Timezone: »

 
Poster
Multi-Sample Training for Neural Image Compression
Tongda Xu · Yan Wang · Dailan He · Chenjian Gao · Han Gao · Kunzan Liu · Hongwei Qin

@

This paper considers the problem of lossy neural image compression (NIC). Current state-of-the-art (SOTA) methods adopt uniform posterior to approximate quantization noise, and single-sample pathwise estimator to approximate the gradient of evidence lower bound (ELBO). In this paper, we propose to train NIC with multiple-sample importance weighted autoencoder (IWAE) target, which is tighter than ELBO and converges to log likelihood as sample size increases. First, we identify that the uniform posterior of NIC has special properties, which affect the variance and bias of pathwise and score function estimators of the IWAE target. Moreover, we provide insights on a commonly adopted trick in NIC from gradient variance perspective. Based on those analysis, we further propose multiple-sample NIC (MS-NIC), an enhanced IWAE target for NIC. Experimental results demonstrate that it improves SOTA NIC methods. Our MS-NIC is plug-and-play, and can be easily extended to neural video compression.

Author Information

Tongda Xu (Tsinghua University)
Yan Wang (sensetime)
Dailan He (The Chinese University of Hong Kong)
Chenjian Gao (SenseTime Research)
Han Gao (University of Electronic Science and Technology of China)
Kunzan Liu (MIT)
Hongwei Qin (SenseTime)

More from the Same Authors

  • 2022 Poster: Flexible Neural Image Compression via Code Editing »
    Chenjian Gao · Tongda Xu · Dailan He · Yan Wang · Hongwei Qin
  • 2023 Poster: Idempotent Learned Image Compression with Right-Inverse »
    Yanghao Li · Tongda Xu · Yan Wang · Jingjing Liu · Ya-Qin Zhang
  • 2022 Spotlight: Multi-Sample Training for Neural Image Compression »
    Tongda Xu · Yan Wang · Dailan He · Chenjian Gao · Han Gao · Kunzan Liu · Hongwei Qin
  • 2022 Spotlight: Lightning Talks 3B-1 »
    Tianying Ji · Tongda Xu · Giulia Denevi · Aibek Alanov · Martin Wistuba · Wei Zhang · Yuesong Shen · Massimiliano Pontil · Vadim Titov · Yan Wang · Yu Luo · Daniel Cremers · Yanjun Han · Arlind Kadra · Dailan He · Josif Grabocka · Zhengyuan Zhou · Fuchun Sun · Carlo Ciliberto · Dmitry Vetrov · Mingxuan Jing · Chenjian Gao · Aaron Flores · Tsachy Weissman · Han Gao · Fengxiang He · Kunzan Liu · Wenbing Huang · Hongwei Qin
  • 2022 Spotlight: Lightning Talks 2B-2 »
    Chenjian Gao · Rui Ding · Lingzhi LI · Fan Yang · Xingting Yao · Jianxin Li · Bing Su · Zhen Shen · Tongda Xu · Shuai Zhang · Ji-Rong Wen · Lin Guo · Fanrong Li · Kehua Guo · Zhongshu Wang · Zhi Chen · Xiangyuan Zhu · Zitao Mo · Dailan He · Hui Xiong · Yan Wang · Zheng Wu · Wenbing Tao · Jian Cheng · Haoyi Zhou · Li Shen · Ping Tan · Liwei Wang · Hongwei Qin
  • 2022 Spotlight: Flexible Neural Image Compression via Code Editing »
    Chenjian Gao · Tongda Xu · Dailan He · Yan Wang · Hongwei Qin