Vijay K Madisetti
Biographic Data
| ID | 10054195 |
|---|---|
| NAME | Vijay K Madisetti |
| GIVEN NAMES | Vijay K |
| FAMILY NAME | Madisetti |
| SIGNATURE | MADISETTI V K |
| AFFILIATIONS | Georgia Institute of Technology |
| ORCID | 0000-0002-6539-6769 |
| VERIFIED | Yes |
| TOTAL WORKS | 3 |
| TOTAL CITATIONS | 0 |
| WORKS AS AUTHOR | 3 |
| WORKS AS EDITOR | 0 |
| FIRST PUBLICATION YEAR | 1997 |
| LATEST PUBLICATION YEAR | 2026 |
| H-INDEX | 0 |
Kair
We present knowledge-aware iterative retrieval (KAIR), an innovative framework for LLM agents driven by large language models (LLMs) that iteratively refines user queries and filters contextual evidence using dynamically evolving knowledge to facilitate better systems. A key feature of this system is its separation of external information sources from an internal knowledge cache, which is continuously updated to guide query formulation and eviden…
Cosmic
We present COSMIC, a leader-driven multiagent coordination framework that extends the AutoGen GroupChat paradigm to support more stable and context-consistent collaboration among large language model (LLM) agents. COSMIC introduces: 1) a leader agent that decomposes tasks and issues an explicit termination decision; 2) a shared rolling memory buffer that synchronizes recent context across agents; and 3) a speaker selector agent (SSA) that enforce…
A case study in the development of multi-media educational material
While industry and academia have been aware of the need for an intensive study of embedded digital system design, resource constraints, fuzzy objectives, and short-time horizon have handicapped progress. The $150M Rapid Prototyping of Application Specific Signal Processors (RASSP) program is a major DARPA/Tri-Service effort designed to overcome these limitations. This effort has developed a number of new technologies that will lead to dramaticall…
No prominent works on this page.
A case study in the development of multi-media educational material
While industry and academia have been aware of the need for an intensive study of embedded digital system design, resource constraints, fuzzy objectives, and short-time horizon have handicapped progress. The $150M Rapid Prototyping of Application Specific Signal Processors (RASSP) program is a major DARPA/Tri-Service effort designed to overcome these limitations. This effort has developed a number of new technologies that will lead to dramaticall…
Kair
We present knowledge-aware iterative retrieval (KAIR), an innovative framework for LLM agents driven by large language models (LLMs) that iteratively refines user queries and filters contextual evidence using dynamically evolving knowledge to facilitate better systems. A key feature of this system is its separation of external information sources from an internal knowledge cache, which is continuously updated to guide query formulation and eviden…
Cosmic
We present COSMIC, a leader-driven multiagent coordination framework that extends the AutoGen GroupChat paradigm to support more stable and context-consistent collaboration among large language model (LLM) agents. COSMIC introduces: 1) a leader agent that decomposes tasks and issues an explicit termination decision; 2) a shared rolling memory buffer that synchronizes recent context across agents; and 3) a speaker selector agent (SSA) that enforce…
Multi-agent system (2 works) · Advancements in PLL and VCO Technologies (1 work) · AI-based Problem Solving and Planning (1 work) · Analog and Mixed-Signal Circuit Design (1 work) · Compiler (1 work) · Computer Science (1 work) · COSMIC cancer database (1 work) · Curriculum (1 work) · Embedded system (1 work) · Engineering (1 work)