Depthwise Discrete Representation Learning
This addresses the challenge of representing latent factors like words or shapes with discrete variables for researchers in machine learning, though it appears incremental as it builds on existing VQVAE methods.
The paper tackles the problem of learning discrete representations by applying vector quantization along the feature axis, hypothesizing that this maps codebook vectors to the marginal distribution of the prior feature space. This approach achieves a 33% improvement over previous discrete models and similar performance to state-of-the-art autoregressive models on benchmarks like CIFAR-10 and ImageNet.
Recent advancements in learning Discrete Representations as opposed to continuous ones have led to state of art results in tasks that involve Language, Audio and Vision. Some latent factors such as words, phonemes and shapes are better represented by discrete latent variables as opposed to continuous. Vector Quantized Variational Autoencoders (VQVAE) have produced remarkable results in multiple domains. VQVAE learns a prior distribution $z_e$ along with its mapping to a discrete number of $K$ vectors (Vector Quantization). We propose applying VQ along the feature axis. We hypothesize that by doing so, we are learning a mapping between the codebook vectors and the marginal distribution of the prior feature space. Our approach leads to 33\% improvement as compared to prevous discrete models and has similar performance to state of the art auto-regressive models (e.g. PixelSNAIL). We evaluate our approach on a static prior using an artificial toy dataset (blobs). We further evaluate our approach on benchmarks for CIFAR-10 and ImageNet.