“In 2024, Google utilized research from a 2019 paper by William Wang, published at ICCV for the VATEX dataset, to build video understanding components for its Gemini model.”