AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Accelerating Deep Learning with Millions of Classes

Zhuoning Yuan, Zhishuai Guo, Xiaotian Yu, Xiaoyu Wang, Tianbao Yang

ECCV 2020 · Architectures and efficiency › RNNs and normalization · approach: Efficient network compression and training · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase random projection size m from 10 to 100 (ablation)RP softmax · MS-Celeb-1M average verificationaccuracy: 0.934 -> 0.965Improved
increase random projection size m from 100 to 1000 (ablation)RP softmax · MS-Celeb-1M average verificationaccuracy: 0.965 -> 0.960No clear effect
increase random projection size m from 10 to 500 (ablation)RP softmax · Penn Treebankperplexity: 137.59 -> 83.98Improved
increase random projection size m from 10 to 500 (ablation)RP softmax · Google Billion Wordsperplexity: 122.16 -> 46.34Improved
add random projected softmax in place of full softmaxResNet50 · MegaFace2 LFWverification accuracy: 95.58 -> 96.60Improved
add random projected softmax in place of full softmaxResNet50 · ImageNet valtop-1 accuracy: 75.69 -> 73.98Hurt
add random projected softmax with sampled softmaxLSTM · Google Billion Wordsspeedup: 2.97 -> 12.5Improved
add recovery of full softmax classifier after projected training (ablation)RP softmax · Penn Treebankperplexity: 83.98 -> 74.78Improved
add recovery of full softmax classifier after projected training (ablation)RP softmax · ImageNet valtop-1 accuracy: 73.98 -> 74.56Improved
add recovery of full softmax classifier after projected training (ablation)RP softmax · Google Billion Wordsperplexity: 46.38 -> 46.56No clear effect
add adaptive projection size (larger for head classes) to RP (ablation)RP softmax · Google Billion Wordsperplexity: 46.38 -> 74.3Hurt
add adaptive projection size (larger for head classes) to RP (ablation)RP softmax · MS-Celeb-1M AgeDBverification accuracy: 95.22 -> 95.42Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy