Ajitesh Srivastava
Papers
4
Total Citations
37
H-Index
4
About
Ajitesh Srivastava is a researcher at the intersection of efficient AI hardware and human-robot interaction, making notable contributions to both reinforcement learning acceleration and socially aware robotics. His work on QTAccel, a generic FPGA-based design for Q-Table reinforcement learning, addresses the computational bottlenecks of traditional neural network approaches by providing a hardware accelerator that significantly outperforms them for tractable state spaces. This work, which has garnered citations for its practical impact, demonstrates his ability to bridge algorithm design with real-world deployment constraints. Beyond hardware, Srivastava has ventured into the novel domain of robotic humor, developing machine learning pipelines for laughter classification and joke success assessment. His studies on adaptive robotic comedians, informed by surveys of human comedians, aim to equip social robots with the ability to read and respond to human reactions during playful dialog. By combining efficient AI systems with emotionally intelligent interaction, Srivastava’s research pushes the boundaries of how machines can learn, adapt, and engage with people in both computational and social contexts.
Research Focus
Key Achievements
Top Papers
- 1
- 2QTAccel8 citations · 2020
- 3
- 4A Robot Walks into a Bar: Automatic Robot Joke Success Assessment4 citations · 2021