Ademi Adeniji

Papers

Total Citations

H-Index

About

Ademi Adeniji is a researcher whose work sits at the intersection of reinforcement learning (RL), robotics, and biologically inspired control systems. Their primary contributions address fundamental challenges in training agents for complex, real-world tasks, particularly the problem of sparse rewards. In their 2023 paper "Language Reward Modulation for Pretraining Reinforcement Learning" (4 citations), Adeniji critically examines the limitations of learned reward functions (LRFs) and proposes a novel framework for using language-derived signals to guide pretraining, enabling more efficient learning in environments where task rewards are scarce. Earlier, in their 2019 work "Recurrent Control Nets as Central Pattern Generators for Deep Reinforcement Learning" (3 citations), Adeniji introduced a groundbreaking approach by modeling recurrent neural networks after Central Pattern Generators (CPGs)—biological circuits that produce rhythmic motion. This work directly bridges neuroscience and deep RL, offering a robust method for generating coordinated locomotion in robotic agents without explicit rhythmic input. While their citation counts reflect an emerging career, Adeniji’s research is notable for its creative synthesis of language models, biological principles, and RL, positioning them as an innovative voice in advancing autonomous learning systems.