Amandianeze Nwana
Research Interests
- Automatic Image Tagging
- Information Retrieval
- Social and Information Network Inference and Extraction
- Machine Learning
- Computer Vision
Current Research
Personalizing Automatic Image Tagging/Annotation: There is so much rich information available on the web through images, and tags provide us a means through which to sort these sorts of information. Because of the massive amounts of images that exist on the web, it is infeasible for every user to tag all the images that interest them, and this leads to many images being left untagged. This makes it hard for users to find images that they might be interested in but have been hitherto left untagged. Since each user tends to have their own unique way of describing and tagging images, if we aim to automatically tag images for users in order to aid the recall or discovery of images that they are interested in, we must take into account these nuances. My current research direction is to explore novel ways of inferring user tag preferences, and ways to create and exploit personalization profiles.
Awards and Honors:
- Bouchet Honor Society Fellow (Inducted 2015)
- Best Presentation Award in Sciences, Bouchet Conference '15
- Cornell Sloan Fellowship (2011-2014)
- Qualcomm Innovation Fellowship Finalist (2014)
Selected Publications:
- A. Nwana and T. Chen, “Towards Understanding User Preferences from User Tagging Behavior for Personalization,” IEEE International Symposium on Multimedia (ISM), 2015.
- A. Nwana, S. Avestimehr and T. Chen, “A Latent Social Approach to YouTube Popularity Prediction”, Globecom 2013 - Symposium on Selected Areas in Communications (GC13 SAC), IEEE Globecom’13 (2013)