MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202611113302 A) filed by Ajay Kumar Garg Engineering College on September 22, 2026, for A Decentralized Multi-Agent System And Method For Autonomous Machine Learning Pipeline Automation.
Inventors include Sachin Parashar; Sachi Jain; Anmol Sharma; Areeb Khan; and Dr. Sachin Jain.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: Disclosed is a decentralized multi-agent system (100) and method for autonomous machine learning pipeline automation. The system comprises a data ingestion interface (102), a data quality profiling agent (124), a decentralized multi-agent engine (104) having a pool of autonomous agents (105) that dynamically self-allocate tasks via a competitive market bidding module (106) and a dynamic role allocation module (108), a three-layer shared memory canvas (110) comprising a Global Blackboard (112), Agent Workspace (114) and Consensus Ledger (116), a sandbox execution environment (118), a self-healing reaper module (120) for automatic fault recovery, a peer-validation module (122), an ensemble selection module (126), a feedback loop module (128), and a deployment packaging module (130). The components are synergistically interconnected, as illustrated in FIG. 1, to eliminate central points of failure, enable fluid role adaptation, and produce peer-validated, resilient, production-ready, containerized machine learning pipelines without human intervention. (Reference figure for publication: FIG. 1)
Disclaimer: Curated by HT Syndication.
