AI Paper Atlas

Research map / Architectures and efficiency / Quantization

HLHLp: Quantized Neural Networks Training for Reaching Flat Minima in Loss Surface

Sungho Shin, Jinhwan Park, Yoonho Boo, Wonyong Sung

AAAI 2020 · Architectures and efficiency › Quantization · approach: Post-training weight quantization · 5 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 fine-tuning with HLHLp training2-bit ResNet-14 · CIFAR-10test accuracy: 89.20 -> 90.64Improved
replace fine-tuning with HLHLp training2-bit ResNet-32 · CIFAR-10test accuracy: 92.32 -> 93.05Improved
replace fine-tuning with HLHLp training2-bit ResNet-20 · CIFAR-100test accuracy: 64.47 -> 66.44Improved
replace fine-tuning with HLHLp training2-bit MobileNetV2 · CIFAR-100test accuracy: 74.97 -> 75.51Improved
add HLHLp second stage vs HL only on contaminated ImageNet (ablation)2-bit ResNet-18 · ImageNet-C noiseaverage noise accuracy: 39.08% -> 45.99%Improved
quantize PTB GRU with HLHLp instead of Wang et al. ternary2-bit GRU · WikiText-2test PPL: 132.49 -> 105.96Improved
quantize LSTM with HLHLp 4-level instead of Xu et al.2-bit LSTM · WikiText-2test PPL: 106.10 -> 105.5Improved
use HLHLp for CTC acoustic model instead of Lee et al.2-bit LSTM AM · WSJ testCER: 9.76 -> 8.21Improved
use HLHLp for CTC acoustic model instead of Lee et al.2-bit LSTM AM · Aurora-4 noisyCER: 51.65 -> 48.6Improved
ablation (B): float cyclic learning-rate training for 110 epochs then 2-bit (ablation)GRU PTB LM · PTBPPL: 96.97 -> 115.92Hurt
ablation (C): replace 8-bit high-precision step with 2-bit (ablation)GRU PTB LM · PTBPPL: 96.97 -> 97.84Hurt
ablation (D): monotonic decreasing learning rate fine-tuning only (ablation)GRU PTB LM · PTBPPL: 96.97 -> 112.08Hurt

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