Partha Maji
Papers
1
Total Citations
3
H-Index
1
About
Partha Maji is a researcher whose work sits at the intersection of Bayesian deep learning and model efficiency, with a particular focus on the practical deployment of uncertainty-aware systems. His key research area addresses a critical tension: how to maintain reliable uncertainty estimates when neural networks are compressed through quantisation for real-world, resource-constrained environments. In his most cited work, "On the Effects of Quantisation on Model Uncertainty in Bayesian Neural Networks" (2021, 3 citations), Maji systematically investigates how quantisation degrades the calibrated uncertainty that makes Bayesian neural networks (BNNs) so valuable for high-stakes decision-making. This contribution is significant because it bridges the gap between theoretical Bayesian methods and practical edge deployment, revealing that naive quantisation can dangerously distort a model’s over- or under-confidence. By identifying these failure modes, Maji’s research provides essential guidance for engineers building trustworthy AI systems in fields like autonomous driving or medical diagnosis, where knowing what a model does not know is as important as its predictions. His work marks an important step toward making principled uncertainty quantification computationally feasible outside of research labs.
Research Focus
Key Achievements
Top Papers
- 1