AI Paper Atlas

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

Face Model Compression by Distilling Knowledge from Neurons

AAAI 2016 · Label-efficient and robust learning › Knowledge distillation · approach: Few-shot and metric distillation · 219 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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
add neuron selection over top hidden-layer targets (ablation)student mimicking T1 · LFWAUC: 97.73 -> 98.18Helped
add neuron selection (ablation)student mimicking T2 · LFWAUC: 97.90 -> 98.12Helped
add neuron selection (ablation)student mimicking T1 ensemble · LFWAUC: 98.07 -> 98.25Helped
use hidden-layer neuron targets instead of soft targets (t=10)student of T1 · LFWAUC: 97.15 -> 97.73Helped
raise soft target temperature from 1 to 10student of T1 · LFWAUC: 94.42 -> 97.15Helped
add attributes as extra supervision to 1-of-K studentS-1-of-K · LFWaccuracy: 91.23 -> 91%No clear effect
add unsupervised fine-tuning on teacher features (ablation)S-selection mimicking T1 ensemble · LFWAUC: 98.25 -> 98.43Helped
compress T1 ensemble into small studentT1 ensemble · CPU inferencespeed-up: 90x speed-up, 51.6x compressionHelped

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