Are Robots Ruining Your Marketing and Advertisements?


Good afternoon! Hi guys! I have a fabulous
slogan for you. You ready? I’m ready. Legend that thing. It’s not quite that
great, is it? Noo. So, there was a study done recently that seven years of Nike
commercials were run through AI artificial intelligence and that is the
slogan it came up with legend that thing you might want to go back to the drawing
board. Right. You know, it’s interesting artificial intelligence it has its place
and right now there’s so much data. Data is like… King. (laughs) It is the king, but it
has its place in marketing. You know you can make data-based decisions. However,
it’s not for everything. It has its place. It’s a tool. Not a replacement. Exactly.
You don’t want artificial intelligence developing your videos. You don’t want
artificial intelligence developing your ads. A really good example is this
company. We love this company We use it all the time: Lumen5. They do great
videos they have great stock images but if you just put a link of your website
into lumen5 and you let it, you know, make a video for you, and you don’t do
any editing, you’re gonna get the Nike video that we’ve posted the link
below. Watch it. Legend that thing. Yeah, it just doesn’t quite make sense, and it’s
the same thing with your ads. Look advertising it’s meant to connect with
humans. That’s your client base and you don’t want robots making decisions for
human connection based marketing. If you’re gonna do marketing, have AI, but
make sure you have human eyes looking at your ads, at your content, at your videos,
and, you know, if you need any help with that we’re happy to look over your ads
and see if we can improve it at all. Thanks guys! See you tomorrow. See you
guys.

You Ask, I Answer: Most Exciting 2020 Marketing Trend?


Christopher Penn: In today’s episode, Laura asks, What is the industry trend you’re most excited about in 2020? Well, the continued growth Unknown: and Christopher Penn: the practicality of what’s happening in artificial intelligence and machine learning and in marketing data science force. There’s three specific trends that I think are so exciting and have been exciting for quite some time. Number one, continued growth in natural language processing. 2018 was sort of a watershed year for a lot of natural language processing in 2019 brought some of the largest models available that we can use to process text, to understand it better to be able to in some cases, generated and that is stuff that I expect to see continue in 2020 not just at the cutting edge of research but at the ability for the average machine learning practitioner to access it and deploy it. So by that i mean it’s it’s relatively friendly, relatively easy to use it is not the incredibly challenging you know, super high tech stuff although that that stuff is exciting too. But for the average well versed practitioner in data science and machine learning to be able to pick up something in NLP and use it. Number two, there are some really amazing things happening in the world of audio with wavenet a few years back and now Mel net bunch of other technologies that are making it easier for machines to understand audio, particularly the spoken word and Then able to replicate it and synthesize it. And this is some amazing stuff again, there’s if you look at the milnet demo, you can listen to a machine synthetically generate voices, synthetic degenerate music. I think music composition is has come a long way and continues to accelerate. It’s not going to replace humans anytime soon. But it’s getting there. And it I would say it’s at the point now where AI can generate, like elevator music, right? stuff if you put it on in the elevator like no one would be offended. But certainly it’s not going to win any Grammys. But as with all things in, in machine learning, the technology improves, the models get better. And whereas two years ago, it was it sounded like something like your three year old would You’re hammering pots and pans with spoons. Now it sounds like elevated music. So expect in time for it to become a credible alternative for people who want access to music composition and don’t have the skills themselves can’t play an instrument but can direct or conducting an orchestra of machines. Number three, and this is a bit more esoteric, but there will be more and better pre trained models in 2020. We saw within the natural language processing field open AI is GPT to released its mega model, the 1.5 billion hyper parameter model in 2019. And that was a really big deal for that company because that was the model they were so concerned about falling into the wrong hands. And they said they didn’t see any evidence of its its use and partly because it’s such a beast try and get to us. But one of the major Overall meta trends and AI in the last couple of years has been, instead of generate your own models, pick up a pre trained model and then fine tune and expect to see more of that happening in music, image recognition, video, natural language processing, good old fashioned statistical computing all these things. We’re seeing more and better models for better just you pick it up and you work with it. Now, there are some challenges with that in the chat. One of the big challenges, of course, is that you are relying on a pre trained model and you you don’t know for sure how that model was trained. You after take on faith to some degree, the model was trained well, so that’s going to be a consideration. Which brings me to three trends I’m concerned about number one is interpreter ability and explain ability, the ability for machine learning and AI practitioners to explain what is the models are doing, which is really important and something that The industry is struggling with and will struggle with for some time. Partly because interpretability is so much more expensive than explain ability, but is the gold standard for being able to say, watch and see what the machine is doing any step of the process. Number two, the things I’m worried about is, is bias. And this goes hand in hand with interpretability. If we don’t know what the machine is doing, we don’t know if it’s doing something wrong. There have been many, many, many examples of machines making decisions that on the surface seem okay. But when you get inside the model, or when you see the post hoc results, you’re like, That’s not right. So bias is a major concern. And it’s one that the industry is making strides on. But the industry as a whole is not going fast enough. Not just fast enough to allay some of the fears that people have. And, and set aside the misconception that’s important. And number three, and by far the one I’m most concerned about in 2020, because it is an election year is the misuse or malicious use of artificial intelligence technology for things like deepfakes is the most popular cited example not the one that I think is probably the most prevalent. I honestly think that bots with minimal natural language processing capabilities are much bigger problem because there’s so much easier to scale deepfakes don’t scale well. Right. deepfakes require a lot of computational power. And yes, you can rent it for pennies on the dollar from like a Google Cloud or something. But to do so, then also makes you not anonymous, right? Because once you sign into something like when a big text clouds every single thing you do is tracked and can be identified but long for So in a lot of cases, if you’re doing something malicious, you need to be doing it in the dark away from the prying eyes of every major tech company ever. So things like, you know, mass armies of Twitter bots and Facebook bots and things are a much more practical application. And very easy, very cheap. And they have the ability thanks to the hyper partisan world that we live in, to really manipulate people. And it’s not the machines fault that humans are gullible and that humans like to have confirmation bias out the wazoo. But it’s the machines enable hostile actors to do more faster and better. And, at least in the context of the United States of America, our defenses have been largely dismantled the last couple years with the abolition of the cyber security Council and things like that. So we are in a case where AI can be maliciously used. And that’s very concerning to me as a practitioner because, again, we want people to trust this technology. If the technology is being used for malicious means really hard to build trust around it. So that’s sort of the opposite of is exciting trend. That’s the the most worrying trend but that’s what we have to look forward to in 2020. For marketing data science for machine learning for artificial intelligence, would love your thoughts, leave them in the comments box below. Subscribe to the YouTube channel on the newsletter, I’ll talk to you soon take care what help solving your company’s data analytics and digital marketing problems. This is Trust insights.ai today and let us know how we can help you

