Dear {Contact_First_Name},
Last year we treated ourselves to a robot vacuum cleaner. It’s the best invention ever and now that I have one, I couldn’t be without it. At the time I was telling my mother how strange it felt to be sitting on the sofa reading a book while the vacuum cleaner was doing my housework. My mum told me that she had felt exactly the same when she got her first electric washing machine in the early sixties!
While I don’t think much AI (artificial intelligence) went into the washing machines of the sixties, it’s a different story for robot vacuum cleaners. The implementation of AI technologies for the robot vacuum scene is relatively new, but the market is expected to grow significantly over the next few years as consumers take advantage of AI technologies for vacuum cleaning, floor cleaning, and lawn mowing. Indeed, the uses for AI are manifold.
Helmut Küchenhoff, Professor of Statistics in the Faculty of Mathematics, Computer Science and Statistics at LMU Munich, goes as far as to say that AI may have its place in political context: “In complex political situations, artificial intelligence can play an important role. AI systems can support human decision-making, for example, by computing scenarios and forecasts without taking the actual decision away from the humans involved”. (Mirage, 3 February 2022).
He goes on to explain that AI and machine learning are already making significant contributions to climate science. He also sees potential for the use of AI in health research, though questions around data quality and goal setting remain. Reliable training data and clearly defined criteria for the goals to be met are essential for AI implementation to be successful.
If you don’t mind sub-titles you can hear (read?) his talk here.
If you are keen on the subject of AI, don’t miss Dr Mariarosaria Taddeo’s talk “AI and Data Science for Social Good” tonight, 8pm (AEDT) as part of Monash-Prato dialogue. The Monash Prato Dialogue is a Distinguished Lecture series on Artificial Intelligence and its impact on society. Marioarosaria will present principles and recommendations to support a governance strategy to leverage the opportunities offered by AI for the climate crisis in a responsive, evidence-based, and ethically sound manner.
Details of the event are available below.
Kind regards
Marie-Louise Rankin
SSA Executive Officer
Read newsletter in your browser |
|
|
64th World Statistics Congress The Chair of the Scientific Programme Committee (SPC), Mark Podolskij, invites the statistical community to prepare proposals for the Invited Paper Sessions (IPS) programme.
The 64th WSC 2023, to be held 16–20 July in Ottawa, Canada, will highlight the developments and contributions of statistics, statistical science, and data science in all aspects of life, particularly the well-being and welfare of people. The WSC 2023 will host talks and presentations on a wide variety of topics, with the overall goal of presenting a balanced programme that provides a sense of the current state and future direction of statistics and their applications.
Read the full announcement on the WSC 2023 website, as well as the Guidelines for Session Organisers, Chairs, Presenters and Discussants.
General enquiries about the Scientific Programme should be directed to scientific@isi2023.org. |
|
|
|
|
David Spiegelhalter on "Desert Island Discs"
Legendary statistician David Spiegelhalter was interviewed last week on Desert Island Discs, a long-running UK radio show on BBC Radio 4, where guests choose the eight tracks and book they would take with them if they were marooned on a desert island. It is a huge honour to be on the show and a mark of national treasure status. It’s a fascinating insight into David’s life and family, and he also talks about the tremendous work he’s done on communicating difficult statistical concepts to the public. In theme with the show, they’ve even a story about real life pirates!
Amongst other things, David revealed that in the early days of the COVID-19 pandemic he had felt overly optimistic about the outcome. He admitted that he “didn’t take it seriously enough” and conceded that he probably would not make a good government advisor. He said: “I think it’s very important that we have to acknowledge that we can never take an objective view about evidence. We always bring our, I think, personalities into it, and mine is unfortunately very optimistic and that’s why I’m very glad I’m not a government advisor. I don’t think I’d be very good at it because I do tend to hope for the best and sort of expect the best as well.”
