Mastering 2048 with Delayed Temporal Coherence Learning, Multi-Stage Weight Promotion, Redundant Encoding and Carousel Shaping

TL;DR

This paper introduces advanced reinforcement learning techniques, including delayed temporal coherence, multi-stage weight promotion, redundant encoding, and carousel shaping, to significantly improve AI performance in the 2048 game, setting new benchmarks.

Contribution

It presents novel enhancements to temporal difference learning with n-tuple networks, achieving the strongest 2048 AI to date by leveraging parallelism and innovative function approximation strategies.

Findings

Achieved state-of-the-art 2048 playing performance

Enhanced learning efficiency through delayed updates and lock-free parallelism

Validated effectiveness of proposed methods for discrete-state Markov decision processes

Abstract

2048 is an engaging single-player, nondeterministic video puzzle game, which, thanks to the simple rules and hard-to-master gameplay, has gained massive popularity in recent years. As 2048 can be conveniently embedded into the discrete-state Markov decision processes framework, we treat it as a testbed for evaluating existing and new methods in reinforcement learning. With the aim to develop a strong 2048 playing program, we employ temporal difference learning with systematic n-tuple networks. We show that this basic method can be significantly improved with temporal coherence learning, multi-stage function approximator with weight promotion, carousel shaping, and redundant encoding. In addition, we demonstrate how to take advantage of the characteristics of the n-tuple network, to improve the algorithmic effectiveness of the learning process by i) delaying the (decayed) update and applying lock-free optimistic parallelism to effortlessly make advantage of multiple CPU cores. This way, we were able to develop the best known 2048 playing program to date, which confirms the effectiveness of the introduced methods for discrete-state Markov decision problems.