Це відео не доступне.
Перепрошуємо.
Algorithms That Control Our Lives - with David Sumpter
Вставка
- Опубліковано 17 сер 2024
- In a world of fake news, Facebook scandals and psychological warfare, mathematician David Sumpter investigates how much influence algorithms truly have over our behaviours.
Subscribe for regular science videos: bit.ly/RiSubscRibe
David's book "Outnumbered: From Facebook and Google to Fake News and Filter-bubbles - The Algorithms That Control Our Lives" is available now - geni.us/qrbsZn
Watch the Q&A: • Q&A: Algorithms That C...
Algorithms permeate our modern lives and analyse our online behaviours constantly, so much so that it's hard to know if you can trust your digital devices and social media accounts.
But just how statistically sound are algorithms like those used by Cambridge Analytica? How well can they really categorise us and influence our behaviours?
David Sumpter is Professor of Applied Mathematics at the University of Uppsala, Sweden. His scientific research analyses life across the scales, from the inner workings of fish schools and ant colonies, to analysing passing networks of football teams, to segregation in society, machine learning and artificial intelligence.
This talk and Q&A was filmed in the Ri on 17 May 2018.
---
Thank you for our Patreon supporters who help us make more videos like this:
Ashok Bommisetti, Avrahaim Chein, Greg Nagel, Lester Su, Rebecca Pan and Will Knott.
---
The Ri is on Patreon: / theroyalinstitution
and Twitter: / ri_science
and Facebook: / royalinstitution
and Tumblr: / ri-science
Our editorial policy: www.rigb.org/ho...
Subscribe for the latest science videos: bit.ly/RiNewsle...
Product links on this page may be affiliate links which means it won't cost you any extra but we may earn a small commission if you decide to purchase through the link.
Just wanted to say thank you for taking the time to put all these talks together in a format those of us with a decent internet connection can watch and learn from. RI's lectures are my favorite.
In the future algorithms will be the threat . " Where you gonna run , where you gonna hide, no where because there's no one like you left" 😨
Just start the video when the lecture starts.
Please stop with the random intro clips!
Marc Marc - Seconded!
Thank you for the feedback. We're trialling short clips at the beginning of hour long talks as a way of increasing engagement, kind of like providing a teaser of what's yet to come. We've had to do this on a large enough number of videos to make sure our results are conclusive. We're running the reports at the end of today to see if they have been successful, so fingers crossed they're an abject failure
The Royal Institution - Thanks for reading the comments! The key word in the OP's point was 'random' - it was not clear that a teaser was what we were seeing - I felt like the film crew had arrived twenty minutes late and I'd missed the beginning of the talk!
Also, since this is UA-cam and not TV, we already decided we want to watch when we clicked 'play', so a teaser is not really necessary as that role is served by the thumbnail and catchy title (perhaps you might want to experiment with teasers for other videos in your library? Just make it clear with your editing that that is what we're watching and the main talk hasn't started yet).
Thanks K1lostream for the clarification!
The Royal Institution I get the purpose of it. But maybe you should fade in with a time, date & location just before the short clip. It's too sudden. It comes across as video with the start chopped off.
On the website, I found a list of upcoming talks, along with an option to book a seat. Is it possible to live telecast the talks, either through youtube or through your website? I would pay for it. I believe there are many others who would like to pay and attend the talks. The Ri has fans from all over the globe. People from other countries, like me, cannot always make it to the UK. That's why I'm requesting a live telecast.
We have actually considered it (and we did do a very basic trial run last year - ua-cam.com/video/QpGy7wqiOvA/v-deo.html) but currently we haven't been able to make it work out financially. While we are able to film and edit these events relatively cheaply, the cost of doing a livestream at the same quality would nearly triple the cost, whether we'd try to do it in-house or outsource it. Live stream cost estimates vary, but say we got it on the lower end and it costs us $1000 per event, we'd have to get 100 people to buy online watching tickets at $10 to break even. We're an independent charity so we're walking a very fine tightrope financially, but we can promise you that if we can figure out a way to not lose money on this, we'll seriously consider it. Or it would also work if we got a big donation to cover all costs! If you know of any rich philanthropists with a love for science, we're all ears.
Money is a big problem in science. I wish I could help you. If I come across people who are willing to patronise science, Ri will be the first in my list. I appreciate the amount of work you guys do for science. Hats off to the hard work and enthusiasm.
Al they have to do is livestream it. Any child could do it nowadays lets face it. So they just dont wanna do it
We don't want to do it badly.
@@TheRoyalInstitution have this problem been fixed?
Nailed it, the purpose of Scientific research (including Mathematics) is to remove as much passion as possible from the abstract natural behaviour of existing systems. Therefore the contrast of methodology and appreciation valuation for the situation is enhanced?
If allowing for different Player styles and techniques is the passionate part, demonstrating what it is possible to achieve in strategies is definitely part of the game.
