MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641085242 A) filed by Cmr Technical Campus; Cmr College Of Engineering & Technology; and Cmr Institute Of Technology on July 11, 2026, for Hierarchical Multi-Agent Learning Framework With Emergent Task Decomposition Via Latent Intent Encoding And Cross-Agent Policy Distillation.
Inventors include G. Aravind; V. Tejaswini; Mr. Y. Mahendra Reddy; Mr. D. Ranadeep Reddy; Mr. D. Jyothi; and Gorige Sangeetha.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present invention discloses a system (100) and method for hierarchical multi-agent learning with emergent task decomposition via latent intent encoding and cross-agent policy distillation. The system (100) comprises a memory (102), a processor (104), and a communication module (106), wherein the processor (104) executes a latent intent encoding module (108), a task decomposition module (110), an agent coordination module (112), and a cross-agent policy distillation module (114). The latent intent encoding module (108) generates latent intent representations from task objectives and contextual information. The task decomposition module (110) dynamically partitions objectives into hierarchical sub-tasks for allocation among multiple autonomous agents. The agent coordination module (112) facilitates collaborative execution and communication, while the cross-agent policy distillation module (114) transfers learned behavioral policies between agents to improve learning efficiency and convergence. The disclosed framework enables adaptive task management, scalable coordination, and optimized distributed learning across complex computational environments.
Disclaimer: Curated by HT Syndication.
