AI Paper Atlas

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

A Distributional Approach to Controlled Text Generation

Muhammad Khalifa, Hady Elsahar, Marc Dymetman

ICLR 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 39 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 GDC instead of PPLM for positive sentiment controlGPT-2 · Positive sentimentsentiment expectation: 0.52 -> 0.56Improved
Use GDC instead of PPLM for positive sentiment controlGPT-2 · Positive sentimentperplexity: 29.26$\pm$22.07 -> 13.53$\pm$3.18Improved
Use GDC instead of PPLM for negative sentiment controlGPT-2 · Negative sentimentsentiment expectation: 0.14 -> 0.51Improved
Use GDC instead of PPLM for negative sentiment controlGPT-2 · Negative sentimentperplexity: 27.72$\pm$23.95 -> 13.59$\pm$3.84Improved
Add a female-rate distributional constraint targeting 50%GPT-2bio · Biography generationfemale biography rate: 07.4 -> 36.7Improved
Add multiple profession-rate distributional constraintsGPT-2bio · Biography generationScience rate: 01.5 -> 20.1Improved
Add multiple profession-rate distributional constraintsGPT-2bio · Biography generationArt rate: 10.9 -> 31.6Improved
Add multiple profession-rate distributional constraintsGPT-2bio · Biography generationBusiness rate: 10.9 -> 10.2Improved
Add multiple profession-rate distributional constraintsGPT-2bio · Biography generationSports rate: 19.5 -> 11.9Improved
Add adaptive proposal updates to DPG (ablation)DPG for rare-word constraints · Vampire, Paris, and US word constraintsconvergence speedImproved
Use TVD rather than KL divergence to decide proposal updates (ablation)DPG-KLD · Vampire, Paris, and US word constraintsconvergence speedNo clear effect
Remove adaptive proposal updates (ablation)DPG · Vampire, Paris, and US word constraintsconvergence speedHurt

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