russir2018-health-search-tutorial

The RUSSIR 2018 Health Search Lecture Series

Welcome to the homepage for the Health Search Lecture Series at RUSSIR 2018.

Quick links:

• RUSSIR 2018 tutorial slides
• RUSSIR 2018 hands-on worksheet

Announcements during RUSSIR

• 31/08/2018: updated part 3 slides to remove content not covered in lecture, added link to additional word embeddings from online drug reviews (thanks to Elena Tutubalina); updated part 4 slide deck.
• 30/08/2018: updated part 2 slides to remove content not covered in lecture; uploaded part 3 and 4 slides plus minor updates.
• 29/08/2018: Corrected slides Part 1; uploaded slides part 2; updated practical part hand-out.
• 28/08/2018: Download the docker image we will use from Tuesday at https://hub.docker.com/r/ielabgroup/health-search-tutorial. Make sure you download before Tuesday afternoon.
• 27/08/2018: Note from Monday’s session: we arrived up to slide 32 of part 1 slide deck. We will cover the remaining material (Tasks in Health Search) during Tuesday’s lecture. After that, we will start looking at methods to address tasks and the problems in health search.

Motivation & Overview

With modern medicine increasingly reliant on information technology, the demand for IR systems that search medical content has grown significantly. The increasing need to retrieve medical advice (by both consumers and clinicians), and the adoption of electronic medical records are two factors driving the demand for health search. IR research has much to offer here by developing new tools and techniques specific to this domain.

The range of health information available - from electronic medical records, to medical literature, to health advice on the web - all leads to complex requirements that often require novel solutions to these different problems.

The key challenge in health search is how to bridge the sematic gap: the mismatch between the raw data and the way a human being interprets it. Although particularly prevalent in health search, the semantic gap problem is found in all domains; advances in health search can thus advance the whole field.

Key challenges include:

• How to leverage semantics and domain-knowledge resources for a better representation of documents and information need.
• What characterises relevance, in particular how topicality is complemented by other dimensions of relevance (understandability, authoritativeness, etc.).
• How bias and time pressure affects perception of relevance and decisions.
• How these influences the search process and evaluation.

Objectives of the Tutorial/Lecture Series

The main aims of RUSSIR HS2018 will be to:

• Summarise the basics of search in the health domain;
• Present the different end user requirements for multiple user groups interested in health search, including tasks;
• Provide an overview of the current use of IR techniques in the health domain;
• Provide a hands-on introduction to domain-specific tools which can be exploited in health search;
• Present resources and campaigns for evaluation in health search, including novel evaluation approaches;
• Present challenges and opportunities for further research in the health domain and discuss how these could be met.

This knowledge will allow IR researchers to identify promising ways of applying their work to the health domain, allowing them to contribute to a domain of rapidly growing importance.

Presenter Bio

Guido Zuccon

Guido is a senior lecturer within the School of Electrical Engineering and Computing Science at the Queensland University of Technology. His research interests include formal models of search and evaluation methods, in particular applied to health search. Guido has actively contributed to the areas of document ranking, search result diversification, formal models, and evaluation. Guido is the IR Task leader for the CLEF eHealth Evaluation Lab, a shared-task that aims to evaluate systems for consumer health search. He has already presented tutorials at SIGIR 2015, CIKM 2016 and ICTIR 2016 on formal models of search.

Guido received a Ph.D. in Computing Science from the University of Glasgow in 2012. Before joining the Queensland University of Technology as a lecturer in 2014, he was a postdoctoral research fellow at the CSIRO, Australia, working on health search technology at intersection of academic research and applied research with real customers.

In October 2018, Guido will join the University of Queensland.