AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Learning to Screen for Fast Softmax Inference on Large Vocabulary Neural Networks

Patrick Chen, Si Si, Sanjiv Kumar, Yang Li, Cho-Jui Hsieh

ICLR 2019 · Architectures and efficiency › RNNs and normalization · approach: Non-autoregressive translation · 9 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
Use learned context-cluster screening and exact softmax over the selected candidate subset instead of full softmaxL2S · PTB-Smallspeedup: 10.6xImproved
Use learned context-cluster screening and exact softmax over the selected candidate subset instead of full softmaxL2S · PTB-Largespeedup: 45.3xImproved
Use learned context-cluster screening and exact softmax over the selected candidate subset instead of full softmaxL2S · NMT: DE-ENspeedup: 20.4xImproved
Use L2S instead of FGD in beam-search decodingNMT: DE-EN, beam 5 · NMT: DE-EN, beam 5speedup: 2.9x -> 13.4xImproved
Use L2S instead of FGD in beam-search decodingNMT: EN-VE, beam 5 · NMT: EN-VE, beam 5speedup: 4.6x -> 20xImproved
Use L2S for perplexity computation with low-rank logits outside the candidate setPTB-Small language model · PTB-Small perplexityspeedup: 1x -> 5.69xImproved
Use L2S for perplexity computation with low-rank logits outside the candidate setPTB-Large language model · PTB-Large perplexityPPL: 78.32 -> 80.09Hurt
Increase the number of clusters from 50 to 100 while adjusting the time budget (ablation)L2S · cluster-count experimentP@1: 0.997 -> 0.998No clear effect
Increase the number of clusters from 100 to 250 while adjusting the time budget (ablation)L2S · cluster-count experimentP@1: 0.998 -> 0.994No clear effect
Replace end-to-end learned screening with spherical k-means clustering (ablation)L2S · PTB-Smallspeedup: 10.6x -> 4xHurt
Replace end-to-end learned screening with spherical k-means clustering (ablation)L2S · PTB-Largespeedup: 45.3x -> 6.9xHurt
Replace end-to-end learned screening with spherical k-means clustering (ablation)L2S · NMT: DE-ENspeedup: 20.4x -> 13.8xHurt

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