AI Paper Atlas

Research map / Architectures and efficiency / Quantization

Adaptive Low-Precision Training for Embeddings in Click-Through Rate Prediction

Shiwei Li, Huifeng Guo, Lu Hou, Wei Zhang, Xing Tang, Ruiming Tang, Rui Zhang, Ruixuan Li

AAAI 2023 · Architectures and efficiency › Quantization · approach: Ultra-low-bit LLM quantization · 21 citations (OpenAlex)

Read the paper

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
learn a feature-wise step size instead of using a fixed step size (ablation)LPT(SR) to form ALPT(SR) · AvazuAUC: 0.7927 -> 0.7951Improved
learn a feature-wise step size instead of using a fixed step size (ablation)LPT(SR) to form ALPT(SR) · CriteoAUC: 0.8123 -> 0.8144Improved
use stochastic rather than deterministic rounding (ablation)ALPT · CriteoAUC: 0.8096 -> 0.8144Improved
use stochastic rather than deterministic rounding (ablation)LPT · AvazuAUC: 0.7654 -> 0.7927Improved
use stochastic rather than deterministic rounding (ablation)LPT · CriteoAUC: 0.7966 -> 0.8123Improved
use feature-wise adaptive step sizes rather than fixed step sizes (ablation)LPT(SR) · Avazu 2-bitAUC: 0.7829 -> 0.7851Improved
use feature-wise adaptive step sizes rather than fixed step sizes (ablation)LPT(SR) · Criteo 2-bitAUC: 0.8009 -> 0.8033Improved
increase embedding dimension to 32 (ablation)ALPT(SR) · Avazu, d=32AUC: 0.7951 -> 0.7955Mixed
use ALPT instead of LPT (ablation)CTR embedding quantization · Criteo, threshold=2AUC: 0.8115 -> 0.8126Improved
use 2-bit ALPT instead of 2-bit LSQCTR embedding quantization · Avazu 2-bitAUC: 0.7912 -> 0.7851Hurt
use 2-bit ALPT instead of 2-bit LSQCTR embedding quantization · Criteo 2-bitAUC: 0.8089 -> 0.8033Hurt

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