If it's going to rain within the next 90 mins, this very British AI system can warn you
- Reference: 1632983709
- News link: https://www.theregister.co.uk/2021/09/30/ai_model_rain_deepminnd/
- Source link:
Traditional forecasting methods rely on solving complex equations that take into account various weather conditions, such as air pressure, moisture, and the temperature of Earth’s atmosphere. The trouble is, at least in Blighty, these systems tend to [1]predict what lies in store for us whole days or weeks ahead.
Deep-learning models are better suited for making more near-term forecasts – such as within the next couple of hours – according to a paper [2]published by the aforementioned boffins in Nature on Wednesday. There are advantages to using AI algorithms; they don’t have to solve thermodynamic equations and are less computationally intensive than other predictive techniques.
[3]
The team led by DeepMind trained a generative adversarial network (GAN) to produce a sequence of maps indicating where it's going to rain. Each of these precipitation maps show where moisture is accruing and moving in the atmosphere, each one covering a region measuring 1,536 × 1,280 km. Millions of examples of these maps were gathered from radar observations from 2016 to 2018. Data gathered in 2019 was reserved for testing.
[4]
Examples of weather maps generated by the neural network ... Image Credit: Ravuri et al
The model was fed a sequence of map examples, each one capturing weather data over five minute intervals during the training stage. It learned to pick up on common patterns describing how clouds spread in the sky and whether they produced rain or not.
In the testing stage, the system was asked to generate the next series of maps to predict rainfall in five minute intervals up to 90 minutes given four previous examples. Essentially, it’s a bit like feeding a system a short video clip and training it to predict the next frames.
[5]AI cleans up sat radar images so scientists can better spot warning signs before volcanoes go all Mount Doom
[6]Here comes an AI that can predict hurricane strength. Don't worry, NASA made it so it probably actually works
[7]Typical. Crap weather halts work on subsea fibre-optic cable between UK and France
[8]Trouts on a plane: Utah drops fish into lakes from aircraft and circa 95% survive
Its performance is judged using a number of factors, including how smooth and gradual the changes between the frames or maps were. Instead of calculating a straightforward method of determining the model’s accuracy, the team relied on asking a panel of 50 expert meteorologists to rank the predictive maps produced by the GAN and compared these with the maps produced from other types of more traditional numerical weather prediction systems.
“Using a systematic evaluation by more than 50 expert meteorologists, we show that our generative model ranked first for its accuracy and usefulness in 89 per cent of cases against two competitive methods,” the paper stated.
[9]
[10]
But how accurate is the model really? It’s difficult to say, Niall Robinson, co-author of the study and head of partnerships and product innovation at the Met Office, told The Register .
We decided that in some senses, the purest way to see if our approach had value was to simply ask the end users – our own meteorologists. In a blind study, they overwhelmingly preferred our new approach to other algorithms
“Machine learning algorithms generally try and optimise for one simple measure of how 'good' it’s prediction is. However, weather forecasts can be good or bad in lots of different ways; perhaps one forecast gets precipitation in the right location but at the wrong intensity, or another gets the right mix of intensities but in the wrong places, and so on. Ultimately, 'good' depends on what quality happens to be useful to whatever problem is at hand.
"We went to a lot of effort in this research to assess our algorithm against a wide suite of metrics, so we showed it was 'good' in several different ways. Moreover, we decided that in some senses, the purest way to see if our approach had value was to simply ask the end users – our own meteorologists. In a blind study, they overwhelmingly preferred our new approach to other algorithms. To our knowledge, this kind of really comprehensive assessment of multiple different kinds of 'good' hasn’t been done before to assess the use of AI in meteorology. I think it’s quite unusual in machine learning research full stop."
The research project is more of a proof-of-concept effort at the moment, and the Met Office won’t be using AI algorithms in real-world forecasting any time soon. There are all sorts of other tools the weather agency has to build before something like a GAN can be used, Robinson said:
[11]
“Once a new bit of useful research is published, there is still a lot of work to do to create an operational service. For instance, we need to carefully consider how new tools are deployed and maintained, the best user interfaces for our meteorologists, and how it fits in with all the other forecasts we provide. That way, we can ensure we provide the reliability that is at the core of the Met Office purpose.”
The Met Office said it’s exploring solving other AI research problems with the help of industry and academia. ®
Get our [12]Tech Resources
[1] https://www.metoffice.gov.uk/weather/guides/about-forecasts
[2] https://www.nature.com/articles/s41586-021-03854-z
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YVWK6Ey3XNvHC33zfqGbkgAAANg&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[4] https://regmedia.co.uk/2021/09/29/gan_deepmind_map.jpg
[5] https://www.theregister.com/2020/10/16/volcano_ai_prediction/
[6] https://www.theregister.com/2020/09/03/ai_hurricane_prediction/
[7] https://www.theregister.com/2021/09/28/crosschannel_fibre_delay/
[8] https://www.theregister.com/2021/07/14/fish_on_a_plane/
[9] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YVWK6Ey3XNvHC33zfqGbkgAAANg&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[10] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YVWK6Ey3XNvHC33zfqGbkgAAANg&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[11] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YVWK6Ey3XNvHC33zfqGbkgAAANg&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[12] https://whitepapers.theregister.com/
Don't need AI for this!
