MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202641081113 A) filed by New Horizon College Of Engineering on July 01, 2026, for Fault Tolerant Hardware Learning Based Distributed System.
Inventors include Revatty Krishnamurthy; Anisha Anandkumar Ukarande; J. Harshaa Vardhini; Jomet Elise George; and Akshaya M.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: The present invention discloses a volunteer computing infrastructure that incorporates a Reinforcement learning scheduler that can be used to train machine learning models across volatile, heterogeneous volunteer nodes. Traditional volunteer computing suffer from training stalls due to high node dropout and heterogeneous hardware. This architecture employs a proximal policy optimization agent to dynamically optimize job node assignments. During training jobs, nodes are continuously penalized and monitored for dropout risks and anomalies. Combined with P2P federated aggregation and adaptive checkpointing the system transforms unpredictable hardware into resilient, low cost training network. The system further incorporates heterogeneity-aware workload partitioning and reinforcement learning-controlled checkpoint placement to improve fault tolerance and training continuity in volatile volunteer computing environments.
Disclaimer: Curated by HT Syndication.
