Edoardo Conti

Papers

Total Citations

264

H-Index

About

Edoardo Conti is a researcher specializing in deep reinforcement learning and evolutionary computation, with a particular focus on improving the efficiency and effectiveness of black-box optimization methods for training deep neural networks. His most notable contribution lies at the intersection of evolution strategies and novelty-based exploration, where he has helped advance the understanding of how population-based approaches can compete with — and in some ways surpass — traditional reinforcement learning methods like Q-learning and policy gradients. Conti's seminal work, "Improving Exploration in Evolution Strategies for Deep Reinforcement Learning via a Population of Novelty-Seeking Agents," has accumulated over 260 citations across its 2017 and 2018 versions, demonstrating substantial influence within the research community. A core insight of this research is that evolution strategies can train neural networks significantly faster than conventional RL approaches — reducing training time from days to hours — by exploiting massive parallelism. By incorporating novelty search into the evolutionary process, Conti and his collaborators addressed one of the field's persistent challenges: the tendency of agents to become trapped in local optima during exploration. This work has proven valuable to researchers seeking scalable, computationally efficient alternatives to gradient-based reinforcement learning.