White-Box Transformers via Sparse Rate Reduction: Compression Is All There Is?

• URL: http://arxiv.org/abs/2311.13110v4
• Date: Fri, 6 Sep 2024 07:40:40 GMT
• Title: White-Box Transformers via Sparse Rate Reduction: Compression Is All There Is?
• Authors: Yaodong Yu, Sam Buchanan, Druv Pai, Tianzhe Chu, Ziyang Wu, Shengbang Tong, Hao Bai, Yuexiang Zhai, Benjamin D. Haeffele, Yi Ma,
• Abstract summary: We show a family of white-box transformer-like deep network architectures, named CRATE, which are mathematically fully interpretable. Experiments show that these networks, despite their simplicity, indeed learn to compress and sparsify representations of large-scale real-world image and text datasets.
• Score: 27.58916930770997
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: In this paper, we contend that a natural objective of representation learning is to compress and transform the distribution of the data, say sets of tokens, towards a low-dimensional Gaussian mixture supported on incoherent subspaces. The goodness of such a representation can be evaluated by a principled measure, called sparse rate reduction, that simultaneously maximizes the intrinsic information gain and extrinsic sparsity of the learned representation. From this perspective, popular deep network architectures, including transformers, can be viewed as realizing iterative schemes to optimize this measure. Particularly, we derive a transformer block from alternating optimization on parts of this objective: the multi-head self-attention operator compresses the representation by implementing an approximate gradient descent step on the coding rate of the features, and the subsequent multi-layer perceptron sparsifies the features. This leads to a family of white-box transformer-like deep network architectures, named CRATE, which are mathematically fully interpretable. We show, by way of a novel connection between denoising and compression, that the inverse to the aforementioned compressive encoding can be realized by the same class of CRATE architectures. Thus, the so-derived white-box architectures are universal to both encoders and decoders. Experiments show that these networks, despite their simplicity, indeed learn to compress and sparsify representations of large-scale real-world image and text datasets, and achieve performance very close to highly engineered transformer-based models: ViT, MAE, DINO, BERT, and GPT2. We believe the proposed computational framework demonstrates great potential in bridging the gap between theory and practice of deep learning, from a unified perspective of data compression. Code is available at: https://ma-lab-berkeley.github.io/CRATE .

Related papers

• Rethinking Decoders for Transformer-based Semantic Segmentation: Compression is All You Need [3.218600495900291]
  We argue that there are fundamental connections between semantic segmentation and compression. We derive a white-box, fully attentional DEcoder for PrIncipled semantiC segemenTation (DEPICT) Experiments conducted on dataset ADE20K find that DEPICT consistently outperforms its black-box counterpart, Segmenter.
  arXiv  Detail & Related papers  (2024-11-05T12:10:02Z)
• High-Performance Transformers for Table Structure Recognition Need Early Convolutions [25.04573593082671]
  Existing approaches use classic convolutional neural network (CNN) backbones for the visual encoder and transformers for the textual decoder. We design a lightweight visual encoder for table structure recognition (TSR) without sacrificing expressive power. We discover that a convolutional stem can match classic CNN backbone performance, with a much simpler model.
  arXiv  Detail & Related papers  (2023-11-09T18:20:52Z)
• AICT: An Adaptive Image Compression Transformer [18.05997169440533]
  We propose a more straightforward yet effective Tranformer-based channel-wise auto-regressive prior model, resulting in an absolute image compression transformer (ICT) The proposed ICT can capture both global and local contexts from the latent representations. We leverage a learnable scaling module with a sandwich ConvNeXt-based pre/post-processor to accurately extract more compact latent representation.
  arXiv  Detail & Related papers  (2023-07-12T11:32:02Z)
• White-Box Transformers via Sparse Rate Reduction [25.51855431031564]
  We show a family of white-box transformer-like deep network architectures which are mathematically fully interpretable. Experiments show that these networks indeed learn to optimize the designed objective.
  arXiv  Detail & Related papers  (2023-06-01T20:28:44Z)
• The Devil Is in the Details: Window-based Attention for Image Compression [58.1577742463617]
  Most existing learned image compression models are based on Convolutional Neural Networks (CNNs) In this paper, we study the effects of multiple kinds of attention mechanisms for local features learning, then introduce a more straightforward yet effective window-based local attention block. The proposed window-based attention is very flexible which could work as a plug-and-play component to enhance CNN and Transformer models.
  arXiv  Detail & Related papers  (2022-03-16T07:55:49Z)
• CSformer: Bridging Convolution and Transformer for Compressive Sensing [65.22377493627687]
  This paper proposes a hybrid framework that integrates the advantages of leveraging detailed spatial information from CNN and the global context provided by transformer for enhanced representation learning. The proposed approach is an end-to-end compressive image sensing method, composed of adaptive sampling and recovery. The experimental results demonstrate the effectiveness of the dedicated transformer-based architecture for compressive sensing.
  arXiv  Detail & Related papers  (2021-12-31T04:37:11Z)
• Communication-Efficient Federated Learning via Quantized Compressed Sensing [82.10695943017907]
  The presented framework consists of gradient compression for wireless devices and gradient reconstruction for a parameter server. Thanks to gradient sparsification and quantization, our strategy can achieve a higher compression ratio than one-bit gradient compression. We demonstrate that the framework achieves almost identical performance with the case that performs no compression.
  arXiv  Detail & Related papers  (2021-11-30T02:13:54Z)
• Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers [149.78470371525754]
  We treat semantic segmentation as a sequence-to-sequence prediction task. Specifically, we deploy a pure transformer to encode an image as a sequence of patches. With the global context modeled in every layer of the transformer, this encoder can be combined with a simple decoder to provide a powerful segmentation model, termed SEgmentation TRansformer (SETR) SETR achieves new state of the art on ADE20K (50.28% mIoU), Pascal Context (55.83% mIoU) and competitive results on Cityscapes.
  arXiv  Detail & Related papers  (2020-12-31T18:55:57Z)
• Permute, Quantize, and Fine-tune: Efficient Compression of Neural Networks [70.0243910593064]
  Key to success of vector quantization is deciding which parameter groups should be compressed together. In this paper we make the observation that the weights of two adjacent layers can be permuted while expressing the same function. We then establish a connection to rate-distortion theory and search for permutations that result in networks that are easier to compress.
  arXiv  Detail & Related papers  (2020-10-29T15:47:26Z)
• Unfolding Neural Networks for Compressive Multichannel Blind Deconvolution [71.29848468762789]
  We propose a learned-structured unfolding neural network for the problem of compressive sparse multichannel blind-deconvolution. In this problem, each channel's measurements are given as convolution of a common source signal and sparse filter. We demonstrate that our method is superior to classical structured compressive sparse multichannel blind-deconvolution methods in terms of accuracy and speed of sparse filter recovery.
  arXiv  Detail & Related papers  (2020-10-22T02:34:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.