NLP Without a Ready-made Labeled Dataset

Explore Natural Language Processing (NLP) techniques for scenarios without pre-existing labeled datasets in this comprehensive conference talk by Sowmya Vajjala, a researcher at the National Research Council of Canada. Delve into practical approaches for addressing real-world NLP challenges when faced with the absence of annotated data. Learn strategies for dataset creation, including methods for finding existing datasets, data annotation techniques, automatic data labeling processes, data augmentation, and transfer learning applications. Gain valuable insights from Vajjala's decade-long experience in NLP across various roles, and discover how to overcome the common hurdle of lacking a ready-made labeled dataset in NLP projects.