Understand Machine Learning Basics for Small Businesses

Hey everyone welcome to understand machine learning, basics for small business hosted by girl with google ill start by introducing myself. My name is bassem one of the instructors here at growth, google and im so excited to have you all here with us today um. So we want this to be as helpful as possible, so make sure to post your questions directly under this live stream. Um, as were going along well choose those top quoted questions so make sure to up vote. Your favorites, and so today were, of course were here talking about machine learning.

And how you can use it to enhance your business practices if we dont get to answer all of your posted questions, that’s, okay, our team will reply and answer all those questions directly in the q a shortly after the virtual event concludes also, if you’d like to Access the additional resources for this client um, this class click the resources tab at the top of the screen.

And youll also be able to access the replay of this virtual event in the future by clicking on demand on the growth google, on-air homepage and also dont forget to share on social using grow with google. Finally, you also have the option to turn on youtubes closed. Captioning to follow along, if you want to to do so, click the closed captioning button directly on the video player on your screen. Alright lets get started looking at our agenda today, so well explore the basics of machine learning and discover how to get started using

Machine learning at your business or for any other purpose. So first we will define machine, learning and discover uses for machine learning in the real world and then well go over some examples to understand how machines learn and identify ways. You might use machine learning yourself to solve a problem or overcome a challenge. Finally, well consider the implications of machine learning from an ethical perspective, so uh to succeed today, you dont need to have a technical background or any know-how on how to write code or even have strong

Foundations for mathematics uh, so when we talk about, you know getting started with machine learning. This is for anyone who has general interest in machine learning or is curious about how machine learning can improve their business processes and help them to become more profitable all right. So today, youll leave with a broad basic understanding of how machine learning works, so that, if you decide that machine learning is something you want to pursue for you can grow your business.

Youll have an idea of where to get started and where to start researching all right. So lets start with the basics. What do you think about when you hear the words machine learning so im going to? Let you mull on that little bit and kind of think about it all right. So machine learning is a science of using data to train computers.

So they can make decisions, perform tasks automatically and even improve um experience right. So, basically, with machine learning, computers dont need to be told exactly what to do. They can be given an objective and a lot of data and theyll teach themselves how to make connections uh to help find that answer or even to perform a task. So a little history here when you think about a typical like how typical computer works, a computer programmer writes a set of instructions for the computer to follow.

Right and then that computer follows its instructions right, so that’s, simple computer programming, uh, but so when you talk about by contrast, machine learning takes a bunch of examples, figures out patterns that explains the examples and then uses those patterns to make predictions about new examples. The computer programmer doesnt actually write the instructions or algorithms for the computer. Instead, the machine comes up with the algorithms and identifies the patterns itself, so it learns the patterns on its own and and changes those algorithms based on

The data the programmer gives it – and this is useful because we live in a world that produces a lot of data right. So when we talk about that its almost impossible to comprehend how much data is available in the world today, in fact the amount of data uh – you know the amount of information humans produce every single day is about 2 5 quintillion bytes of data that’s a lot Right and then most of the data is of course, anonymized

Which means its not tied to a particular person, but it can still be used to identify patterns or gain insight about the world uh, so that data can be almost anything um. You know what kinds of search terms people are using travel patterns uh. What people are watching and listening to even health data like a persons, steps or physical activity right? These are all different sources of data that can be tapped into uh for machine learning.

Right and so as the amount of data humans produce continues to grow exponentially, machine learning can help us to make sense of of it and then use it to help improve peoples lives, a computers, ability to process analyze, uh and learn from that information and then visualize. It and be able to understand it comprehend it and then interact with the world around it. This can be very, very valuable and help people to solve many real world problems. So youve probably already seen machine

Learning at work right, uh can, can you think of any business who you that who you use, that might already be using machine learning to improve your customer experience or even style problems. There are so many ways: businesses use machines, learning, learning models every day and and actually over the next couple slides. We will take a look and talk about some of the of those different ways and youve probably already encountered

Most of them, but really just kind of dive into how machine learning is actually already in our everyday life, so traffic predictions right, so youve all probably used gps applications that rely on machine learning. These models can identify the best route and even predict traffic conditions and travel time based on current and evolving road conditions. Like im sure you either use. You know google maps of traffic predictions to any route, optimization or even waze right, another google product and when you say