How You Can Increase Your Salary as a Software Developer


Hey, this is the Daily Overpass, my name is
Eric and I make apps! Now today, I wanna talk about how you can
increase your salary as a software developer! Alright! So today I’m coming to you from Green Park
in Reading. It’s this big industrial business park. I was here for a seminar which just let out
a little while ago and I thought I’d take a little walk around, and this place is nice! One day, you see something like this and you
think one day I want Overpass to be at an office complex like this because there’s a
big lake in the middle, all these big glass buildings, there’s walkways here, you see
a bunch of young technology people with their day badges on walking around. Kind of like when you see the Google complex
and everything like that, and it’s really nice! In fact, we are moving to a new office in
Wantage here soon, just yesterday looking at some new offices, and we will be moving
within the next month, something a bit nicer now that we’re getting some employees and
stuff, try to grow a little bit! Anyway I wanted to start off today by thanking
everyone for the live stream last week. I did my first live stream of Thursday night
and I was really nervous, I wasn’t sure if I was gonna have enough stuff to talk about. I appreciate everybody who showed up and asked
questions and everything, and we’re gonna do it again this Friday at 8 pm UK time, which
I think is noon on the pacific coast in California, so if it’s something you’re interested in,
attending the next live stream, it would be really great to have you there, and if you
have any questions that you wanna ask, we can put it to the group so it’s not just me,
because I’m not an expert on anything. But we can put it to the group and if you
want us to look at your Google Play page, or anything like that, that might be kind
of fun too! But, if you’re brave enough for everyone to
see it and give a bit of constructive criticism, but it’s a really good group! Anyway, that’s there. And also I wanna say thanks everyone for liking
and subscribing and sharing the content here on the channel – I’m supposed to say that
at the beginning! So, today I wanna talk about developer salaries! I was talking to a friend…. Eeek, I am so sorry for this awkward transition,
I filmed that entire video out at Green Park, it was beautiful scenery, people were walking
by, everything, and the microphone failed halfway through it, so I got a lot of footage
of me just “mu mu mu”, like an idiot with no audio. Sorry about that sorry for this awkward transition. So what I was telling you about was a couple
of days ago I was contacted by a friend of mine that I used to work with when I worked
at an investment bank as a contractor. Se this must have been about twelve, thirteen,
fourteen years ago, and it was one of my first contracts, my rate kept going up and up and
I was making about 375 pounds a day working at this investment bank, which I thought was
really good, and it is really good! So as a developer, back then it was a lot
of ASP stuff, it wasn’t even ASP.net, it was a lot of Vb, you know, com components, stuff
like that. And there was a guy who came in as a consultant
to work on this one project, on this one trading system that we worked on together. So it was like an older legacy system and
I was writing the new glue that connected other systems, kind of the piping, and everything
like that. And we talked a lot at lunch and he just knew
this one technology which is an older technology, unic space, and we talked and I kept trying
to tell him he should learn ASP or ASP.net because there’s more job opportunities in
that language that I have. And I was very much like “you spend some much
time looking for work, I would say you should learn this technology because there’s lots
of jobs! I’m thinking you should do the thing where
there’s lots of jobs for. And then one day I found out how much he was
being paid, like we were just having a conversation, I was making 375 a day and he was making over
a thousand pounds a day! I couldn’t believe it! I think he was making three times more than
me, it was more than a thousand pounds a day. And it turns out, because there were not a
lot of people who could do what he could do – the market was full of ASP developers, was
full of PHP developers, full of all these developers, everyone who goes looking for
a job, you look for the one with the most opportunities, but that meant that the supply
was bigger than the demand, so that they could lower down the rate. So this guy was making loads because there
was only a handful of people in the world who could do what he did! Then I started thinking about this, thinking
“wow, I didn’t think I was actually on a low rate, so I started asking around. There was another team that were mostly Java
developers, Java and some older technology, I can’t remember what it was, but they used
some older technologies and everything like that and I started asking them and they were
making more too because it was hard to find the skills that they had, even though the
skills they had used to be mainstream, but they were older technologies. It wasn’t Java – it was…I can’t remember
what it was but it was an old technology! And these guys, they knew it so they were
able to charge more! And it made me start thinking about the distribution
curve. So I don’t know if you’ve ever seen the distribution
curve, where its like the bell curve where you’ve got the early adopters, the early majority,
the late majority and the laggard that come up behind. It made me start thinking about technology
and being able to raise your rate as a developer! So it seems to me, if you were in the early
majority and you have a skill that’s not very widely used yet, you might not find a lot
of jobs in that. So right now maybe flutter, flutter might
be a good one now. So if you were just becoming a developer and
you went to go learn flutter, you might have a hard time finding a job. But, you’d be able to charge a lot more because
there’s fewer people in that market. As soon as it gets to the early majority,
right now the early majority in my opinion is react native , ionic, native – all that
kind of stuff! And then you’ve got the late majority, which
is definitely PHP, I could throw a stick out the window and hit a PHP developer! There are so many out there, so their rates
are much lower. We’ve got PHP, c sharp, that kind of stuff. Then you’ve got the late majority which is
like the laggers, which is like pearl and all this ASP, classic ASP, VB6, that kind
of stuff. Those people could probably charge a lot more
because it’s hard to find those developers with those skills to look at the old stuff. SO,my advice always to new developers is to
look at the center section for the job that you’re gonna go for. Because you don’t wanna learn a technology
which you can’t find a job in, but at same time focus on the early majority, because
that’s…if you can learn those technologies while you learn the other thing, while you’re
learning react native, also learn something that’s cutting edge. And I say flutter but there are probably things
that are more cutting edge than that now! There’s always machine learning and AI and
all the other cool stuff that’s coming out. But if you were to lean that early majority
stuff right now, eventually it might become the majority and then the reins will go down,
there’s more people flooding into the market, but if you hold onto those skills as you get
older, you might be able o charge more again at the late majority. So anyway, those are just my opinions after
talking to him it was kind of funny, it was funny, he was doing ok. I was worried about him thinking if you can’t
find any jobs, you should learn ASP, and he had no problems! So anyway, sorry about the audio problems
today. That;s it for today, I’ll talk to you tomorrow”

