Overview

Natural Language Interfaces (NLInt) offer an easy way for humans to interact with complex systems underneath without understanding the technical details of the system. With the advent of deep learning in the past few years we have begun to scratch the surface of complex NLInt tasks like question answering, dialog systems, semantic parsing etc. that have also paved way in the industry to build virtual assistants like Alexa, Google Home and Siri that are being used by millions of users worldwide. However, there is a need for systems that understand language on a deeper level. NLInt come in a variety of communication forms like speech, image and/or text that users are most familiar with to carry out their day to day tasks. This could include having conversations with agents to find important pieces of information from (multi-modal) knowledge bases, interacting with Web service APIs, composing emails with suggested autocomplete, interacting with a car, excel files and/or web browsers to name a few. With the advent of large language models like ChatGPT, GPT-4 and in context learning, the community has started to move towards zero-shot generalization abilities that can further enable the development of a larger number of NLInt.

The aim of this workshop is to bring together researchers in both industry and academia who are interested in the advancement of NLInt. Historically, a lot of work on NLInts has happened within the industry, but this workshop aims to bring together both academic and industry researchers to combine their expertise and knowledge. Researchers who work in Natural Language Processing (NLP) and Human Computer Interaction (HCI) are welcome to discuss the existing issues of today’s NLInt and suggest improvements that will help in the advancement of the field. We welcome papers along both the modeling domains as well as the interfaces domain. For modeling, researchers could focus on highlighting the issues with existing NLP models/algorithms and/or suggest improvements both in terms of absolute performance as well as latency. For HCI, researchers can focus on the issues with the existing NLInt and discuss interesting new ideas to improve the user-experience of NL interfaces.


Dates

All deadlines are AOE(Anywhere on Earth).


Invited Speakers

Stay Tuned for Speaker line up!


Topics

This workshop aims to bring together researchers and practitioners from different communities related to NLIs. As such, the workshop welcomes and covers a wide range of topics around NLIs, including (non-exclusively):


Organizers

organizers

The workshop is organized by Vinayshekhar Bannihatti Kumar (AWS AI Labs), Sopan Khosla (AWS AI Labs)，Rashmi Gangadharaiah (AWS AI Labs), Scott Wen-tau Yih (META AI - FAIR), Ahmed Hassan Awadallah (Microsoft Research), Tania Bedrax-Weiss (Google Research), Dan Roth (AWS AI Labs) and Katrin Kirchoff (AWS AI Labs)

For any questions, please email nli.at.aacl@gmail.com


Affiliations

affiliations

Location

NLI 2023 will be held online, co-located with IJCNLP-AACL (conference registration required).