Niveda Krishnamoorthy

Suggest Changes
Learn More
We are interested in the task of image annotation using noisy natural text as training data. An image and its caption convey different information, but are generated by the same underlying concepts.(More)