Low Resource Language Understanding in Voice Assistants

Subendhu Rongali, Konstantine Arkoudas, Wael Hamza, Andrew McCallum
Last updated on Jan 24, 2023

This was my primary area of research with Prof. Andrew McCallum and the focus of my dissertation. We worked on improving low resource language understanding in voice assistants. This work was in collaboration with Konstantine Arkoudas and Wael Hamza at Amazon Alexa AI.

Our works in this project were published at WWW 2020, AAAI 2021, and IJCAI 2022, and the final piece of work on zeroshot domain adapation for semantic parsing will be published at EACL 2023.

Publications

Low-Resource Compositional Semantic Parsing with Concept Pretraining

Semantic parsing plays a key role in digital voice assistants such as Alexa, Siri, and Google Assistant by mapping natural language to …
Code Dataset Project

Low Resource Language Understanding in Voice Assistants

Voice assistants such as Amazon Alexa, Apple Siri, and Google Assistant have become ubiquitous. They rely on spoken language …

Preprint Project

Training Naturalized Semantic Parsers with Very Little Data

Semantic parsing is an important NLP problem, particularly for voice assistants such as Alexa and Google Assistant. State-of-the-art …
Dataset Project Appendix

Exploring Transfer Learning for End-to-End Spoken Language Understanding

NOTE: Our speech dataset couldn’t clear legal approval at Amazon due to unavailability of some of the annotators. We apologize …

Preprint PDF Dataset Project Poster Slides

Don't Parse, Generate! A Sequence to Sequence Architecture for Task-Oriented Semantic Parsing

Virtual assistants such as Amazon Alexa, Apple Siri, and Google Assistant often rely on a semantic parsing component to understand …
Preprint PDF Project Poster Slides