Pretraining Action Recognition Models on HowTo100M for Cross-Domain Few-Shot Adaptation in EPIC-Kitchens-100

This report presents the technical details of our submission to the EPIC-Kitchens-100 Action Recognition Challenge 2021. To participate in the challenge we deployed spatio-temporal feature extraction and aggregation models we have developed recently: GSF and XViT. GSF is an efficient spatio-temporal feature extracting module that can be plugged into 2D CNNs for video action recognition. XViT is a convolution free video feature extractor based on transformer architecture. We design an ensemble of GSF and XViT model families with different backbones and pretraining to generate the prediction sco

Research goal: What is the impact of pretraining action recognition models on HowTo100M data versus other large-scale video datasets (e.g., Kinetics, ActivityNet) on cross-domain few-shot adaptation accuracy in EPIC-Kitchens-100, measured via mean Average Precision (mAP)?

Autonomous synthesis report generated by SOVEREIGN Research Kernel. Tribunal consensus score: 8.4/10.