AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Light Multi-Segment Activation for Model Compression

Zhenhui Xu, Guolin Ke, Jia Zhang, Jiang Bian, Tie-Yan Liu

AAAI 2020 · Label-efficient and robust learning › Knowledge distillation · approach: Rethinking distillation objectives · 2 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
Replace ReLU with LMA-8Distilled compact image classifiers · CIFAR-10, Students 1–3accuracy: 88.74 $\pm$0.25 -> 90.57 $\pm$0.20; 82.67 $\pm$0.46 -> 85.66 $\pm$0.34; 73.33 $\pm$0.79 -> 77.66 $\pm$0.47Improved
Replace ReLU with LMA-8Distilled WRN-10 students · CIFAR-100accuracy: 69.11 $\pm$0.80 -> 70.92 $\pm$0.42; 63.12 $\pm$1.00 -> 66.31 $\pm$0.68Improved
Replace ReLU with LMA-8Distilled translation models · Ope, Students 1–3BLEU: 13.73 $\pm$0.19 -> 14.11 $\pm$0.12; 10.46 $\pm$0.18 -> 10.94 $\pm$0.18; ~~6.12 $\pm$0.12 -> ~~6.64 $\pm$0.04Improved
Replace ReLU with LMA-8Distilled translation models · WMT13BLEU: 26.89 $\pm$0.05 -> 27.12 $\pm$0.07; 20.39 $\pm$0.09 -> 21.19 $\pm$0.08Improved
Replace ReLU with LMA-8Distilled translation models · Ope and WMT13perplexity: 31.84 $\pm$0.31 -> 30.21 $\pm$0.25; 44.51 $\pm$0.52 -> 41.21 $\pm$0.35; 71.69 $\pm$0.51 -> 67.62 $\pm$0.31; ~~6.44 $\pm$Improved
Use LMA-8 rather than ReLU in quantized distillationQuantized CIFAR-10 student models · CIFAR-10, 4 bitsaccuracy: 85.74 $\pm$0.15 -> 86.31 $\pm$0.41; 77.04 $\pm$0.51 -> 79.48 $\pm$0.79; 65.33 $\pm$0.17 -> 68.85 $\pm$0.99Improved
Use LMA-8 rather than ReLU in quantized distillationQuantized CIFAR-10 student models · CIFAR-10, 8 bitsaccuracy: 87.02 $\pm$0.23 -> 88.56 $\pm$0.52; 80.53 $\pm$0.75 -> 83.37 $\pm$0.51; 70.23 $\pm$0.98 -> 74.47 $\pm$0.74Improved
Replace ReLU with LMA-8 instead of APLU-8Compact models · CIFAR-10 and translation tasksinference memoryImproved
Increase the LMA segment count from 4 to 8 (ablation)CIFAR-10 student models · CIFAR-10accuracyImproved
Increase the LMA segment count beyond 10 (ablation)CIFAR-10 student models · CIFAR-10accuracyHurt
Replace LMA with Maxout-8 at the same Student 1 model size (ablation)CIFAR-10 Student 1 · CIFAR-10accuracy: 88.74$\pm$0.25% -> 87.76$\pm$0.55%Hurt

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