Systematic "corralling" of personal attention on material products and services to the exclusion of passion driven natural satisfaction and self-sufficiency is not going to end well. (If that's the message behind dissatisfaction with the Scientific Methodology, ..getting the blame for the overused strategic applications)
In order to "protect your family with a gun", "we had to destroy the village(country) in order to save it"?
Natural Mathematician's humour (sub context): "I did a study to see if I live in an Echo Chamber".
The circumstances of QM-Time modulation-modules imply that the bio-universe of Pi, e-Pi-i orbital tangency to the zero-axis here-now, is an equivalent sum-of-all-history statement to "this is an experience of living in a natural-mathematical Echo Chamber of quantum information", or it's a Phys-Chem connection of the "One Electron theory" type, temporal superposition..,
..implies the situation that AI is a fairly trivial sub context of the inherent intelligence of natural, general Universal temporal existence.
Data, information, is mined and discovered by relearning the development path of QM-Time; the Phys-Chem re-evolution context of circular logic oscillation pulses, around the Universal singularity.., intelligently? (Comparing the recorded empirical history of event changes in the eternity-now event probabilities, and the Science Methodologies of possibilities. Analysing quantum information made diffuse by pseudo randomness, and distributed in quantized field phase-state, or nanoparticle type qualities)
Two things. Firstly: conservatives doesn't have to search and debate, to get information and arguments. Theirs is the world we live in. Liberals, or like me social democrats, have to find ways to change the system. That is more work, creates more content. Secondly: General AI ain't that far away. Remember that we don't have to create it. Once we have created a smart enough AI, it will do the programming. And thinking at a billion times our speed, it'll be a matter of days.
Lovely talk! Also, great speaking skills - the video is comfortably watchable at x2 speed even for a non-native speaker as myself.
29:50 I think there are a lot of people who are neurotic, but do not fill out tests as if they were neurotics, in most questions it is pretty easy to guess what each answer means, so they may be neurotic on Facebook but not in test.
A big puzle how entities like Cambridge Analytica even exist. analytics have been around for years yet they suddenly overnight get government contracts from nowhere. How does that even happen?
No, you dont get 45m,000 hits, you get 45,000 possible pages of interest. Hits dont actually mean anything since 1980
It has been shown that the color blue is associated with trustworthiness and honesty (at least that is the common opinion within marketing and advertising) - so I suspect that the blue logos of the AI research is to avoid any connotation with the singularity skynet apocalyptic fears that exist in the general population.
31:13 "If I just said randomly: "You are the most neurotic one," I'd have a 50% chance of getting it right". This would be correct only if 50% of people are the most neurotic, which is not the case
Basic Bayesian probability mistakes...
No. If two people are picked randomly from the audience, one of them is more neurotic than the other. If he said, "You are the most neurotic one" to a random one of the two, there's a 50% chance of getting it right. He meant, the more neurotic of the two (assuming the chance of both being equally neurotic is negligible).
thanks, I missed the part "If I picked out 2 random people...".
Maybe the RI should give a slight hint that the beginning of the video is a snippet to entice! I thought the video had missed the start of the lecture. I had to read down the comments to find out. This was frustrating. Then I got a colour slide but instead of an introduction, I was thrust into the lecture minus intro. I am not a scientist so I need an introduction, I need to know a bit about the lecturer. I felt cheated. I was not intelligent enough to discover the contents of "Show More" sooner. By the time I did the lecture was in full swing and I missed quite a bit and had to start again. I needed the information in the "Show More". Great lecturer though.
I would bet that bubble sort is the most used algorithm every day since it was "discovered"
He seems to not understand that even a small statistical advantage in targeting can make a huge difference in a close race.
marginal gains, indeed :)
Algorithms That [Do Not] Control Our Lives
Good talk, although powerful AI came around sooner than a long way off.
Great information.
Can air know us?
And if you went to university in the 1980s or 1990s, you likely had an awkward computer science undergraduate approach you in a hypersocial environment like a coffee house and ask you to participate in a very personal,yet anonymous survey.
Don't actually use the current psychological scientific terms lest Google fire you unjustly...
Wow, I actually love anime, role play games and terry pratchet books. But I think I am not shy xD
Statistics only work on averages of the whole- not in saying something is 100% true for everyone. Frustrating that I heard people calling in on the radio after hearing the overall universal trait women find most attractive in men is being handy around the house & a few women called in irate that they'd rather have men who they can express their feelings too- duh, even if it's 80% true, there's always outliers for everything (not to say we're wierd ;)) but of course companies need trends so they can maximize profits
Pls, subtitles, Russian and other languages
We'd love to, in fact we have started a fundraiser to be able to do exactly that - www.patreon.com/TheRoyalInstitution
Can't afford that. Would be more than willing to translate, free of charge, to Swedish.
4:05 29:10
_-·¹← / ·¹-e→
Red looks too evil :)
You see what he's like? - We haven't changed our mind! - No toast! :)
No it isn't!
stop stealing my data
LHC!!
This guy overthinks everything.
Remember, when Obama did this it's brilliant. When Trump does it it's evil.