Adriana I. Kovashka
- Associate Professor
Dr. Kovashka received her BA degrees in Computer Science and Media Studies from Pomona College, in 2008, and her PhD in Computer Science from The University of Texas in Austin, in 2014. She joined Pitt's Computer Science department in January 2015.
Her primary research area is computer vision, with overlap in machine learning, information retrieval, crowdsourcing, and natural language processing. Her work examines approaches for improving image retrieval with semantic visual attributes, by ensuring that the most useful feedback is given to the retrieval system, and that the system understands correctly potentially ambiguous and subjective attribute terms. Dr. Kovashka is also interested in improving the communication between human and machine to ensure that vision systems receive rich supervision and can visualize their outputs to their human users, for debugging purposes.
A. Kovashka and K. Grauman, "Interactive Image Search with Attribute-based Guidance and Personalization," IEEE Conference on Computer Vision and Pattern Recognition - Computer Vision and Human Computation Workshop (CVPR 2014), pp. 3432-3439, Columbus, OH, June 2014.
A. Kovashka, and K. Grauman, "Discovering Attribute Shades of Meaning with the Crowd," International Journal of Computer Vision (IJCV), pp. 56-73, 2015.
A. Kovashka, D. Parikh, and K. Grauman, "WhittleSearch: Interactive Image Search with Relative Attribute Feedback," International Journal of Computer Vision (IJCV), pp. 185-210, April 2015.
S. Liu and A. Kovashka, "Adapting Attributes by Selecting Features Similar across Domains," IEEE Winter Conference on Applications of Computer Vision, pp. 1-8, Lake Placid, NY, March 2016.
C. Thomas and A. Kovashka, "Seeing Behind the Camera: Identifying the Authorship of a Photograph," IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2016), pp. 3494-3502, Las Vegas, NV, June 2016.
Content-based image retrieval
Interactions between vision and language
Human-machine communication and crowdsourcing for visual recognition