Just 1 line of code:
If location = UK then rain_in_next_90_minutes = true
(I'm sure the AI could be very useful for other locations, however!)
Re: Don't need AI for this!
Reminds me of when visiting Skye years a ago a local person explained the "alogrithm" for predicting rain.
"If you can see the Cullins then its going to rain, if you can't see the Cullins then its raining already"
Re: Don't need AI for this!
The more general form is "If you can see the $LocalHill then its going to rain, if you can't see the $LocalHill then its raining already"
There's a reason why East Anglia is drier than the rest of the UK.
"AI" is the solution
I did like their practical assessment of success, ie ask the users.
Since I only usually care about the rainfall levels in my immediate vicinity and at times when I plan to step outside, I find that the standard trick of hanging up an old pine cone still mostly does the job. I want to know whether to carry a jacket when I leave for work, when I come back from work, or when I want to go shopping. These days, of course, it's no longer necessary to also hang a pine cone up at work and keep a spare jacket there.
Success rates may vary, just like with any other forecasting system.
Re: the standard trick of hanging up an old pine cone
How does this work? Is it that you first go outside to hang up the pine cone, notice what the weather seems to be doing, then go back inside to adjust your garb appropriately? :-)
Re: the standard trick of hanging up an old pine cone
Absolutely. This was invented centuries ago before we had windows.
(on the off chance you're not joking and you've never been near a wood, pine cones react to moist levels in the air and close when it's too moist to protect the seeds they hold)
Huh
It's Yorkshire. I can look out of the window and can 99% guarantee it's going to rain today. And tomorrow.. and the day after that ... :-)
Objective performance?
" with lead times from 5–90 min ahead. Using a systematic evaluation by more than 50 expert meteorologists, we show that our generative model ranked first for its accuracy and usefulness in 89% of cases against two competitive methods "
In all fairness, I've only read [1]the abstract , but i'd really like to know the method's absolute false positive and false negative rates, rather than just its "accuracy and usefulness" rating compared with other methods.
Apart from which, given a 5 minute lead time I reckon any observant human could have a pretty reliable judgement about whether it's going to rain. At the 90 minute lead time, any observant human familiar with a given geographical area might also do pretty well. The most interesting report would therefore be how much better than an observant and informed human this might perform.
We seem to be constantly looking for ways of replacing the capacities of competent humans with complex machines. If successful this might well lead to a general loss of human competence, and there's no logical reason why this could not eventually degrade the competence of the population of creators and trainers of the machines. However, so far each machine at best has only a single narrow specialised skill, so we'd need an awful lot of them to replace the entire gamut of competences of a single competent human.
[1] https://www.nature.com/articles/s41586-021-03854-z
Re: Objective performance?
But your single expert human can only look out of one window at once, and is probably not lookin out the window anyway because they are reading El Reg.
Agree with your general rant about the usefulness of AI though....
British Weather
They're still waiting for it to stop raining more than ninety minutes so they can test it.
Only three rules needed
It has rained.
It is raining.
It will rain.
Radar
Err - Surely you get the same level of accuracy just by looking at the Rainfall Radar - a.k.a. looking out a bigger window?
Helpful solution to DIRECTION of rain
I'm rain averse. But if looking at a metoffice rain radar I can't tell what direction the rain is moving.
SOLVED! By the power of less than 200 lines of good old javascript you can see for yourself in a so-obvious-no-wonder-the met-office-aren't-interested way.
Here is my bench-top prototype. [1]https://vulpeculox.net/misc/jsjq/rain/index.htm
[1] https://vulpeculox.net/misc/jsjq/rain/index.htm
Re: Helpful solution to DIRECTION of rain
Is that moving back-and-forth or forth-and-back? Arrows (for the wind) would be clearer, and simpler to program.
Have they heard of darksky?
"weather forecasts can be good or bad in lots of different ways; perhaps one forecast gets precipitation in the right location but at the wrong intensity, or another gets the right mix of intensities but in the wrong places, and so on. "
It sounds like a job for quantum computing.
By the internet gods!
We might have an actual use case for "AI" unique to a problem the UK faces...
Imagine when this on general release, we can finally be free of the tyranny of having the washing rained on. Our forgetting your umbrella.
Finally a utopia for all in Blighty.
/sarcasm
Still one can dream.
Coat icon because until then you never know when it's going to rain.