You know im looking to leave tomorrow, you know and be and arrive somewhere by lets, say 9am and waze will tell you, based on all of the traffic patterns that they have in that database. When is the best time for you to leave to arrive at your destination from where you are at 9am tomorrow, right and then this is all based on a machine learning to provide you

With the best information you can in order to get to that place on time, another example is manufacturers who use data to detect when equipment or other assets are operating at low efficiency or at risk for failure. Then they take actions before issues can occur, so the machines can see something happening, make predictions and say hey. You need to check this out before something bad happens, and so machine learning has huge implications for the health field as well. An example of machine learning model uh is one that has been trained to look

At a biopsy sample and identify any tumor cells, a tool like that uh for cancer and many other diseases can help doctors, diagnose patients uh quickly right much more quickly and even access to high quality medical care uh for everyone right uh. So one study actually used computers to assist exist, a diagnosis right. So when you talk like a cad when reviewing um early mammogram scans for for women,

Who actually later developed breast cancer right? The computer actually spotted 52 of the cancer as much as a year before women were actually officially diagnosed, which is incredible. Additionally, machine learning can be used to understand risk factors for disease in in large populations right, and so these are just some of the examples we talk about, how machine learning is being used and can be, can be used in

The future another application for machine learning is, of course, market analysis. This is probably one of the the best examples that are big examples. You see when people are using smart machines that can be trained to track trading volatility um or manage wealth and assets on be. You know, on behalf of an investor. These algorithms can identify trends more, you know efficiently than humans, and they can also react uh much quicker in real time and, of course, their voice interfaces.

Like google home to help people to manage tasks, you know activating new systems and access information, so lets take a look at an example of machine learning in action all right all right. So those are just a few examples on how machine learning uh is used in the real world, in fact, you’re surrounded by the effects of machine learning every day. Now that you have a better understanding of what machine learning is and how its used in the real world today lets talk about the process behind it right.

Uh ive been saying that machines learn, but it might not be clear to you exactly how they do so uh and so lets clarify by lets using an example right of how you actually learn right. So take a look at the math problem on this slide and what i want you to do is take a look at the problem and then try to finish this sequence. All right ill. Give you a second to kind of go through it and think you know: can you figure out

What the answer would be take this at my water and give a chance all right times up lets see what the answer is. All right. Did anyone get 81 right all right, so the answer is 81, but now for those of you who arrived at the answer, you pre probably figured it out for those who didnt um well lets kind of talk through it right. So, first you look at the first three responses and then identify the pattern right in the action. So its pretty much you know, do

Something what do i have to do two three: to make: none right, all right, so there’s, a couple things i can do and i figured out all right and then so now i need to what do i have to do to 4 to make it 16 right And so we know that we have, you know 3 times, 3 4 times 4, and then we started to figure out a pattern: 8 times 8 make 64 and then so what do we need to do? Um to

This 9 right and then so. The pattern says: you’re pretty much multiplying that number by itself or squaring it whatever you’re going to call it right and then that’s the the logic applied right. So you figure that out and and based on the data provided what that number should be, and then so you apply the same learning of patterns to a to any new situation right. You use your

Previous experience uh your previous knowledge to apply it to a new situation and that’s. You know pretty much how machine learning so machine learning uh replicates that same process, uh that you and your brain go through to get the answer to this math problem, and you know youve, learned and understood what the pattern was representing and you applied that learning to Your example right or number four, so machine learning focuses on developing computer programs that use data to create what we call models. These models build their own logic and insight based on the information

Or data that they receive and then so machine learning is trying to teach machines to learn from experience. Just like we did with our example right uh so by you, know, there’s a couple different ways right by fighting finding patterns in data making important connections and then applying that knowledge to future challenges right and then so for lets. Take a quick look at an example of that. So first thing to really kind of talk about is we need to remember that computer

Learning does not it does. It does have a few uh limitations, uh and if youve ever tried and failed to unlock your phone uh with fake face recognition, you know it doesnt always work perfectly. You know um. If, at any point you know the data uses has changed right, so your face has changed, maybe you’re growing, a beard or you’re wearing sunglasses, uh or todays. You know world the mask, you’re wearing masks and its like. Ah you know your facial unlock doesnt work anymore.

