Mitsubishi Electric Corporation has introduced the manufacturing industry’s first multi-agent AI technology, utilizing an argumentation framework to automate complex decision-making. This innovative system generates adversarial debates among expert AI agents, ensuring rapid and expert-level decision-making with transparent reasoning. Designed under the Maisart® AI program, the technology addresses the growing complexity of decision-making in areas like security risk assessment and production planning, which often relies on individual expertise and long consensus-building processes.
Challenges such as opacity in AI reasoning have hindered adoption in critical sectors. By incorporating adversarial generation from Generative Adversarial Networks (GANs), Mitsubishi Electric’s technology allows AI agents to compete in deriving improved conclusions, facilitating evidence-based decision-making. This solution not only enhances operational efficiency but also supports specialized tasks in security analysis and risk management. For more details, visit www.MitsubishiElectric.com/news/.
For inquiries, contact Mitsubishi Electric’s Information Technology R&D Center.
Source link