AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

D-Separation for Causal Self-Explanation

Wei Liu, Jun Wang, Haozhao Wang, Ruixuan Li, Zhiying Deng, YuanKai Zhang, Yang Qiu

NeurIPS 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Group-robust classification · 2 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 MCD instead of FRFR · correlated BeerAdvocate Aroma S~10F1: 52.1 -> 65.8Improved
use MCD instead of FRFR · correlated BeerAdvocate Appearance S~10F1: 52.6 -> 63.9Improved
use MCD instead of FRFR · correlated BeerAdvocate Palate S~10F1: 42.3 -> 53.1Improved
use MCD instead of FRFR · correlated BeerAdvocate Appearance S~30F1: 62.3 -> 73.4Improved
use MCD instead of FRFR · correlated BeerAdvocate Appearance S~20F1: 79.6 -> 82.3Improved
use MCD instead of FRFR · HotelReviewF1: 44.8 -> 47.8Improved
replace KL-divergence with JS-divergence in MCD criterion (ablation)MCD · correlated BeerAdvocate Appearance S~10F1: 63.9 -> 65.9Improved
replace KL-divergence with JS-divergence (ablation)MCD · correlated BeerAdvocate Palate S~30F1: 42.9 -> 37.1Hurt
replace KL-divergence with JS-divergence (ablation)MCD · correlated BeerAdvocate Aroma S~20F1: 72.8 -> 66.2No clear effect
increase mask correlation (skew 65 to 75) (ablation)FR · Palate decorrelatedF1: 63.4 -> 51.0Hurt
increase mask correlation (skew 65 to 75) (ablation)MCD · Palate decorrelatedF1: 62.6 -> 63.1No clear effect
swap GRU encoder for pretrained ELECTRA (ablation)MCD · decorrelated Beer-AppearanceF1: 80.1 -> 85.2Improved

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