UnifiedQA-v2: Stronger Generalization via Broader Cross-Format Training

02/23/2022

∙

We present UnifiedQA-v2, a QA model built with the same process as UnifiedQA, except that it utilizes more supervision – roughly 3x the number of datasets used for UnifiedQA. This generally leads to better in-domain and cross-domain results.

READ FULL TEXT

UnifiedQA-v2: Stronger Generalization via Broader Cross-Format Training

Learning to Generalize for Cross-domain QA

Multi-Source (Pre-)Training for Cross-Domain Measurement, Unit and Context Extraction

Cross-Domain Ensemble Distillation for Domain Generalization

Bridge Data: Boosting Generalization of Robotic Skills with Cross-Domain Datasets

TTT-UCDR: Test-time Training for Universal Cross-Domain Retrieval

On the limits of cross-domain generalization in automated X-ray prediction

Grand challenges and emergent modes of convergence science

UnifiedQA-v2: Stronger Generalization via Broader Cross-Format Training

Related Research

Learning to Generalize for Cross-domain QA

Multi-Source (Pre-)Training for Cross-Domain Measurement, Unit and Context Extraction

Cross-Domain Ensemble Distillation for Domain Generalization

Bridge Data: Boosting Generalization of Robotic Skills with Cross-Domain Datasets

TTT-UCDR: Test-time Training for Universal Cross-Domain Retrieval

On the limits of cross-domain generalization in automated X-ray prediction

Grand challenges and emergent modes of convergence science