In recent years, social scientists have begun using computational methods to study images, launching a new “images-as-data” research field (Joo & Steinert-Threlkeld 2018). Either through home-grown classifiers or pre-trained image labeling services, it is now possible to analyze image content at large scales. These methods open new avenues of research in a variety of sub-disciplines in the social sciences, including communication and behavioral studies. At the same time, researchers are actively working on addressing known biases in computational image categorization.

To take stock of new methodological and empirical developments in this exciting field, we invite interested scholars to participate in a 2-day workshop in Dublin.

We encourage both substantive and methodological papers from all sub-disciplines of the social sciences, as long as they share a focus on applying image-as-data methods to questions of interest to these disciplines. Topics include, but are not limited to:

• Using images to study (political) communication• Developing image training data
• Image classification algorithm development
• Gender and ethnic bias in image classification

• Emotion detection in images
• Validation of image classifiers
• Applications and ethics of synthetic (“deepfake”) video and audio
• Linking qualitative and quantitative approaches to the study of images

The workshop will take place on 28-29 May 2020 on the campus of Trinity College Dublin in Dublin, Ireland starting 28 May around noon and ending 29 May at around the same time.

We strongly encourage applications from women, racial and ethnic minorities, and junior scholars. We also encourage participation from scholars who rely on mixed methods in the study of images.