Work From Home Data Scientist: Day in the Life


hey guys Ken here back with another
video for you today I thought it would mix it up and do a V log style what it’s
like to work as a work from home data scientist as always it really helps when
I get interaction with my videos so please like if the video is interesting
to you and subscribe if you want to see more weekly content like this alright what’s up guys headed to the gym
this morning usually get up pretty early around 5:15 out the door by 5:40 when I
go to the gym I I go about three times a week usually Monday Wednesday and Friday
I start off with a little stretching routine for about twenty to thirty
minutes got to loosen up when you’re sitting at a desk for most
of the day you got to make sure that you’re pretty limber and you’re pretty
loose then I do so you know a little bit more fast-paced stuff depends on the day
I try and mix it up I usually hit the big muscle groups and then occasionally
I’ll go hit a couple golf balls after then I’ll go back home getting a little bit of exercise of
movement during the day is for me one of the most important parts helps me keep
an active mind especially when I’m working on pretty technical work I will
say that on the days I don’t go to the gym I have a little spin bike that you
probably see in there in the back of my videos and I’ll hop on that for 15-20
minutes in the morning and go through my same stretching routine alright so I
just got back from the gym after I work out I usually steam and get rid of all
the toxins in my body and then I take a cold shower to really wake me up get me
fired up for the day I start each morning by writing in my handy notebook
I write down all of my goals and a couple of things that I am planning to
do during that day I try and schedule out as much as I can
leaving a little bit of room for anything that could really happen today
I’m going to start with that I have a stand-up at around 9:45 with my team and
then for the rest of the day I’m doing a lot of QA on a dashboard that one of my
guys put together and I’m also working on thinking about how I can evaluate a
new model with that being said let’s do this so that is what a day my life is like
thank you so much for watching if you have any questions about what it’s like
to work from home what it is like to work as a data scientist please leave
them in the comments section below I like to post weekly I’m trying to get
into that cadence so look out for content every Thursday thank you so much
and have a great one

