerbicki and his Team on this achievement. We invite you to read the article: it is available HERE.