Welcome fellow researchers to the DeCLaRe Lab! We hereby DeCLaRe our quest to breathe cognitive and language skills of human-like depth into machines by solving challenging NLP problems, such as, dialogue comprehension and generation, commonsense reasoning, multimodal understanding, and more. Addressing such open research problems requires powerful, scalable, and data-hungry algorithms. As such, we develop cutting-edge neural models, based on sound linguistic concepts. To know more about our work, please browse our catalog of publications.

> Intrigued by our logo? Read the story behind it.



We have a number of openings to join us at the DeCLaRe Lab as postdocs and research assistants/associates. Drop an email to Soujanya


The DeCLaRe lab has been awarded programmatic funding by AME. We are grateful to AME for their generosity. The project will focus on building knowledge graph-based question answering.


The DeCLaRe lab secured funding from DSO for a project on complex question answering. Thank you, DSO!

New paper

New paper on transfer learning - Hazarika, Devamanyu, Soujanya Poria, Roger Zimmermann, and Rada Mihalcea. "Emotion Recognition in Conversations with Transfer Learning from Generative Conversation Modeling. [Download the paper].