Waching daily May 14 2018

I think that the United States has many problems,

but one of its largest problems that go unrecognized is its prison problem.

There are currently over 9 million people incarcerated in prisons worldwide.

Half of the number is held just in three countries: The U.S, China, and Russia.

The U.S by far has the most prisoners with 2.3 million behind bars.

The problem with the prison system, is that it doesn't work.

For example you're basically putting everyone who's ever committed a crime

into one building. A building where they can talk to each other and learn how to commit more complex crimes.

70% of prisoners are rearrested within 3 years of being released, and 77% are rearrested withing 5 years.

Though America does have a horrible reoffending rate, other countries aren't far behind.

In the UK, 47% of prisoners reoffend within 3 years.

Norway is officially the 6th safest country in the world according to the

Legatum Prosperity Index

Norway only puts 70/100,000 people behind pars, compared to America who puts 716/100,000 people behind bars.

Norway's reoffending rate is only 20%,

the reason why this is, is because their prison system isn't like other countries.

Halden is a maximum security prison in Norway.

Inmates get their own cell, and they have unrestricted windows, with plenty of natural light.

There's a flatscreen TV in every room, a private shower, a minifridge, and other luxuries.

There's cookery school where inmates are taught how to cook for themselves,

there's a shop and kitchen so inmates can buy ingredients and make their own meals, or cook for their friends.

Prisoners are free to leave their cells at anytime and enjoy the prisons various facilities,

such as: a gym, a sports hall, an outdoor sports field, a music studio, a woodworking studio,

but not only are prisoners encouraged to leave their cells, theyre actually paid to do so.

Prisoners are given $7 an day to leave their cells and socialize.

All inmates are offered the opportunity of education and paid jobs within the prison.

Nobody has ever tried to escape from Halden and incidents of violence are extremely rare.

Norwegian prisons focus solely on rehabilitation and repairing the mental damage that may have caused one to commit crime in the first place.

Maximum prison sentence for Norway is just 21 years.

Even terrorists are only assigned 21 years [lol ok]

Personally, I think that Norway is doing something right with their prison system and

that America can take after them.

Though the prison systems have their flaws,

I think that lower crime rates make up for it.

HAHA YES

FINALLY

gracias al cielo!!

im so done.

oH my god it hurts to blink

For more infomation >> Greatest History Project - NORWAY IS BETTER THAN US IN EVERY WAY - Duration: 2:22.

-------------------------------------------

AI is on the roadmap, where do I start? - Innovation Leaders 2018 - Duration: 15:50.

so Good evening, everyone Tonight I'm going to speak about AI

Yes, this is another conference about AI .

But looking at these huge audience tonight

makes me think that yeah well AI is an interesting topic but also maybe that AR

is not crystal clear for everyone.

So I'm going to talk about AI but not generic

facts about AI, neither already seen use cases.

I'm gonna talk about AI for all

AI for you

And this without math nor theory don't be worried

so I hope I will convince you to treat today that

AI is concrete and not only for global or

high-tech data-driven companies but for all businesses

and I'm going to show you a real working example

that will dymistify the whole thing

Just a few word about wazzabi because you see wazzabi and you don't know what it is

I've been working for 15 years in AI for various company and dozens of industries

and I founded two companies in this field and the last one being wazzabi and

in wazzabi we do AI for all, for SMEs as well as larger companies

so when I visit my customers for the first time

I often hear the board has asked us to

do something with AI we need to show the markets that we are ahead of competition

and the next sentence often is well I don't know where to start

can you help us ?

and that's normal I would say okay

AI projects have not been implemented in a lot of companies maybe a few of them

but still there is no rules no best practices in place so maybe you decide

to go and find skills over the internet and you search for

contractors doing AI and there too is very hard to find the talents that know

what it takes to implement such AI services. It's not like a traditional IT

project where all the contractors are well identified

here to find information is hard

who are the players in Switzerland?