While not much good can be said about the pandemic, Spiegelhalter did point out that for a certain demographic in the UK COVID-19 actually did save lives. “If you look at people between 15 and 30 in 2020, 300 fewer died than would normally have died, and that includes the 100 that died from Covid, sadly." The lock-downs protected people in more ways than one. |
|
|
|
|
RSS International Conference 2022 - Submissions are open for contributed talks and posters to the RSS International Conference 2022, 12-15 September 2022, Aberdeen.
Submit on any topic related to statistics and data science. Initial deadline 5 April.
The SPC welcomes submissions on any topic related to statistics, data science and the use of data. Not only are they keen to have presentations on novel methodology and advances in analytical techniques, but also on the interesting and impactful stories which can be told via the use of statistics and data science.
Submissions are now open for the following types of presentation, with the relevant deadlines:
You are asked to refer to the guidelines for submissions before commencing your submission. The guidelines include information about the format and content of the abstract.
Please note:
Please do not make a submission if you are not able to meet these criteria. |
|
|
|
|
Open letter from SSA past presidents Dennis Trewin and Nick Fisher: Have statisticians and, by implication the SSA, failed their civic duty to promote good statistical practice, in relation to the pandemic?
The Covid-19 pandemic is the biggest challenge that Australian Governments have faced in decades. It is one where statistical science could make an incredibly valuable contribution. However, our professional society has been largely silent, seemingly not wanting to get involved in public policy.
Certainly, individual statisticians have been involved and made some very important contributions. There have been some Covid-related talks at Branch meetings (but not many) but this is very different from the role played by the Royal Statistical Society (RSS) for example. The RSS set up a Covid-19 Task Force, released a series of statements on the statistical aspects of new Covid-19 policy initiatives, prepared briefing notes on Covid statistics for public consumption, made submissions to public inquiries, engaged with the modellers and hosted major lectures.
We believe a 21st professional society must have a policy and advocacy role. We are a small organisation and must be selective as to where we get involved but surely Covid-19 was sufficiently important to warrant special SSA input. The SSA needs to establish some principles on when and how to get involved. There is a lot to be learnt from the RSS and the American Statistical Association who have also been very active in public policy interventions.
We tried to gain some publicity about what statisticians had to offer by hosting a Zoom workshop last August, following the release of the report of the Doherty Institute that informed the National Cabinet decisions for ‘opening up’ when vaccination rates reached certain levels. Whilst the Doherty report is a fine report in many respects, epidemiologists are not statisticians and do not have the same understanding of appropriate ways to cope with uncertainty. This led to issues when Doherty had to make assumptions on the effectiveness of vaccines, public health interventions and the test and trace systems with rather scant data and uncertainty was clearly prevalent.
We put a lot of effort into this workshop, and an agreed statement of Findings, so were deeply disappointed when the SSA Executive decided it was not the place of the Society to comment on this work in the absence of allegations of misconduct. Surely we should be prepared to make a contribution when we can add value to public policy in a constructive way. In fact, the workshop was rather prescient in its findings. - Two of the main findings were: The level of public health measures should be based on both vaccination levels and the real time effective reproduction number. As it happened, public health measures were being reduced at the same time as the effective reproduction number was increasing.
- Data from surveillance testing do not provide an accurate picture of the state of the pandemic and this is likely to become a bigger problem as contact tracing becomes only partially effective. An ongoing national scientific random testing process needs to be put in place to properly understand the progress of the virus.
We are Past Presidents and Life Members of the SSA and have been professionally accredited since the AStat program was introduced. We want the SSA to prosper. It is becoming increasingly obvious that Australia needs clear, disinterested and well-informed professional statistical contributions to matters of public interest, such as how to manage a pandemic. If the SSA chooses not to provide such contributions, it will only have itself to blame if another body far less well-equipped provides such opinions. We urge the SSA to introduce a suitable process as a matter of urgency. It is ultimately a question of acting in a responsible way on behalf of the profession.
Nicholas Fisher
Dennis Trewin |
|
|
|
|
Response by Jessica Kasza and Adrian Barnett (SSA President and Immediate Past President): The SSA welcomes constructive criticism and feedback from our members, such as that contained in the letter from our past Presidents Nick Fisher and Dennis Trewin.
