PreCNet: Next-Frame Video Prediction Based on Predictive Coding

TL;DR

PreCNet, a deep learning model inspired by predictive coding theory, achieves state-of-the-art results in next-frame video prediction by faithfully implementing a neuroscience-inspired architecture.

Contribution

This work introduces PreCNet, a novel deep learning model based on predictive coding, demonstrating its effectiveness in next-frame video prediction tasks.

Findings

Achieved state-of-the-art performance on urban video prediction benchmarks.

Performance improved with larger training datasets, highlighting data limitations.

Demonstrated that neuroscience-inspired architectures can excel in machine learning tasks.

Abstract

Predictive coding, currently a highly influential theory in neuroscience, has not been widely adopted in machine learning yet. In this work, we transform the seminal model of Rao and Ballard (1999) into a modern deep learning framework while remaining maximally faithful to the original schema. The resulting network we propose (PreCNet) is tested on a widely used next frame video prediction benchmark, which consists of images from an urban environment recorded from a car-mounted camera, and achieves state-of-the-art performance. Performance on all measures (MSE, PSNR, SSIM) was further improved when a larger training set (2M images from BDD100k), pointing to the limitations of the KITTI training set. This work demonstrates that an architecture carefully based in a neuroscience model, without being explicitly tailored to the task at hand, can exhibit exceptional performance.