Kowndinya Boyalakuntla

Rutgers, The State University of New Jersey

Papers

Total Citations

H-Index

About

Kowndinya Boyalakuntla is a researcher pushing the boundaries of robotic manipulation through the integration of language understanding, semantic reasoning, and generative AI. His primary research areas center on semantic object rearrangement, affordance prediction, and task planning for autonomous systems. In his highly innovative work, "LGMCTS: Language-Guided Monte-Carlo Tree Search for Executable Semantic Object Rearrangement" (2023, 3 citations), Boyalakuntla introduced a novel framework that enables robots to generate actionable rearrangement plans directly from natural language descriptions. This approach overcomes the limitations of prior methods like StructFormer by combining language understanding with Monte-Carlo Tree Search to produce executable, semantically coherent object layouts. Expanding into fine-grained manipulation, his work "DAP: Diffusion-based Affordance Prediction for Multi-modality Storage" (2024, 1 citation) tackles the challenging problem of precise 6D object placement in storage tasks. By leveraging diffusion models for affordance prediction, this work addresses the critical need for accurate orientation and positioning in constrained environments. Boyalakuntla’s contributions are particularly notable for bridging high-level semantic reasoning with low-level robotic control, offering practical pathways toward more intelligent and adaptable autonomous systems.