|
Data to Decisions CRC brings together leading edge Research & Development, Innovation and Education & Training to tackle Big Data challenges in Australia. This seminar will include two D2D CRC PhD students, John Calvo Martinez and Xuanyi Dong, outlining and presenting findings from their research.
Who's it for?
Anyone who is interested in learning about event extraction over multiple text streams and efficient neural architecture search.
Our presenters
Xuanyi Dong is a second-year PhD candidate at the Centre for Artificial Intelligence, University of Technology Sydney (UTS). He has over two years of industrial experience working on computer vision projects and has published 12 papers to date. Xuanyis focus is on automated machine learning, especially its application for computer vision.
John Calvo Martinez is a PhD Candidate with 10 years of experience in Data Science and ICT project implementations. He has experience through Data Architect roles in consulting and multi-national companies, extending from Data Engineering and Visualization, to Data Mining and real time NLP. Johns research focus is on distributed stream mining.
What's it about?
Xuanyi will present on his research into Efficient Neural Architecture Search. This presentation will explore representation learning, which is a fundamental research problem. Representation learning is fundamental because it benefits downstream computer vision applications, such as detection and segmentation however because it has undergone a transition from feature engineering to architecture engineering, expert knowledge and computational resources are needed to secure architecture for good feature representations. This presentation will provide insight into algorithm design to discover robust architectures to save GPU resources and increase speed.
John will present on his research into Event Extraction over Multiple Text Streams. Event extraction is the complex task to identify, extract and categorise information from raw text data. The goal is to automatically recognise events (what happened to who and when) from text, extracting persons, organisations, locations and dates from raw sentences. Specifically, the aim of the research is to create novel machine learning techniques to deal with event extraction, text categorisation and entity classification in real time, by using stream mining frameworks. This presentation will explore ways to extract such events using real-time machine-learning as well as how to increase accuracies when dealing with multiple sources of information. It will also discuss a novel rule learning approach made to deal with wrong event combinations.
|
|
 |
LocationBase64 (View)
64 North Terrace
Kent Town 6067
Australia
Categories
| Kid Friendly: No |
| Dog Friendly: No |
| Non-Smoking: Yes! |
| Wheelchair Accessible: Yes! |
|
Contact
|
