Patents
Inventions filed at IBM, grouped by invention rather than by filing — one entry may have issued in several jurisdictions. Filed under my full legal name, Jayaram Kallapalayam Radhakrishnan.
10 granted inventions (18 issued patents across jurisdictions), 8 pending applications. Years shown are the disclosure year. Where a PDF link is shown, it points to the full document on USPTO. Grants without one had no public PDF that could be verified.
Granted
- Federated Learning with Partitioned and Dynamically-Shuffled Model Updates CN ZL202210537839.3·JP 7740842
- Trusted and Decentralized Aggregation for Federated Learning US 12541708 PDF·CN ZL202280032564.6·JP 7841826
- Data Subset Selection for Federated Learning US 12632795
- Secure Federation of Distributed Stochastic Gradient Descent CN ZL202080079660.7·GB 2606867·JP 7513359
- Hyperparameter Determination for a Differentially Private Federated Learning Process US 11941520 PDF
- Application Aware Resource Allocation for Deep Learning Job Scheduling US 11586475 PDF
- Determining Optimal Compute Resources for Distributed Batch Based Optimization Applications US 11263052 PDF·CN ZL202010724291.4
- Scalable Multi-Framework Multi-Tenant Lifecycle Management of Deep Learning Applications US 11269728 PDF·US 11196547 PDF
- Graphics Processing Unit Resource Sharing US 9830677 PDF·US 9830678 PDF
- Application Deployment and Monitoring in a Cloud Environment to Satisfy Integrity and Geo-Fencing Constraints US 10228924 PDF
Pending applications
- Extracting and Generalizing Subtask-Level Knowledge for Autonomous Agents
- Trajectory-Based Memory Generation and Contextual Retrieval
- Storage and Memory Management for Multimodal AI Systems
- Deployment-Aware Memory Packaging for Artificial Intelligence Agents
- Conversational Workflow Building using Large Language Models
- In-Context Example Ordering for Large Language Models
- Using Secure and Reliable Benefit-Analysis Matchmaking to Select Federated Learning Candidates
- Grouped Aggregation in Federated Learning
Other disclosures
- Purpose-Directed Gisting for Agentic Memoryfiled
- Federated Learning Participant Selection through Label Distribution Clusteringfiled
- Method and System for In-Field Malfunctioning Detection of AI Systemsfiled
- Serverless Federated Learning: Unifying Cross-Device and Cross-Silo Learningdefensive publication