Sunday, October 27, 2019

Bob Edlin: $13m granted to teach Siri to speak Te Reo

The Taxpayers Union promptly picked up on the spending of millions of dollars of public money, almost beating the Point of Order Trough Monitor to sound an alert.

The union focussed on just one of four projects to be funded from a trough labelled “Strategic Science Investment Fund”.  This project – to receive $13 million of taxpayers’ money – aims to teach Siri to speak Te Reo.

Siri (Wikepedia explains) is a virtual assistant that is part of Apple Inc.’s iOS, iPadOS, watchOS, macOS, tvOS and audioOS operating systems.   
Research, Science and Innovation Minister Megan Wood made special mention of Siri in her press statement.
The Taxpayers Union, however, challenged the wisdom of this spending in a press statement headed $13 million teaching Siri to use Te Reo is an IT boondoggle.
boondoggle (for those unfamiliar with the expression) is a project that is considered a waste of both time and money, yet is often continued due to extraneous policy or political motivations.
Dr Woods said:
“The four projects range from teaching Siri to speak Te Reo to crunching large environmental data sets collected via satellites.
“The projects were chosen on excellence, and will help New Zealand to address some of our greatest challenges like bringing Te Reo into everyday digital interactions, and building a picture of climatic and ecological trends,” Megan Woods says.
“To harness the benefits of data science, New Zealand needs to be at the forefront of emerging data science technologies. The new investment will significantly lift New Zealand’s capability, provide fresh and ambitious thinking to support international and national collaborations in cutting edge data science research.”
Collectively the programmes provide a wide coverage of data science research and make good use of unique New Zealand data sets, Dr Woods said.
“They have strong domestic and international collaborations involving a number of New Zealand universities and research organisations with strong links to end users.
“Greater use of advanced data science across the economy, environment and society will create new ways of working, helping to position New Zealand for a sustainable, prosperous and inclusive future.”
 The four successful programmes are:
  • A language platform for a multilingual Aotearoa: Starting with Te Reo. This will develop a multilingual language platform to that will enable New Zealanders to engage with technology in the language they use or aspire to use every day. World-leading data scientists from New Zealand, Cambridge University, Oxford University, Mozilla and Māori communities will work on this project in a unique collaboration. (Te Hiku Media and Dragonfly Data Science, $13 million over seven years).
  • A data-science driven evolution of aquaculture for building the blue economy: This will develop innovative data science techniques that will enable the aquaculture industry to produce high quality, low-carbon protein efficiently and at large scale, without impacting the environment. (Victoria University of Wellington, $13 million over 7 years)
  • Time-Evolving Data Science / Artificial Intelligence for Advanced Open Environmental Science: This will focus on developing new methods to deal with environmental datasets that are collected in large volumes over time, ranging from broad scale satellite images to single point measurements on the ground, in the water or air. (University of Waikato, $13 million over 7 years)
  • Beyond Prediction: explanatory and transparent data science for life and social sciences: This will develop new methods that discover, gather and integrate useful data with minimal human intervention. (University of Auckland, $10 million over 7 years).
Dr Woods explained that data science involves extracting useful information through different methods of recording, storing, and analysing data.
The Strategic Science Investment Fund’s purpose is to establish and support longer-term research programmes of mission-led science critical to the future of New Zealand’s economy, environment and wellbeing.
This is the first SSIF investment in a fund dedicated to data science research.
The Taxpayers Union said the $13 million research grant for Te Reo voice recognition technology is a poor use of taxpayer money.
It said the project is hoped to allow people to ask Siri or Alexa things like how to “find a choice as kai of panipopo” . This will “ensure a New Zealand identity is firmly embedded in the digital world”.
Taxpayers’ Union spokesman Louis Houlbrooke was dismissive:
“This is a classic case of taxpayer money being poured into a shiny, fashionable IT project that is a nice-to-have at best. These projects should be left to private companies who have an incentive to keep costs under control and develop technology that people actually use.
“Someone on the median wage would have to pay income tax for 1,500 years to cover the cost of this project. This money could have been used to provide more much-needed measles vaccines, or even tax relief.
“The entire Strategic Science Investment Fund appears to be rife with fuzzy spending initiatives and corporate welfare. This is the same fund that last year allocated $5.4 million for research into shoe leather. The fact Minister Megan Woods is proudly putting out press releases about this kind of spending leads us to wonder whether she has taxpayers’ interests at heart.”
The Point of Order Trough Monitor noted another aspect of the $13 million spend worthy of comment – the identities of the beneficiaries.
The project will be led by Te Hiku Media and Dragonfly Data Science.
Te Hiku Media, in Northland, is a charitable media organisation, collectively belonging to the Far North iwi of Ngāti Kuri, Te Aupouri, Ngai Takoto, Te Rārawa and Ngāti Kahu.
The station is an iwi communications hub for iwi radio, online TV and media services.
Māori language revitalisation is a core focus of Te Hiku Media, along with archiving and training.
So why didn’t the money come from the millions being invested in the revitalisation of Te Reo?
Dragonfly Data Science, based in Wellington, boasts expertise in statistical analysis.
Our speciality is modern Bayesian methods, which allow uncertainty to be carried all the way through from analysis to decision making.  
Our statistical work is supported by excellent technical database and software development skills. We develop operational systems to allow organisations to get ongoing value from their data. 
But Cambridge University, Oxford University and Mozilla will work on this project too.
This suggests we can’t support NZ scientists alone to do what must be done and will be sending money to the UK and the US to help grow the fortunes of their boffins.
There’s the prospects of some overseas junkets for the New Zealanders involved, of course, although this will be at the expense of carbon reduction.
What about the other projects?
With regard to aquaculture (Victoria University of Wellington, another $13 million over seven years), we wonder about the environmental implications.
Salmon farms – for example – are linked with the spread of parasites such as sea lice, while pesticides threaten other ocean creatures and plants. Antibiotics, used to keep disease at bay, leach into the surrounding environment. This boosts the risk of drug resistance and can give rise to so-called super-bacteria.
Stocks of wild fish harvested to feed salmon risk being depleted and salmon excretions can smother life on the sea floor.
Developing new methods to deal with environmental datasets that are collected in large volumes over time, ranging from broad scale satellite images to single point measurements on the ground, in the water or air, sounds attractive.  In this case, the University of Waikato gets $13 million over seven years.
We wonder if improved weather forecasting might be one of the consequences…
Maybe not.  NZ already has two State agencies – NIWA and MetService – competing to do the same job because competition – ha! – improves weather-forecasting efficiency.
That leaves the development of new methods that discover, gather and integrate useful data with minimal human intervention. (University of Auckland, $10 million over seen years. 
So the ProvinciaL Growth Fund aims to generate jobs; the Strategic Science Investment Fund invests in a project to reduce them.
Bob Edlin is a veteran journalist and editor for the Point of Order blog HERE.

No comments: