Noelia Ferruz receives €1.5 million to develop AI with reasoning skills

Noelia Ferruz, a chemist, has received €1.5 million from the European Union to develop an AI system with advanced reasoning capabilities. This project aims to create a digital mind that can explain its reasoning, a significant step towards understanding AI's decision-making processes. Ferruz will collaborate with a team of bioinformaticians and molecular biologists to train the AI and test the new molecules it designs. The AI's ability to explore uncharted territories of protein design could lead to groundbreaking discoveries in biochemistry. The project addresses the current limitations of AI, which often operates as a 'black box,' making it difficult to understand how it reaches conclusions. Ferruz emphasizes the importance of explanatory AI, which seeks to clarify the reasoning behind AI-generated outcomes. By analyzing the activation of artificial neurons in the AI, researchers hope to gain insights into its learning processes and improve its ability to generate protein sequences. The challenge lies not only in creating this advanced AI but also in comprehending its inner workings. Current models, like ChatGPT, lack the ability to articulate their reasoning, which poses a barrier to their effective application in scientific research. Ferruz's initiative represents a pioneering effort to bridge this gap, potentially leading to a more human-like intelligence in AI systems. This funding from the European Research Council (ERC) highlights the project's significance, as it is one of only 494 selected for support. The implications of Ferruz's work could extend beyond protein design, influencing various fields that rely on AI for complex problem-solving.