Would you like to help NICTA as part of a trial in the ACT?
We want you!
NICTA’s Machine Learning Research Group is teaming up with solar energy experts from ANU to develop ways of predicting the expected power output from rooftop solar energy systems.
We are aiming to develop software that will forecast fine-grained predictions of solar output using low cost devices installed on individual homes that can be deployed all across Australia. In order for this to be possible, NICTA needs to collect large amounts of data from regions all over the ACT. We need your help to do this.
What does the trial involve?
By letting NICTA install a low-cost, non-invasive monitoring device with your solar inverter system, we can build the dataset needed to develop solar forecasting algorithms. We need 2-4 households per suburb in the ACT to sign up as volunteers for this two-year project. If you have a solar power system installed on your rooftop, or are in the process of installing one, then please consider volunteering for this project. The monitoring device and installation costs will be borne by NICTA.
For more information on the project, visit the project page: https://www.nicta.com.au/research/machine_learning/projects/distributed_solar_prediction
To sign up, please register your details via the Solar Monitoring Portal: https://www.nicta.com.au/business/security_and_environment/solar_monitoring_portal.
If you have any questions, send the project team an email at email@example.com.
The paper "Gaussian Process Factorization Machines for Context-aware Recommendations" by Trung V. Nguyen, Alexandros Karatzoglou, and Linas Baltrunas has been accepted for publication at SIGIR'14. Part of the work was done when Trung was an intern with Telefonica R&D.
Hanna Suominen short-listed for the 2013 L'Oréal Australia and New Zealand For Women In Science Fellowship
A big congratulations to Hanna Suominen who has been shortlisted for a 2013 L'Oréal Australia and New Zealand For Women In Science Fellowship. Out of 230 applications, the judges have selected a shortlist of 34 candidates from which they will select the three 2013 fellows. The very best of luck to our Hanna!
Bob Williamson was invited to present on Machine Learning at the DBCDE meeting
Prof Bob Williamson presented a talk on Machine Learning at the Department of Broadband, Communications and the Digital Economy meeting on 16 May.
Results for the first two tasks of CLEFeHealth have been published
The CLEF evaluation lab is a benchmarking activity aiming to develop processing methods and resources to enrich difficult-to-understand health text as well as their evaluation setting. It consists of an evaluation campaign and one-day workshop in CLEF2013 organised in Valencia, Spain on 23-26 September, 2013. Hanna Suominen from NICTA CRL is co-chairing the lab.
Task 1a: Boundary detection of disorder named entities
Twenty-one teams from Australia, China, France, India, Ireland, Republic of Korea, Spain, and the USA participated.
Task 1b: Named entity recognition and normalization of SNOMED disorders
Sixteen teams from Australia, China, Ireland, Spain, and the USA participated.
Task 2: Normalization of acronyms/abbreviations to UMLS codes
Five teams from Australia, China, France, and the USA participated.
Research Cooperation - Tsinghua University
A team from ANU and NICTA including Lexing Xie, Scott Sanner, Mark Reid and Wray Buntine visited Tsinghua University on 7-8th May to promote research cooperation. Some future student visits are planned.
Aaron Defazio offered to join Microsoft Research Cambridge
Congratulations to Machine Learning PhD student Aaron Defazio who got accepted for a fixed-term position of Research Intern with Microsoft Research Cambridge.
National Youth Science Forum
February - March 2013
Julien Epps receives funding from the ARC for a Discovery Project
A team including NICTA ML's Julien Epps has successfully secured $360K from the ARC Discovery Projects scheme, for a project entitled 'Affective Sensing Technology for the Detection and Monitoring of Depression and Melancholia', running 2013-2016. Led by Dr Roland Goecke (University of Canberra) and including collaborators from the Black Dog Institute, UNSW and CMU, the project will develop affective sensing technology to study affect in facial expressions and speech, and evaluate it by developing the first automated, objective measure of depressive disorders.
This funding augments existing efforts by the same team (under NHMRC and other funding) since before 2011, and by providing a focal point for the affective sensing technology, capitalises on the unique data resources and expertise of the Black Dog Institute and helps establish the team among leaders in the emerging area of objective measures of mental health. ML PhD student Nicholas Cummins, whose topic falls directly within the scope of the project, has already made key advances. The speech-based depression measures build on earlier NICTA research into cognitive load estimation from speech.