Data Scientist Salary 2019


Data scientists continue to be one of
the most in-demand jobs for 2019. In this data scientist salary guide 2019, find
out what you can expect to be paid based on market predictions and hiring data. Hi,
I’m Jen. Welcome to the channel! Glassdoor ranked data scientist as the number one
job for 2018 and there’s an expectation that will carry through into 2019 as
well. With a substantial salary, high job satisfaction, and an ever increasing
number of open positions, data science is a great career to enter or continue
growing in during 2019. If you’re curious to find out the difference between a
data scientist and a data analyst, check out the linked video. I talked about data
analyst salaries last week. I’ll link to that video in case you’re really
interested in a data analyst position. Entry-level data scientist salaries for
2019 consistently hover around $100,000 with $103,000 as the median base pay. This doesn’t factor in bonuses, benefits, and
other forms of compensation. At this level, you’re likely new to the role and
still developing your skills. You have some sort of an educational background
that’s prepared you for the role. You may also be slightly more experienced, but in
an area where there are few data scientist jobs or at a small company who
may not be paying as much to start out. Mid-level data scientist salaries for
2019 have a median pay of $121,000 per year.
At this level, you should have some experience and be fully capable of
performing the job on your own. You may have to consult someone with more
experience in extremely complex situations, but most of the day-to-day
work is stuff that you can do independently. Data scientists with much
higher levels of experience or specialized experience can expect to
make more. This is also true of markets where there are a lot of job openings
and few qualified data scientists to fill the role. This could also apply if
you’re entering a large company and especially one where there’s a lot of
complexity in the day-to-day work. For these data scientists, the median salary
is $147,000. In extremely competitive markets or ones
that require a significant amount of experience or multiple, specialized
certifications, pay can reach or exceed $200,000 per year.
All of these salaries are just base pay. This means benefits, bonuses, profit
sharing, those sorts of things, are all on top of this and they can easily add up
to tens of thousands of dollars or more. These salary levels also don’t account
for location differences which may have little difference or in certain
locations can have massive differences. For instance, if you’re in New York City
working as a data scientist, salary levels are a full 40%
than the numbers that I just shared. In contrast, if you’re one of the few data
scientists working in El Paso, you’re making or can expect to make about
30% less than these salaries. If you’re working as a data scientist
today, how does your salary compare? Maybe your pay is lower, but you have benefits
or bonuses that make up some of the difference. If you want to know the total
value of your compensation, check out the compensation worksheet in the link below.
You can input all of your financial compensation and it will automatically
calculate the value. If you’re watching this video as research because you’re
considering a data scientist job offer, check out our negotiation coaching. I’ll
work with you one on one to create a custom negotiation plan, walking you
through the entire process and teaching you the negotiation skills and
tactics to help you land a higher salary or more benefits and bonuses. This can
make the difference of literally hundreds of thousands of dollars
throughout the course of your data scientist career.

Automated trucks: Blue-collar disaster or economic win? | Andrew Yang


The big misconception about the impact of
technology in the workforce is thinking that it’s around the corner. Instead it’s been with us for years. If you look at the last 20 years or so, we’ve
automated away 4 million manufacturing jobs in Michigan, Ohio, Pennsylvania, Wisconsin,
Missouri, Iowa, all the swing states that Donald Trump needed to win in 2016 and did
win. Then my friends in Silicon Valley and my friends
who work in technology know that what we did to the manufacturing workers we are now going
to do to the retail workers, the call center workers, the fast food workers, the truck
drivers, and then even bookkeepers, accountants, insurance agents, lawyers, and on and on through
the economy. So what happened to the manufacturing worker
is a very clear sign of what’s going to happen to these other workers moving forward. And I talked a little bit about retail workers,
the most common occupation in the economy. Thirty percent of Main Street stores and malls
are going to close in the next five years because Amazon is soaking up $20 billion of
commerce every year. And many of these workers are making $11 to
$12 an hour and don’t have a huge savings cushion to be able to make meaningful adjustments. Being a truck driver is the most common job
in 29 states. There are 3 and 1/2 million truck drivers
in this country, average age 49, 94% male, average education high school or one year
of college. They’re making about $46,000 a year right
now. It’s one of the higher paying blue collar
jobs in this country. And on the west coast, you have my friends
in Silicon Valley who are trying to automate truck driving. And the reason they’re doing that is because
of the money — $168 billion in financial incentives for automating away truck drivers. And that’s not just labor savings. That’s also equipment utilization because
a truck never needs to stop whereas human-driven trucks have to stop every 14 hours; fuel efficiency
because trucks can convoy together in lower wind resistance and so robot trucks would
be able get places with less fuel, fewer accidents because truck drivers right now kill about
4,000 other motorists a year in accidents and that number would come down if you had
automated freight. So there’s a very, very powerful set of incentives
to try and automate truck driving as an occupation. Again, though, you have these 3 and 1/2 million
truckers, and only 13% of them are unionized. So there’s not going to be a grand negotiation. So imagine being a trucker who’s taken out
$50,000, $60,000 loan to lease your truck and it’s your livelihood and your means of
support, and then all of a sudden, you have to compete with a robot truck that doesn’t
need to sleep. And that is what is around the corner for
hundreds of thousands of truckers in this country in the next five to 10 years when
robot trucks start to hit our highways. And Amazon is testing out robot trucks as
we speak, right now in the Midwest.

