Online Learned Continual Compression with Stacked Quantization Module

By Lucas Caccia, Eugene Belilovsky, Massimo Caccia, Joelle Pineau. 2019

6 citations recorded by the source corpus.

We introduce and study the problem of Online Continual Compression, where one attempts to learn to compress and store a representative dataset from a non i.i.d data stream, while only observing each sample once. This problem is highly relevant for downstream online continual learning tasks, as well as standard learning methods under resource constrained data collection. To address this we propose a new architecture which Stacks Quantization Modules (SQM), consisting of a series of discrete autoencoders, each equipped with their own memory. Every added module is trained to reconstruct the latent space of the previous module using fewer bits, allowing the learned representation to become more compact as training progresses. This modularity has several advantages: 1) moderate compressions are quickly available early in training, which is crucial for remembering the early tasks, 2) as more data needs to be stored, earlier data becomes more compressed, freeing memory, 3) unlike previous methods, our approach does not require pretraining, even on challenging datasets. We show several potential applications of this method. We first replace the episodic memory used in Experience Replay with SQM, leading to significant gains on standard continual learning benchmarks using a fixed memory budget. We then apply our method to online compression of larger images like those from Imagenet, and show that it is also effective with other modalities, such as LiDAR data.

Subfields: Artificial Intelligence, Computational Mechanics, Ocean Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computational Mechanics

In Ocean Engineering

Citations in the corpus

Towards Continual Reinforcement Learning: A Review and Perspectives (2020, background): “efficient, though now only consisting of approximate recollections (Riemer et al., 2017; Caccia et al., 2019).”
Online Fast Adaptation and Knowledge Accumulation: a New Approach to Continual Learning (2020, methodology): “…2017; Isele & Cosgun, 2018; Rolnick et al., 2019; Aljundi et al., 2019a), constrained optimization based on those samples (Lopez-Paz & Ranzato, 2017)), or in a compressed form, e.g., via a generative model (Aljundi et al., 2019a; Caccia et al., 2019; Ostapenko et al., 2019; Lesort et al., 2019a).”
Regularization Shortcomings for Continual Learning (2019, background): “• Rehearsal: In order to maintain knowledge from past learning experiences, the algorithms save a subset of old data as memory [21, 19, 1, 2, 8, 28, 7, 3, 26].”