And most of our clients can't find this information

so they're struggling they are postponing and even

sometimes abandoning their AI initiatives

so you hear about AI everyday

everyway: in the newspapers, at the radio, on television, there are dozens of conferences about AI everywhere

about so just stop because it's what is the feeling of our custom today

too much information and maybe not the right information

so in this conference you see use cases

what our competitor did and it's beautiful

and most of the time they come from the US by the way

and they are far away from our daily concerns

so and you cannot get started

Maybe you're even frustrated okay so

everyone tells you that AR is the Holy Grail and you're like this little kid

looking at this amazing toy and who thinks he can never have it

but actually yes you can have it

and there it is

And now you're probably thinking

what is this guy with its green tie talking about?

This is Excel this has nothing to do with AI

But we'll see in a little bit yes actually this excel sheet

is using AI to make some predictions

but let me first introduced the use case because

I said I want to be concrete tonight

so I'll share with you this part of the optimization we did for a customer

so suppose you have customer requests with

different complexities that you receive each day

could be claims for an insurance company

could be loan applications

could be tax returns

customs declaration

subscription to a service your company's offering

and on your side you have team members

with different skills and capacities

how do you divide the workload among your team members ?

some might speak English in French but not German

others are working on the whole week others not on Wednesdays and so on

so it's really hard to find a system that will do that

so how would you do that in order to have the smaller time,

the processing time small but of high quality

so if you take the traditional approach

you will probably design a rule-based system

that will decide on fixed rules

we saw that it could be a lot of rules for this kind of systems

and you can see the disadvantage of this approach

it will result in a static system okay

so what if your skills inside that your team is changing

what if the behavior of your customer changes

so this has to be re-engineered

I'm going to show you now what we did with machine learning

and how this can be adapted in real-time

so first let's go back to this excel sheet

and I'll show you if I'm able to go here

live what is happening in this beautiful excel sheet

we have an excel sheet online

on the top here this is the value for one specific request

so we have the customer ID the age and the revenue

and here at the bottom we have the predicted complexity of this request

complexity goes from zero not complex to for the highest complexity possible

so this means that

if your system predicts that this request is of complexity four it needs

an experienced employee to process this request

now let's test it live

and I hope it will work

it should be working

I'm going change a few values here

let's put 1,000

Let's go back to this one and change this value here

let's go to 1000 and it should be

Predicted value should be one ok

and one means that this request is less complex and

can be automated or processed by less experienced employee

how we did it ?

I said this is AI

Actually Excel here is connecting to a web service in the cloud

that will make predictions on the complexity of each request

and the thing is

you can use it directly from Excel when it's working

so now I'm gonna show you how we did it

and how we designed this

I'm going to open something that is called Microsoft Azure machine learning studio

which is one of the well integrated machine learning tool available today

and I decided to show you the use case with this tool because

it's very visual and intuitive so you don't need any background

or any knowledge of machine learning

don't be worried I said there will be no theory

to understand what I'm going to tell you now

so we we've learned that to train model we need some historical data

so here first thing we are going to do is

to input some historical data and as you can see you can visualize this

we'll find the same information that we found in the excel sheet

we have 20,000 lines here ok and we have some columns and

here at the end for each of these 20,000 requests

we have its complexity

that means that the known data that's what

will be used to train our system and help him decide the complexity of a

given request

we have to do some data manipulation could be cleaning

could be feature engineering that's a specific field

we're gonna create some new features that will help the system to learn faster for instance

then we're gonna split the data

we've heard that in the first talk

we're gonna use some part of the data for the trainings and we'll keep

separate sets for the testing

to know if our algorithm is able to generalize well

so we decided here to train two types of algorithms

the first one is a true class neuronal network

don't be afraid

and here is a decision tree that's another kind of algorithm

but what is great here that we're able

to compare the accuracy of the two algorithms and decide which one is the best

so here we have neuron neural network here we have the decision tree

and we can see here that the accuracy of the decision tree is higher, maximum being one,

is higher than the neuron

what we're gonna do here is just click on this one

they say okay I wanna use this algorithm in production I'm gonna just

click here to generate the web service and we are done we then have access here

to the web service and we can play with it directly from Excel that's exactly

what I shown you just before

or if you want to use this web service inside any

of your application then you just need to call the web service with this

information available here it's all documented and very easily connected

through any kind of application

well so I won't go into more details so as you

can see this seems to be not rocket science okay

even if it looks straight forward you have to be very careful here

because that's the problem of most

companies that doesn't haven already implemented their services

it seems easy and straight forward so you can create an Asure account in five minutes