AI Impact on Jobs & the Skills of the Future


Hi everyone, and welcome back. So we’ve been hearing more and more talk
about A.I. and how A.I. and automation will impact jobs in the future. From Trevor Noah to the Financial Times everyone
is talking about it, so we wanted to add our opinion into the mix. So how is A.I. for business going to affect
the future of work? And more specifically, can we identify those
jobs that are at more risk of being taking over by A.I. and automation? There’s a very wide dissonance on this. We started analyzing a bunch of different
reports that have been shared online like McKinsey reports, OECD studies, and, my personal
favorite, an Oxford study that said 47 percent of U.S.
jobs are at risk of automation over the next few years. Meanwhile, we see that the general population
and workers think differently. A recent study, conducted by Marist College,
actually identifies that 97% percent of workers believe that most jobs will be automated,
but not their own. This suggests that the general public needs
to be educated on which jobs are susceptible to this risk, which are not and businesses need to be aware of the forthcoming
skills gap. Of course, not all jobs are equal. The Oxford study that we cited a moment ago
actually highlights this. They examined 700 participants and found that
the generalist occupations that require creative knowledge or innovation are at least risk. The same is true for occupations in education,
healthcare, media and arts jobs. On the flip side, jobs like telemarketers,
junior lawyers, accountants are at most risk. In short, there is a simple rule of thumb:
if your job is in some way predictable or routine, the risk of automation is much higher. If a job doesn’t require innovation or creativity
than the return on investment for companies is higher on machines than real time employees. Machines are faster, can’t be distracted and
can work 24/7. This is actually good for creative marketers,
because A.I. and automation can serve to augment their jobs, rather than substituting them. Last month, McKinsey and the World Economic
forum published a white paper about the impact of emerging technologies on the creative economy. They stated that artificial intelligence is
changing creative content from beginning to end. By 2030, A.I. will be able to write high school
essays, code in Python, compose top 40th chart songs and make creative videos. But all these advancements also comes with
risks and costs. Take a look at this report by the Global Commission
on The future of work. In the absence of effective transition policies,
many people will have to accept lower-skilled and lower-paying jobs. High-skilled workers are taking less cognitively
demanding jobs, displacing less educated workers, And this is already happening! Also, technological dividends are being unevenly
distributed among firms. A very limited amount of companies tend to
dominate when it comes to “big data”. Just think about Google and Facebook. Today, they alone are responsible for 70%
of the referral marketing traffic and receive more than 50% of total, global advertising
budget. So the question is: can businesses, workers,
and social institutions go in the same direction? If companies and public policy leaders can
understand the evolving landscape they can help the workforce anticipate the
upcoming challenges. Technology and the demographic changes are
leading to a smaller workforce, compared to the previous generation, and the workforces has to pursue many careers
during their time of work. We need to provide workers with an environment
where they can continuously upskill and grow. Governments will have to re-evaluate the educational
system we will have to continuously learn and grow and companies will have to redesign their
structure and their culture around technologies Just like during the industrial revolution,
we are heading into a new age. In the great transformation that we are about
to see by 2020, it is estimated that 20% to 25% of the labour force will be displaced
within 10–20 years. However, this is also an opportunity for us
to get ahead. We have to find ways to attract and retain
highly skilled workers and allow them the time to upskill themselves, even during work
hours. We think that a good way to start is to develop a learning community so you can benefit from each other. And also to use technology to supplement your
goal tracking and your effort, instead of as a distraction. In short: what are you doing to bridge the
dissonance. Have you made a map of how A.I. and automation
will affect your industry and your company? If this is an economic imperative, how do
you feel about committing yourself to a lifelong approach to knowledge? Now, what is your opinion? Please, let us know your thoughts in the comments
below and stay sharp!

