Ankur Mali

About

Ankur Mali is at the forefront of a paradigm shift in artificial intelligence, pioneering brain-inspired computational models that move beyond traditional deep learning. His research centers on developing biologically plausible learning algorithms, with a particular focus on predictive coding and neural generative coding (NGC) as alternatives to error backpropagation. Mali’s major contributions include the introduction of "Active Predictive Coding," a backpropagation-free framework for reinforcement learning that enables agents to learn from sparse rewards in robotic control tasks—a significant step toward more efficient, adaptable AI systems. His work, such as the 2023 paper "Brain-inspired Computational Intelligence via Predictive Coding" (12 citations), challenges the status quo by designing agents built entirely from predictive processing circuits, facilitating dynamic, online learning. With over 20 citations across his most-cited works, Mali’s research is gaining traction for its potential to bridge neuroscience and AI, offering a path toward more robust, energy-efficient intelligence. His achievements highlight a commitment to rethinking foundational learning mechanisms, making him a key voice in the quest for truly intelligent machines.

Research Focus

Key Achievements

Top Papers

1. 1
   Brain-inspired Computational Intelligence via Predictive Coding

   12 citations · 2023
2. 2
   Active Predictive Coding: Brain-Inspired Reinforcement Learning for Sparse Reward Robotic Control Problems

   7 citations · 2023
3. 3
   Active Predicting Coding: Brain-Inspired Reinforcement Learning for Sparse Reward Robotic Control Problems

   2 citations · 2022

Key Collaborators

Contact & Links