ADM+S research featured in leading search engine conference SIGIR 2023
Author Loren Dela Cruz
Date 24 August 2023
ADM+S PhD Students Marwah Alaofi and Kaixin Ji, and Associate Investigator Dr Damiano Spina from RMIT University recently presented their research on generating user-centred approaches to evaluating and understanding information retrieval and interaction at the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR).
SIGIR is the premier international forum for the latest developments and discussions in the Information Retrieval domain. In its 46th edition, SIGIR ’23 featured keynotes delivered by figures from both academia and industry, including representatives from tech companies such as Google and Microsoft, with a specific focus on the impact and the potential of Large Language Models (LLMs) in Information Retrieval research and development.
Damiano and Kaixin presented their work on understanding how our bodies, through pulse, sweat levels, and pupil size, are reacted and impacted during information-processing tasks such as reading, listening, speaking and typing. Their study informs machine learning models that train on this type of data. This research is a result of a collaboration with an ADM+S Research Fellow Dr Danula Hettiachchi and supervised by ADM+S Investigators Prof Flora Salim, Prof Falk Scholer and Dr Damiano Spina.
Read the paper: Examining the Impact of Uncontrolled Variables on Physiological Signals in User Studies for Information Processing Activities
Marwah presented her latest findings in using LLMs to generate synthetic search queries. Her research quantifies the similarity between human and LLM-generated queries and how they contribute to document pooling. Developed in collaboration with Luke Gallagher from RMIT University and supervised by ADM+S Investigators Prof Mark Sanderson and Prof Falk Scholer, and Paul Thomas from Microsoft, this research is part of Marwah’s ongoing research into understanding and measuring the impact of variability across users.
Read the paper: Can Generative LLMs Create Query Variants for Test Collections? An Exploratory Study
This research is supported by the Australian Research Council (CE200100005, DE200100064, DP180102687).