Trung Nguyen and Edwin Bonilla – Papers accepted to AISTATS 2013
Machine Learning's Trung Nguyen and Edwin Bonilla had their paper entitled 'Efficient Variational Inference for Gaussian Process Regression Networks' accepted to the 16th International Conference on Artificial Intelligence and Statistics (AISTATS) 2013, for an oral presentation. Congratulations!
Scott Sanner to present tutorial at AAAI-13
Well done to Scott Sanner who had a four hour tutorial titled "Advances in Symbolic Methods for Inference, Optimization, and Decision-making" accepted for presentation at AAAI-13 in Seattle, USA. AAAI is one of the top two international AI conferences, along with IJCAI.
Wei Liu – AI 2012 Best Paper Award
Congratulations to Wei Liu for receiving a Best Paper Award at the AI 2012 Conference for his paper entitled ‘Efficient Adversarial Learning Strategy for Constructing Robust Classification Boundaries’. This paper is a joint work with Sanjay Chawla, James Bailey, Christopher Leckie and Rao Kotagiri. The award will be received at the conference in early December.
Fifth Asian Conference on Machine Learning (ACML) 2013 will be held in Canberra
ACML 2012 in Singapore ran from 4-6th November, had Machine Learning's Bob Williamson as invited speaker and Wray Buntine as Programme Co-Chair.
It was also announced that ACML 2013 will be hosted in November at ANU in Canberra run by NICTA and ANU
Discovery Early Career Researcher Award to Mark Reid
Congratulations to Mark Reid for receiving a three year Discovery Early Career Researcher Award for a project titled "Composing Machine Learning via Market Mechanisms".
This project aims to better understand connections between learning algorithms and markets as aggregators of information and develop new, principled techniques for combining predictions. This will improve our ability to construct systems that make predictions based on multiple, complex and structured sources of data.
Scott Sanner receives a 'Top Supervisor Award"
Congratulations to Scott Sanner for receiving a Top Supervisor Award signed by Prof Mandy Thomas from the ANU.
Commonwealth ICT Senior Executive Forum - Prof Bob Williamson invited as a guest speaker to talk about Machine Learning.
Prof Bob Williamson is invited to the second Commonwealth ICT Senior Executive Forum for 2012, to be held in Canberra on 29th of August. He will talk about Machine Learning, which is the science of making sense of data.
Scott Sanner to give invited talk at the Brazilian Conference on Intelligent Systems in October.
Congratulations to Scott Sanner who is an invited speaker at Brazilian Conference on Intelligent Systems in October:
Title: Recent Advances in Continuous Planning (Tutorial) and Data Structures for Efficient Inference and Optimization in Expressive Continuous Domains (Talk).
AOARD Grant Awarded to Fang Chen.
Congratulations to Fang Chen for receiving a $120K AOARD grant for ‘Robust Multimodal Cognitive Load Measurement’ project. Novel fusion methodologies for integrating cognitive load information from multiple sources in a manner that optimizes their collective use will be researched on the grant.
In-hospital surveillance of invasive fungal infections.
A video about text mining system for in-hospital surveillance of invasive fungal infections released by the eHealth/Machine Learning group for the HIC Health Innformatics Conference 2012.
Wray Buntine talks to the ABC World Today about AI.
See the next entry. ABC World Today story is here.
How Many Computers to Identify a Cat? 16,000.
NICTA Graduate, Quoc Le, makes the New York Times with his ground breaking research on simulation of the human brain to learn about images.
Scott Sanner awarded prize for reviewing at AAAI.
Scott Sanner awarded "Outstanding PC Award" for his AAAI 2012 conference reviewing.
Fang Chen's group reported in Sydney Morning Herald.
"Technology gives voice to worker brain strain". Read more
6 papers accepted at the International Conference on Machine Learning.
NICTA Machine Learning Research Group gets 6 papers accepted at the International Conference on Machine Learning, to be held June 2012, the leading international machine learning conference.
PhD student Mahsa Salehi member of winning Stethocloud team.
Mahsa Salehi was part of a team that won the Australian final of the Microsoft ImagineCup competition for their electronic stethoscope application, which aims to detect pneumonia in children. "Smart stethoscope on cloud nine". Read more
NICTA leads $5M geothermal ‘big data’ analytics initiative involving the use of machine learning.
