A5002711436- Topic Modeling
- Natural Language Processing Techniques
- Information Retrieval and Search Behavior
- Complex Network Analysis Techniques
- Expert finding and Q&A systems
- ICT in Developing Communities
- Speech and dialogue systems
- Text and Document Classification Technologies
- Sentiment Analysis and Opinion Mining
- Advanced Text Analysis Techniques
- Semantic Web and Ontologies
- AI in Service Interactions
- Multimodal Machine Learning Applications
- Human Mobility and Location-Based Analysis
- Opinion Dynamics and Social Influence
- Spam and Phishing Detection
- Mental Health via Writing
- Social Media and Politics
- Health Literacy and Information Accessibility
- Child Development and Digital Technology
- Meta-analysis and systematic reviews
- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Geographic Information Systems Studies
- Biomedical Text Mining and Ontologies
The University of Queensland
2021-2023
Qatar Foundation
2021
RMIT University
2015-2020
The use of social media has profoundly affected and political dynamics in the Arab world. In this paper, we explore Arabic microblogs retrieval. We illustrate some challenges associated with microblog retrieval, which mainly stem from different dialects that vary lexical selection, morphology, phonetics lack orthographic spelling conventions. present required processing for effective retrieval such as improved letter normalization, elongated word handling, stopword removal, stemming
A reader of a news article would often be interested in the comments other readers on an article, because give insight into popular opinions or feelings toward given piece news. In recent years, social media platforms, such as Twitter, have become hub for users to communicate and express their thoughts. This includes sharing articles commenting them. this paper, we propose approach identifying “comment-tweets” that comment articles. We discuss nature comment-tweets compare them subjective...
Pseudo Relevance Feedback (PRF) is known to improve the effectiveness of bag-of-words retrievers. At same time, deep language models have been shown outperform traditional rerankers. However, it unclear how integrate PRF directly with emergent models. This article addresses this gap by investigating methods for integrating signals rerankers and dense retrievers based on We consider text-based, vector-based hybrid approaches investigate different ways combining scoring relevance signals. An...
The Future Conversations workshop at CHIIR'21 looked to the future of search, recommendation, and information interaction ask: where are opportunities for conversational interactions? What do we need get there? Furthermore, who stands benefit? was hands-on interactive. Rather than a series technical talks, solicited position statements on opportunities, problems, solutions in search all modalities (written, spoken, or multimodal). This paper -co-authored by organisers participants workshop-...
Current pre-trained language model approaches to information retrieval can be broadly divided into two categories: sparse retrievers (to which belong also non-neural such as bag-of-words methods, e.g., BM25) and dense retrievers. Each of these categories appears capture different characteristics relevance. Previous work has investigated how relevance signals from could combined with those via interpolation. Such interpolation would generally lead higher effectiveness.
Increasingly, people go online to seek health advice. They commonly use the symptoms they are experiencing identify conditions may have (self-diagnosis task) as well determine an appropriate action take (triaging task); e.g., should emergent medical attention or attempt treat themselves at home? This paper investigates effectiveness of two most common methods for self-diagnosis and triaging: symptom checkers traditional web search engines. To this end, we conducted a user study with 64...
Geolocating Twitter users—the task of identifying their home locations—serves a wide range community and business applications such as managing natural crises, journalism, public health. Many approaches have been proposed for automatically geolocating users based on tweets; at the same time, various evaluation metrics to measure effectiveness these approaches, making it challenging understand which is most suitable this task. In article, we propose guide standardized user geolocation by...
Pseudo-Relevance Feedback (PRF) assumes that the top results retrieved by a first-stage ranker are relevant to original query and uses them improve representation for second round of retrieval. This assumption however is often not correct: some or even all feedback documents may be irrelevant. Indeed, effectiveness PRF methods well depend on quality signal thus ranker. aspect has received little attention before.
We analyze fifteen Twitter user geolocation models and two baselines comparing how they are evaluated. Our results demonstrate that the choice of effectiveness metric can have a substantial impact on conclusions drawn from an experiment. show for general evaluations, range metrics should be reported to ensure complete picture system is conveyed.
Pseudo Relevance Feedback (PRF) is known to improve the effectiveness of bag-of-words retrievers. At same time, deep language models have been shown outperform traditional rerankers. However, it unclear how integrate PRF directly with emergent models. In this article, we address gap by investigating methods for integrating signals into rerankers and dense retrievers based on We consider text-based vector-based approaches, investigate different ways combining scoring relevance signals. An...
We present SCC, a test collection for evaluating search in chat conversations. Chat applications such as Slack, WhatsApp and Wechat have become popular communication methods. Typical requirements these revolve around the task of known item retrieval, i.e. find information that user has previously experienced their chats. However, capabilities are often very basic. Our aims to support new research into building effective methods conversations search. do so by with 114 retrieval topics...
While large amounts of potentially useful agricultural resources (journal articles, manuals, reports) are available, their value cannot be realised if they easily searched and presented to the agriculture users in a digestible form.AgAsk is conversational search system for domain, providing tailored answers growers questions. AgAsk underpinned by an efficient effective neural passage ranking model fine-tuned on real world growers' An adaptable, messaging-style user interface deployed via...