Women Snowboard Too: Overcoming Bias in Captioning Models & Other Key Research in Computer Vision

Women Also Snowboard: Overcoming Bias in Captioning Models

Kaylee Burns, Lisa Anne Hendricks, Kate Saenko, Trevor Darrell, and Anna Rohrbach. Women also snowboard: Overcoming bias in captioning models. In ECCV, 2018.

This paper addresses the issue of gender bias in captioning models, highlighting the importance of representing women accurately in visual descriptions. The researchers explore strategies to overcome this bias and create more inclusive captions.

ActivityNet: A Large-Scale Video Benchmark

Fabian Caba Heilbron, Victor Escorcia, Bernard Ghanem, and Juan Carlos Niebles. ActivityNet: A large-scale video benchmark for human activity understanding. In CVPR, 2015.

ActivityNet is a comprehensive dataset designed to facilitate research in human activity understanding. It provides a large-scale collection of videos with annotated actions, enabling researchers to develop and evaluate models for tasks like action recognition and video summarization.

Microsoft COCO Captions: Data Collection and Evaluation

Xinlei Chen, Hao Fang, Tsung-Yi Lin, Ramakrishna Vedantam, Saurabh Gupta, Piotr Dollar, and C Lawrence Zitnick. 'Microsoft COCO captions: Data collection and evaluation server.' arXiv preprint arXiv:1504.00325, 2015.

This paper introduces the Microsoft COCO captions dataset, a valuable resource for training and evaluating image captioning models. It describes the dataset's creation process, including the collection of images and human-generated captions, and provides insights into the evaluation methods used to assess the performance of captioning systems.

Uniter: Universal Image-Text Representation Learning

Yen-Chun Chen, Linjie Li, Licheng Yu, Ahmed El Kholy, Faisal Ahmed, Zhe Gan, Yu Cheng, and Jingjing Liu. Uniter: Universal image-text representation learning. In ECCV, 2020.

Uniter is a powerful model for learning universal representations of images and text. It aims to capture the underlying relationships between visual and textual information, enabling applications like image-text retrieval and cross-modal understanding.

Dealing with Disagreements in Subjective Annotations

Aida Mostafazadeh Davani, Mark D´ıaz, and Vinodkumar Prabhakaran. Dealing with disagreements: Looking beyond the majority vote in subjective annotations. Trans. ACL, 2022.

This paper addresses the challenge of handling disagreements in subjective annotations, a common problem in areas like sentiment analysis and opinion mining. The researchers propose methods to go beyond simple majority voting and effectively capture the diversity of opinions expressed in annotations.