Vihang Patil
Papers
1
Total Citations
2
H-Index
1
About
Vihang Patil is a researcher at the forefront of Reinforcement Learning (RL) and robotics, with a focus on developing large-scale action models for real-world deployment. His most notable contribution is the introduction of the **Large Recurrent Action Model (LRAM)**, which leverages the xLSTM architecture to enable fast, efficient inference for robotics tasks. This work directly addresses a critical bottleneck in the field: while Transformer-based agents are powerful, their slow inference speeds limit their use in real-time applications. Patil’s approach offers a compelling alternative, achieving competitive performance with significantly faster decision-making. Although his seminal paper, published in 2024, is early in its citation lifecycle, it represents a forward-looking shift toward practical, deployable RL systems. Patil’s research sits at the intersection of sequence modeling, offline RL, and embodied AI, aiming to bridge the gap between high-performance agents and real-world latency constraints. His work is particularly relevant for students and researchers interested in efficient architectures for robotics, real-time control, and the next generation of action models beyond Transformers.
Research Focus
Key Achievements
Top Papers
- 1