Online Learned Continual Compression with Adaptive Quantization Modules
This addresses the challenge of efficient data storage and representation learning in continual learning settings for AI systems handling streaming data.
The paper tackles the problem of online continual compression, where models must learn to compress and store data from a non-i.i.d. stream with single-pass observation, by introducing Adaptive Quantization Modules (AQM) to control compression variation without pretraining. It shows significant gains on continual learning benchmarks, with demonstrations on larger images, LiDAR, and reinforcement learning environments.
We introduce and study the problem of Online Continual Compression, where one attempts to simultaneously learn to compress and store a representative dataset from a non i.i.d data stream, while only observing each sample once. A naive application of auto-encoders in this setting encounters a major challenge: representations derived from earlier encoder states must be usable by later decoder states. We show how to use discrete auto-encoders to effectively address this challenge and introduce Adaptive Quantization Modules (AQM) to control variation in the compression ability of the module at any given stage of learning. This enables selecting an appropriate compression for incoming samples, while taking into account overall memory constraints and current progress of the learned compression. Unlike previous methods, our approach does not require any pretraining, even on challenging datasets. We show that using AQM to replace standard episodic memory in continual learning settings leads to significant gains on continual learning benchmarks. Furthermore we demonstrate this approach with larger images, LiDAR, and reinforcement learning environments.