Papers
1
Total Citations
7
H-Index
1
About
Michael Poli is a researcher at the forefront of machine learning and dynamical systems, with a primary focus on developing novel architectures for sequence modeling and scientific computing. His major contributions center on advancing state-space models (SSMs) as efficient alternatives to transformers, particularly through his work on the Mamba architecture, which has redefined long-context processing in deep learning. Poli’s research also explores neural hybrid automata for learning dynamics with multiple modes and stochastic transitions, bridging continuous-time control and discrete event-triggered processes. His highly influential work has garnered thousands of citations, reflecting its profound impact on both theoretical understanding and practical applications in AI. Notably, his papers on structured state-space sequences (S4) and Mamba have become foundational in the field, enabling breakthroughs in areas ranging from natural language processing to scientific modeling. Poli’s ability to integrate insights from control theory, differential equations, and deep learning positions him as a leading innovator, with his contributions shaping the next generation of efficient, scalable neural architectures.
Research Focus
Key Achievements
Top Papers
- 1