Abhiram Iyer

Carnegie Mellon University, Moscow Institute of Thermal Technology

Papers

Total Citations

H-Index

About

Abhiram Iyer is a rising star in artificial intelligence and robotics, whose work tackles some of the field’s most fundamental challenges: building systems that can learn continuously and operate reliably in dynamic, high-dimensional environments. His research sits at the intersection of embodied AI, continual learning, and probabilistic state estimation. Iyer’s most influential contribution to date is his 2022 paper, *Avoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments* (54 citations). In this work, he introduced a biologically-inspired neural architecture that uses active dendrites to prevent catastrophic forgetting, enabling an agent to seamlessly adapt to shifting task contexts without overwriting prior knowledge—a breakthrough for real-world robotic deployment. More recently, in *Resampling-free Particle Filters in High-dimensions* (2024), Iyer has taken on a classic robotics bottleneck: the curse of dimensionality in particle filtering. By proposing a resampling-free approach, his work promises more efficient and robust state estimation for complex robotic systems. With a growing citation footprint and a focus on foundational problems, Iyer is establishing himself as a key voice in the next generation of adaptive, intelligent machines.