and play with it

but you need to understand what you are doing

so sure you're not playing with duplo's you dealing with LEGO Technic okay

and AI projects really requires specific skills and experienced people to be successful

so what are these ingredients that we need to

successfully implement an AI project

the first thing is an affordable computing power

and this we know that it's available in the cloud for not a

lot of money you don't have to buy all the infrastructure and put it in place

on your premises that's the old days now you can just provision clusters in five

minutes in the cloud so this is okay

then of course you need sophisticated

algorithms to learn the tasks you want to achieve

and these two in the AI and machine learning fields a great thing is that the open source community is huge

and most of the algorithms are available are free okay

and they even packaged in libraries available for Python or other

There are two languages commonly used in machine learning

so this is okay too we have all this nearly for free,

then it comes to an important one

it comes to data and there is really at the heart of your projects

and is an important success factor

and here I want just to say a few words about data

I won't be too long

you sometimes hear that you need a lot of data for doing machine learning

sometimes yes and sometimes no but

to require a data does not mean to

collect data as much as possible

you're AI project will never start with your data

but it will start with your strategy

what does your company want to, want to achieve with your with AI

no matter what data you have already and

in most companies we went they have the data

okay maybe we have to clean the data that's an important work okay

but don't be afraid:

we need social media data

because I saw coca-cola did that in a workshop, no

Keep Calm and you have historical data

most of your company have a lot of data and you can do already a lot of

things only with what you have

and last but not least the skills

of course we need people that know this technology knows how to manipulate these algorithms

and this data to set AI project

the only problem here is that these skills are scarce today in Switzerland

there are a lot of efforts putting in even here at EPFL with

the new program here but still

and most of you if you want to implement

your first AI project we need to make partnership with contractors, specialized contractors

and find the skills outside of your company

but it's possible because in this field experience helps a lot ok

it's a lot also about feeling ok you need to know the background you need know the

technical aspects but experience is really really important

but we can say we have it here

so I hope to have shown you that AI is concrete and applicable

in many businesses in any businesses

so if we want to make this even more

concrete for your own company

wazzabi has created what we call a AI starter kit

which will allow your company to discover what AI can do

for a specific problem inside your company and in the end you'll have a real working AI

service as we shown you just before

so tonight I would like you to leave this

conference with the understanding that yes AI can be complex that it requires

specific skills experience and an adapted management to be successful but

I also would like you to have the feeling that AI for all can become

AI for you

thank you much

For more infomation >> AI is on the roadmap, where do I start? - Innovation Leaders 2018 - Duration: 15:50.

-------------------------------------------

✅ Bethesda's failure to keep Rage 2 a secret is now complete, as the live action trailer confirms a - Duration: 2:33.

Bethesda's failure to keep Rage 2 a secret is now complete, as the live action trailer confirms a Mad Max style setting

Rage 2 is real, and it's going to revealed in full at E3 next month. That fact had already been leaked by Walmart, but now the teaser trailer below has appeared online as well

According to hints dropped by Bethesda last week the teaser is likely to be made official today but was leaked onto the Internet over the weekend and saved, in GIF form, by the same tipster that broke the Walmart story

Because it's a GIF there's no sound, but the live action clip seem to suggest a setting and tone largely unchanged from the 2011 original

As you can see, that setting is basically Mad Max, to the point where the original also let you race a post-apocalyptic car and compete in races

Whether that element's still in the game isn't clear from the teaser but the original was also semi-open world, so it's likely that aspect will be expanded upon in the sequel

Even if that does start to encroach on Fallout territory (or indeed the official Mad Max game)

Somewhat surprisingly the game is referred to simply as Rage 2, despite the original never having been that successful or famous

It did end very abruptly though, so hopefully the sequel can atone for that at least

What's also interesting about the leak is that it confirms one of the major revelations from the Walmart Canada leak, increasing the likelihood that the other unannounced games – including Gears Of War 5 and Just Cause 4 – are also true

Email gamecentral@ukmetro.co.uk, leave a comment below, and follow us on Twitter