Cloud Conversations: Oracle’s Work with Emerging Tech


Big topic, artificial intelligence, everyone’s
talking about it. Oracle adaptive intelligence, how is this
being manifested in the cloud platform today? Yeah, we’re embedding AI
into every part of our cloud so we can democratize the access
and value from AI so it’s inside our integration cloud service, it’s
inside our analytics cloud service it’s inside our applications, our SaaS applications
so we can give you next best offer, how to do discounting based on what the
market data is saying, etc. So a very interesting discussion on AI,
and now there is an other big part to AI that you definitely want
to talk about. Well, its the autonomous database.
Of course! You know, people are spending so
much time patching database, provisioning databases, operating databases.
We want to take all of that away. With the autonomous database, we’ve created
a database that is self-driving, self-patching, self-scaling, self-healing.
Yes, self healing–exactly! So, the goal is, for example,
for a warehouse you don’t even have to worry about
setting indexes. You don’t even have to worry about
query optimization. Over the last two decades, we’ve learned
how to optimize the platform and we’ve learned how to optimize
the infrastructure upon which it is running So that you can just get
the value of the database.

What is Cognitive Technology? A Look at Real Business Applications


According to a recent analysis from LinkedIn,
2019’s employers are looking for a combination of both hard and soft skills.
Cloud computing and artificial intelligence are topping the list of desired attributes.
So is there a type of technology that combines both of these desired attributes?
The answer is: Yes! These are the so-called cognitive technologies. Hi! I’m Judit, A.I. Trainer and
Coach at Growth Tribe. Offered usually by the AI-first companies within their cloud platforms and by pioneering
startups, these cognitive technologies mimic human abilities such as vision, text analysis
and speech. More importantly, these applications make
use of top performing algorithms trained on loads of data containing images, videos and text. Basically companies like Google, Amazon and
Microsoft have been acquiring this data from society’s behaviour for years, and now, they
are selling this intelligence back to society as machine learning as a service (MLaaS).
They became part of the secret tool kit of the best analytics translators and those decision
makers responsible for matching a business problem with a feasible technology. Let me give you an example: SkinVision is an app that detects skin cancer
melanoma, which is based on computer vision supported by Amazon Web Services.
It has trained a proprietary algorithm to label melanoma pictures that is now being
used by more than 1 million active users. Users can take pictures with their smartphones
and scan their body frequently, under a subscription model. Considering that 1 in 5 citizens develop some kind of skin cancer, the service is now frequently
offered for free by insurers to their clients. So instead of developing an app to book a
visit to a dermatologist, it actually brings the specialist view directly to you in the
form of computer vision. A typical example of artificial narrow intelligence.
But you still need to know how to evaluate these models according to the cost of their
mistakes. It is still the job a human to judge these
mistakes in an analyst-in-the-loop approach. Just like when you buy a car, you need to
know how to compare different models and make diagnostics of engine failures. But you don’t
need to know how to build the entire car yourself. Another very interesting example comes from
DataSine’s Pomegranate image scoring. Their machine learning models have been trained
on images and text examples to understand how different content appeals to humans.
It offers suggestions on words and phrases to replace, as well as colours, themes
and images, depending on the personality of your users. We’ve seen some really nice examples from participants in our Growth & AI traineeship. Marwan from the Mobile Company, recently tested Pomegranate by comparing the suggestions of
the image scoring model with real click through rates of different ad campaigns.
Using this kind of image recognition, a skilled marketer will not need to run too many experiments
in order to find images and copywriting that actually generate a higher conversion rate. Another application of computer vision was tested by Merle from Son of a Tailor.
She used Google’s Auto ML Vision to build a prototype classification model for new t-shirts
based on their attributes, for example crew neck or v-neck.
After you set up the environment on GCP, there are no coding skills required, but the person
running the experiment can observe the performance metrics like precision and recall.
Or even observe how the model would perform on new t-shirts once deployed in an application. Microsoft Azure, similarly, has a set of cognitive services that can be used to detect specific
content in text and images. Another one of our trainees, Niels, from Eyecons,
tested the Content Moderator and review API to detect profanity and negative sentiment
in content. This content could be posted, for example,
by sport fans in forums and comments to articles. Then, a human-in-the-loop could react promptly
with the best communication strategy. Hope you enjoyed these examples of cognitive
technologies and are keen to know more about them, as well as the current top
professional skills: cloud computing and artificial intelligence. Don’t forget to like, subscribe, leave your comments below and see you next
time!