NICTA’s robust research capabilities in machine learning and in the increasingly significant area of big data analytics will be used to locate and characterise geothermal energy sources deep beneath the surface of the Earth. See the media release. Also published in The Register.
Professor Bob Williamson awarded top Australian science honour.
NICTA's Machine Learning Leader Prof Bob Williamson, was elected as a Fellow of the Australian Academy of Science in recognition of his research in developing powerful new methods and theoretical arguments for analysing data.
NICTA's yearly cornerstone event.
We host our annual research and innovation event, Techfest.
NICTA scored within the top six teams in the Text Retrieval Conference, Medical Records Track, Washington DC, USA.
Bob Williamson gave an invited talk at the Max Plank Institute.
Bob talk "Theory of Loss Functions" was held at the Empirical Inference Symposium in honour of Vapnik's 75th birthday.
Scott Sanner co-organises a workshop at Neural Information Processing Systems 2011.
Wray Buntine presents the ALTA 2011 Keynote, "Discovery in Text: Visualisation, Topics and Statistics".
Slides from Wray's ALTA 2011 Keynote,"Discovery in Text: Visualisation, Topics and Statistics".November 2011
ML researchers co-organise a workshop on "Inferring Cognitive and Emotional States from Multimodal Measures"
Fang Chen, Julien Epps, Natalie Ruiz, Eric Choi and and Ronnie Taib jointly organise this workshop as part of the International Conference on Mulltmodal Interfaces 2011.
Wray Buntine appointed co-chair of Asian Machine Learning Conference (ACML).
The 4th Asian Conference on Machine Learning (ACML2012) will be held on November 4-6, 2012, at the Singapore Management University, Singapore
Lawrence Cavedon tells about predictive technology in eHealth, "University adopts predictive technology".
"At NICTA, we are investigating how to perform text analytics (rather than data analytics) across different types of text documents, from tweets to research papers", says Lawrence. Read more at the Sydney Morning Herald.
William Billingsley awarded finalist in Education section for The Australian Innovation Challenge.
Hanna Suominen gives invited talk to iDASH Privacy Workshop in California, USA.
Hanna's talk title "Developing Machine Learning Methods and Language Technologies for Health and Wellness: Experiences Related to Data Dissemination from Australia and the Nordic and Baltic countries" - See the iDASH workshop.
Scott Sanner co-organises the "European Workshop on Reinforcement Learning".
This event was co-located with ECML PKDD 2011.
NICTA’s BrainGauge technology receives a Highly Commended Award.
At the 2011 Sydney Engineering Excellence Awards in the Innovations and Inventions categor, Fang Chen's research project.
January 2011 - August 2011
Hanna Suominen interviewed for the ABC 7pm TV News.
Hanna talks about text classification system for crisis management
Tiberio Caetano co-organised the CVPR Workshop.
The CVPR workshop “Inference in Graphical Models with Structured Potentials’’ held in June 2011.
NICTA’s BrainGauge technology led by Fang Chen and Bo Yin, was a finalist, Innovation Award, CeBIT Australia 2011.
Fang Chen was appointed as Associate Chair for the 2012 International Conference on Intelligent User Interfaces.
View conference details
A joint partnership between NICTA and the ANU
This collaboration included Scott Sanner, Edwin Bonilla and ANU's Peter Christen and Lexing Xie. The team received a Google Grant to carry out social recommendation research on Facebook.
Funding for a joint research.
Project tiled “Enhancing communication in online meetings using multimodal cognitive load measurement”from the National Institute of Informatics (NII), Japan, has won a colloborative grant. The value of the grant is 3 million Japanese Yen for three years (~AUD$37K).
Project “Effects of cognitive load on trust” was accepted by the U.S. Air Force Office of Scientific Research
This project proposal received a funding of US$58,000 per annum.
Wray Buntine organised the 2011 Machine Learning Summer School in Singapore with Wee Sun Lee of the National University of Singapore.
The summer school is part of the machine learning summer school series started in 2002. 2011 Machine Learning Summer School, held 13-17 June in Singapore was a success with over 300 graduate level students and industry participating (June 11).
Joachim Gudmundsson - new appointment
Joachim Gudmundsson got appointed Future Fellow (for 5 years) by ARC in 2011.
Fang Chen's group and NSW Roads and Traffic Authority are embarking on a joint research.
This research agreement will develop a system that would not only improve the management of traffic incidents, but also manage the “cognitive” load of its operators.