Arena: A General Evaluation Platform and Building Toolkit for Multi-Agent Intelligence

Paper Link: arXiv

Accepted for Poster Presentation at AAAI 2020 (2020)

Co-Authors: Yuhang Song, Andrzej Wojcicki, Thomas Lukasiewicz, Jianyi Wang, Zhenghua Xu, Mai Xu, Zihan Ding, Lianlong Wu

Arena is a general evaluation platform for multi-agent intelligence with 35 games of diverse logics and representations providing a building toolkit for researchers to easily invent and build novel multi-agent problems from the provided game set based on a GUI-configurable social tree and five basic multi-agent reward schemes. All the implementations and accompanied tutorials have been open-sourced for the community code link.

Problems & Challenges in AutoML [working draft]

Paper Link: draft

Co-Authors: Shubhi Sareen

In this paper, we focus on the various algorithms that can generalize of different tasks via AutoML. Our survey will cover central areas in AutoML including NAS, Meta-Learning etc. In parallel, we also highlight the unique advantages of various machine learning techniques including bayesian, evolutionary algorithms, reinforcement learning, deep learning and such. To conclude, we describe the key gaps and limitations of the several current areas of research within the field.