Over the past two years there has been continued discussion within the Society’s Executive Committee around the role that the Society has to play in the COVID-19 pandemic. This discussion has considered those issues included in the letter and several issues that were not mentioned therein. These issues included the fact that several of our members were involved in groups working on the modelling for the Government. Thus, criticism of this modelling would involve criticism of our members – and in this case, such criticism was not thought to be productive. Nor was criticism of the qualifications of individual members thought to be appropriate (e.g. the question of epidemiologists vs. statisticians). Further, in discussions with some of our members who were involved, it was clear that much additional modelling was being done that was not appearing in Governmental reports.
We agree that the Royal Statistical Society has done a remarkable job in providing public commentary on the British Government’s handling of the COVID-19 pandemic. There are two key issues to bear in mind. The first is that the Royal Statistical Society is a much larger society than the SSA, with far more resources available to them. The second is that much of the public commentary provided by the Royal Statistical Society was through Sir David Spiegelhalter, who holds a position specifically focussed on enhancing the public understanding of risk.
A key role that the SSA can play in contributing to matters of public interest is to empower our members in their communication of complex statistical ideas. The SSA can do that by recognising and advocating for such work, and by providing training in statistical communication: an integral part of the statistician’s toolkit. Over the past two years the SSA has run several workshops aimed specifically at this, and we intend to continue offering such workshops to our members.
While the SSA does not wish to debate the use of particular analysis methods or statistical models, the Society does advocate for the interests of our members, and will continue to do so. Recent examples of such advocacy include lobbying the ARC and the Government about the use of preprints in funding applications and about political interference in the award of grants. |
|
|
|
|
Statistics are vital: An interview with Manira Ahmad
'Statistics Are Vital' is a joint campaign by the Royal Statistical Society and Significance magazine, celebrating the work of statisticians during the Covid-19 pandemic.
Recently, the RSS spoke to Manira Ahmad, former Head of local intelligence support (LIST) at Public Health Scotland (now Chief Officer at PHS).
Manira and the team at PHS produce insights relating to health, care and community to support improved local decision making and redesign of services. When the pandemic struck in 2020, they had to quickly respond to the needs of frontline services and provide the new data that was urgently needed. How did they do this? |
|
|
|
|
|
|
Scientific Programming Committee members needed for the 2023 Australian Statistical Conference
The Scientific Programming Committee for the 2023 Australian Statistical Conference, co-Chaired by Dr Nicole White (QUT) and Dr Karen Lamb (University of Melbourne), is seeking new members to help plan the program for the conference taking place in December 2023. The conference theme is ‘Statisticians in society’, focussing on the key role statisticians play in communication across diverse areas that are key to our society.
Interested? Please email Nicole White (nm.white@qut.edu.au) by 25th February to tell us a little more about yourself, including brief details of i) the statistical section you align most with (e.g., biostatistics, environmental, official), ii) employment sector (e.g., academic, industry, government), iii) how long you have been working as a statistician, iv) what prior experience you have in conference or event organisation, and v) what you will bring to the committee. Please note that no prior experience in conference organisation is required. We would love to hear from people keen to get more involved in the SSA and strongly encourage people who may not have participated in SSA activities before to reach out (including early career statisticians). We look forward to hearing from you! |
|
|
SSA NHMRC Investigator Grant Fellowship Top-Up Awards
Are you applying for an NHMRC Investigator Grant at Emerging Leadership Level 1 or 2 this year? To help our early-career members in this highly competitive scheme, we are awarding up to five SSA Fellowship Top-Up Awards, worth $3000 each, to support their applications. A panel of senior members and recent fellowship winners will select up to five winners. Please see the website here for details and an application form. |
|
|
|
|
Monash Prato Dialogue: “AI and Data Science for Social Good”
Thursday, 10 February, 8pm-9pm AEDT (7pm AEST), Monash Data Futures Institute
In this lecture, Dr Mariarosaria Taddeo, from the Oxford Internet Institute at the University of Oxford, will present principles and recommendations to support a governance strategy to leverage the opportunities offered by AI for the climate crisis in a responsive, evidence-based, and ethically sound manner.
