AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Stochastic Transformer Networks With Linear Competing Units: Application To End-to-End SL Translation

Andreas Voskou, Konstantinos P. Panousis, Dimitrios Kosmopoulos, Dimitris N. Metaxas, Sotirios Chatzis

ICCV 2021 · Architectures and efficiency › Attention and transformers · approach: Efficient language transformers · 43 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 LWTA activations (U=4) (ablation)Bayesian Transformer for SLT · PHOENIX 2014T testBLEU-4: 22.61 -> 23.65Improved
replace ELU with LWTA activations (U=4) (ablation)Bayesian Transformer for SLT · PHOENIX 2014T testBLEU-4: 22.56 -> 23.65Improved
replace SiLU with LWTA activations (U=4) (ablation)Bayesian Transformer for SLT · PHOENIX 2014T testBLEU-4: 22.33 -> 23.65Improved
use U=4 competing units instead of U=8 per block (ablation)LWTA Transformer · PHOENIX 2014T testBLEU-4: 22.96 -> 23.65Improved
use U=4 competing units instead of U=2 per block (ablation)LWTA Transformer · PHOENIX 2014T testBLEU-4: 22.82 -> 23.65Improved
use U=4 instead of U=16 per block (ablation)LWTA Transformer · PHOENIX 2014T testBLEU-4: 22.52 -> 23.65Improved
use variational Gaussian weights instead of point-estimates (ablation)LWTA Transformer · PHOENIX 2014T testBLEU-4: 22.34 -> 23.65Improved
use encoder-decoder depth 2-2 instead of 1-1 (ablation)LWTA Transformer · PHOENIX 2014T testBLEU-4: 21.85 -> 23.65Improved
use depth 2-2 instead of 3-3 (ablation)LWTA Transformer · PHOENIX 2014T testBLEU-4: 21.29 -> 23.65Improved
apply weight compression at inference (ablation)LWTA Transformer 2-2 · PHOENIX 2014T testBLEU-4: 23.65 -> 23.52No clear effect
apply weight compression at inference (ablation)LWTA Transformer 3-3 · PHOENIX 2014T testBLEU-4: 21.29 -> 20.77Hurt
use ensemble decoding with L=8 instead of single modelLWTA Transformer 2-2 · PHOENIX 2014T testBLEU-4: 23.65 -> 25.59Improved

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