AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Domain Agnostic Real-Valued Specificity Prediction

Wei-Jen Ko, Greg Durrett, Junyi Jessy Li

AAAI 2019 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Statistical alignment and hypothesis testing

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 self-ensembling with consistency loss and distribution regularizationsentence specificity prediction · Twitter, Yelp, and MovieMAE: 0.237 -> 0.113; 0.325 -> 0.105; 0.226 -> 0.114Improved
add mean-std distribution regularization to consistency training (ablation)SE+A · TwitterMAE: 0.151$\pm$0.005 -> 0.113$\pm$0.001Improved
add mean-std distribution regularization to consistency training (ablation)SE+A · YelpMAE: 0.137$\pm$0.005 -> 0.107$\pm$0.003Improved
add mean-std distribution regularization to consistency training (ablation)SE+A · MovieMAE: 0.148$\pm$0.001 -> 0.114$\pm$0.001Improved
add consistency loss without distribution regularization (ablation)SE · TwitterMAE: 0.148$\pm$0.006 -> 0.151$\pm$0.005Hurt
add mean-std distribution regularization without consistency loss (ablation)SE · TwitterMAE: 0.148$\pm$0.006 -> 0.125$\pm$0.001Improved
remove data augmentation (ablation)SE+AD (mean-std) · TwitterSpearman: 0.676$\pm$0.004 -> 0.666Hurt
remove data augmentation (ablation)SE+AD (mean-std) · YelpSpearman: 0.750$\pm$0.016 -> 0.728Hurt
remove data augmentation (ablation)SE+AD (mean-std) · MovieSpearman: 0.702$\pm$0.003 -> 0.669Hurt
use KL divergence rather than mean and standard deviation to regularize the posterior distribution (ablation)SE+AD · Twitter, Yelp, and MovieMAE: Twitter 0.113$\pm$0.001 vs 0.115$\pm$0.002; Yelp 0.107$\pm$0.003 vs 0.105$\pm$0.002; Movie 0.114$\pm$0.001 vs 0.114$\pm$No clear effect
filter general responses by predicted specificity instead of filtering short responsesOpenSubtitles dialogue generation · OpenSubtitles dialogueinformativeness preference: 29.0 vs 38.9Improved
filter general responses by predicted specificity instead of filtering short responsesOpenSubtitles dialogue generation · OpenSubtitles dialogueperplexity: 51.97 -> 46.92Improved

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