They trained the model (U-Net) with frames from numerical solutions of the hydrodynamical equations.
Today, we are bringing other exciting results involving black holes and AI. We released a new paper:
"Black Hole Weather Forecasting with Deep Learning: A Pilot Study"
Work by Roberta Duarte (@import_robs), Rodrigo Nemmen (@nemmen) and João Paulo Navarro (from @NVIDIABrasil).
They trained the model (U-Net) with frames from numerical solutions of the hydrodynamical equations.
We want to investigate if deep learning can be a new method to simulate accurately in less time!
1- The model simulating only one system after learning only from this system
2- The model simulating an unseen system after training with several systems with different initial conditions
The result is that the model can simulate up to 8e4 gravitational time accurately with a speed-up of 30000x faster!
However, they hid one system to understand the generalization power of the model!
It simulated the unseen system for 4e4 gravitational time, showing that the model can generalize the black hole physics presented in the dataset!
More from Science
😭
The new answer to a 77-year-old problem in data analysis, published today in @naturemethods. Instead of significance tests, use estimation graphics. Our software suite DABEST makes it easy for everyone to visualize effect sizes.https://t.co/UzwXJ7EUC5 pic.twitter.com/VtxyY0xaRM
— Adam Claridge-Chang (@adamcchang) June 19, 2019
https://t.co/hm9NoaU4nr
Open letter to journal editors: dynamite plots must die. Dynamite plots, also known as bar and line graphs, hide important information. Editors should require authors to show readers the data and avoid these plots. https://t.co/0GNKEIUCJL pic.twitter.com/OS9ytEFRZN
— Rafael Irizarry (@rafalab) February 22, 2019
https://t.co/8fKDiKjSWc
Couldn't find D3 code for grouped horisontal box plots that show data points so I made this @mbostock @thisisalfie https://t.co/cQjDPhyZdw pic.twitter.com/y6RNmDB2p3
— Ulrik Lyngs (@ulyngs) June 28, 2017
https://t.co/jkaicC1F2x
made a pkg for pirate plots in ggplot: add any of points/means/bars/CIs/violins \u2013 better than ye olde bar/box plotshttps://t.co/Z2m2kW3hsl pic.twitter.com/npAirPQexM
— Mika Braginsky (@mbraginsky) September 28, 2017
https://t.co/PpxWT4Jef4
See the new #PowerBI visual awesomeness for data points & sources, box-&-whisker plots! https://t.co/dOmgoxWfDE pic.twitter.com/HAUOAMJEJW
— Microsoft Power BI (@MSPowerBI) February 1, 2016
2) The leading hypothesis is that the new variant evolved within just one person, chronically infected with the virus for so long it was able to evolve into a new, more infectious form.
same thing happened in Boston in another immunocompromised person that was sick for 155 days.
3) What happened in Boston with one 45 year old man who was highly infectious for 155 days straight before he died... is exactly what scientists think happened in Kent, England that gave rise to #B117.
Immunocompromised 45 year old suffered from #COVID19 for 155 days before he died. The virus was changing very quickly inside the man's body\u2014it acquired a big cluster of >20 mutations\u2014resembled the same ones seen in #B117 & #B1351. (NPR audio Part 1 of 2)\U0001f9f5https://t.co/7kWiBZ1xGk pic.twitter.com/ZJ7AExB78Y
— Eric Feigl-Ding (@DrEricDing) February 8, 2021
4) Doctors were shocked to find virus has evolved many different forms inside of this one immunocompromised man. 20 new mutations in one virus, akin to the #B117. This is possibly how #B1351 in South Africa 🇿🇦 and #P1 in Brazil 🇧🇷 also evolved.
2) NPR report audio part 2 of 2:
— Eric Feigl-Ding (@DrEricDing) February 8, 2021
Dr. Li couldn't believe what they found. "I was shocked," he says. "When I saw the virus sequences, I knew that we were dealing with something completely different and potentially very important." pic.twitter.com/HT3Yt6djFd
5) “On its own, the appearance of a new variant in genomic databases doesn’t tell us much. “That’s just one genome amongst thousands every week. It wouldn’t necessarily stick out,” says Oliver Pybus, a professor of evolution and infectious disease at Oxford.
I'm going to answer the question so many people have been asking this week:
WHAT IS PROJECT X???
Here's the definitive thread to tell you - and show you -precisely what Project X is
Grab a drink, sit down with me and let's #TalkLiberation
<3
1/?
"Project X" is actually called "PanQuake".
Pan means "all". Quake is the huge effect our voices can have when our communications are uncensored and when we have access to brand new functionality that *enhances* our social reach, rather than diminishes it
Here's our logo:
2/?
You can follow the fledgling official PanQuake Twitter account here: @pan_quake and see our super cool new website here: https://t.co/F7wLSeM6aK
You can find our donation page here: https://t.co/VICFnsR0RX
Keep reading this thread to find out why we created it & what it is
3/?
SPOILER ALERT: Much of the content below this point is from my personal slides & speech notes from today's launch event. That stream got totally ruined by (big) tech problems, but I'm happy to report everything is turning out wonderfully
Not one single team member or guest left. We are all still here, smiling not crying, as we record this event and will get it out to you all very soon :)
— Suzie Dawson (@Suzi3D) January 17, 2021
I'm so proud of everyone, what an amazing crewhttps://t.co/RmE0BicIXF
Here are some of our most high profile & dedicated public advocates for PanQuake - many of whom were scheduled to appear at our launch. All of whom stuck around for hours, to do a prerecord of the event, which is being edited, processed & uploaded for you as I write this.
5/?
You May Also Like
These setups I found from the following 4 accounts:
1. @Pathik_Trader
2. @sourabhsiso19
3. @ITRADE191
4. @DillikiBiili
Share for the benefit of everyone.
Here are the setups from @Pathik_Trader Sir first.
1. Open Drive (Intraday Setup explained)
#OpenDrive#intradaySetup
— Pathik (@Pathik_Trader) April 16, 2019
Sharing one high probability trending setup for intraday.
Few conditions needs to be met
1. Opening should be above/below previous day high/low for buy/sell setup.
2. Open=low (for buy)
Open=high (for sell)
(1/n)
Bactesting results of Open Drive
Already explained strategy of #opendrive
— Pathik (@Pathik_Trader) May 27, 2020
Backtested results in 30 stocks and nifty, banknifty.
Success ratio : approx 40-45%
RR average 1:2
Entry as per strategy
Stoploss = Open level
Exit 3:15 PM Or SL
39 months 14 months -ve, 25 +ve
Yearly all 4 years +ve performance. pic.twitter.com/nGqhzMKGVy
2. Two Price Action setups to get good long side trade for intraday.
1. PDC Acts as Support
2. PDH Acts as
So today we will discuss two more price action setups to get good long side trade for intraday.
— Pathik (@Pathik_Trader) June 20, 2020
1. PDC Acts as Support
2. PDH Acts as Support
Example of PDC/PDH Setup given
#nifty
— Pathik (@Pathik_Trader) June 23, 2020
This is how it created long setup by taking support at PDC.
hopefully shared setup on last weekend helped. pic.twitter.com/2mduSUpMn5