The Monash Prato Dialogue is a Distinguished Lecture series on Artificial Intelligence and its impact on society. |
|
|
|
|
|
|
ADSN Data Ethics Workshop #2 - Prof. Rachel Thomas Thursday, 24 February, 1 pm AEDT (12 pm AEST)
This is the second of two ADSN (Australian Data Science Network) data ethics workshops. The aim of this series is to share open questions, spark discussion, and facilitate collaborations. The workshop will be led by Professor Rachel Thomas, Professor of Practice at the QUT Centre for Data Science and Co-founder of fast.ai. Our guests for this workshop are:
Catarina Pinto Moreira (QUT): “An Interpretable Probabilistic Approach for Demystifying Black-box Predictive Models” Yves Saint James Aquino (Univ of Wollongong): “The algorithm will see you now: ethical, legal and social implications of adopting machine learning systems for diagnosis and screening” Michael Evans (Queensland AI Hub): “The Role of Policy in Data Ethics” |
|
|
|
|
|
|
MACH EMCR Research Design Webinar Series
MACH (Melbourne Academic Centre for Health) is pleased to present a four-part webinar series exploring research design tailored specifically for early and mid-career researchers. The series will be conducted by experts from the University of Melbourne’s Methods and Implementation Support for Clinical and Health research Hub (MISCH) team from February-June 2022. The SSA is proud to be able to call three of the presenters below accredited members of the Society.
Seminar 1: Kicking off your research: how to craft a well-defined research question
Presenter: Dr Karen Lamb (Uni Melbourne)
Wednesday, 23rd February 2022, 12.30-1.30pm AEDT, online
Register here
Seminar 2: What’s it going to take to get your study started? Pilot and Feasibility studies
Presenter: Ms Sabine Braat (Uni Melbourne)
Wednesday, 6th April 2022 12:30-1:30pm AEDT, online
Register here
Seminar 3: Statistics for your grant applications
Presenter: Prof Julie Simpson (Uni Melbourne)
Wednesday, 11th May 2022 12:30-1:30pm AEDT, online
Register here
Seminar 4: Economic evaluation alongside clinical trials: principles of study design and decision analysis
Presenter: Dr An Duy Tran (Uni Melbourne)
Wednesday, 22nd June 2022, 12:30-1:30pm AEDT, online
Register here |
|
|
|
|
Statistics for Research Workers using R and R Markdown
7 Mar 2022 (AEDT) – 1 Apr 2022 (AEDT), online
The Statistical Consulting Centre at the University of Melbourne is offering Statistics for Research Workers using R and R Markdown.
This course provides an introduction to foundational statistical methods and ideas used throughout statistics and data science. It uses R statistical software with RStudio and R Markdown.
The expected time commitment is 3 hours a day for 4 days each week. The course runs for four weeks. |
|
|
|
|
|
|
SSA Workshop: Introduction to Machine Learning for Health Data
12 -13 May 2022, 9:00 AM - 5:00 PM each day, Adelaide
Craving an in-person event? SSA is proudly offering a 2 day course with presenter Dr Oscar Perez-Concha, Centre for Big Data Research in Health, UNSW Sydney.
This workshop introduces the basics for understanding and using machine learning algorithms. We will discuss the machine learning workflow, from clearly defining our research question to the rationale behind choosing different machine learning techniques for different scenarios, highlighting questions such as overfitting/ under-fitting, missing data, and interpretability. We will focus on the principles behind some of the most used supervised learning algorithms. However, the detailed mathematics underlying these algorithms will not be discussed. Real healthcare scenarios using Python will be presented.
Dr Oscar Perez-Concha is a health data scientist with over 15 years’ experience in machine learning and statistical modelling. |
|
|
|
|
If you have news from the Australian statistical community to share in Stats Matters and Events, please get in touch with us! We love getting feedback too.
|
|
|
Statistical Society of Australia | PO Box 213 Belconnen ACT 2616 Australia 02 6251 3647 | www.statsoc.org.au |
|
|
|