Or even just you know, its too low light or whatever it is there’s, not enough data uh for that phone to identify you right and so that’s a perfect example of a limitation machine learning uh. So machines provide a lot of value to humans, but theyre not uh brains right, so they are not creative or independent thinkers. Nor can they act on uh. What they’ve learned uh you’re like? What are you talking about right, so you may have been able to identify a friend with a mask or sunglasses. You can

Totally do it, you see someone hey hows, it going. You havent seen a long time, even if theyre wearing a mask or sunglasses, but your phone needs more data to be able to make that decision. Uh that’s, why our understanding of machine learning and best practices are very important when we talk about that all right. So to help you understand how much machine learning is used in the real world im going to tell you about how one business might solve a problem with machine learning. This is going to be a very simplified example.

But at least help you to understand the basics of machine learning and its applications, all right. So imagine um, mallory, right, uh, shes, a mechanic, an expert on all kinds of cars. She enjoys buying previously owned cars, uh that arent in the best shape, and then she repairs them updates them, and then she of course sells them for profit. Shes been really successful. At this, in fact, shes so successful that shes got so busy at work that she decides to. She wants to hire two assistants to help her out right.

However, malory soon discovers there’s a really big problem. Um, although she was, she was quick to be able to determine and knows looking at which cars to invest in her trainees or her new hires just dont have the experience so lets lets. Take a look. First of all, lets take a look at um, the reasons why mallory has been so successful during her years in the repair business

Shes learned how to identify the cars that have the most potential to be profitable, right and so based on her knowledge. You know she goes through and says all right. Let me consider a couple things: uh, the age of a car uh how well the previous owner took care of it. Is there any act in the history of the current mileage? Are these cars known to be reliable, uh the cost of fuel right that’s, a big consideration for some people of ownership and and just basically how the car was primarily driven?

Right, well, you know, that’s. These are some of the examples of things that she takes into consideration when shes looking to decide whether or not this car is going to be worth um. You know the effort going that goes into it. The two people mallory has hired will not have her experience. They dont have enough experience in the market um, but with machine learning um helping she can help them to determine which cars, at least to work on uh and then which ones to sell for profit. So first things first mallory needs to

Provide some data – and this is the this – is the the foundation to machine learning data and really the more data, the better right. So in mallorys case, this means compiling data on previously owned cars that were repaired and resold in her city over the past uh three years right. This image shows mallorys data compiled in a spreadsheet, and this is what we call mallorys

Training data right, its called training data because its the information she can use to train or to teach the computer right, so the data should so to get the data she pulled up all her records, uh of all the cars shes considered buying in the past uh. This data will include details like we talked about before you know. You can see the accident history service records all these different things in the

Different columns and so on now, if you look at column, i this column contains mallorys decision about whether each car uh was a profitable investment. This is the judgment that she wants to teach the computer to make right. So, basically, based on these factors – yes or no so mallory can create an app that uses machine learning to help her employees make informed choices about which cars to work on uh to create this tool. She can use a service like google clouds, ai platform and then she would prepare her training data and then build the machine. Learning.

Algorithm that will use the data to re. You know return the decision about whether or not to purchase this car. The machine learning provider will, of course, help validate the algorithm to make sure its returning good decisions and of course, finally, theyll help her to deploy the machine learning algorithm as the app shes looking for and then so then she her and her team can identify

Potentially valuable cars to repair much more quickly and with more profit in the end now this was a fictitious example, and we can. We will make believe that things went well for her and her new team. However, in the real world uh, she may have encountered some actual roblox uh. Can you think of you know for you car people out there think of some challenges that may have prevented mallorys model from actually working ill. Give you a couple of seconds to

Kind of think about that all right, so first things important to keep in mind uh with machine learning that any model is predicting the future based on historical patterns um in mallorys example, her models, uh, is judge the likely profitability of future automotive investments based on her Past investments, uh, nothing is certain right, that’s important understand, nothing is certain. No one and no computer can predict the future uh and that’s. Really important. 2020 is a perfect example of what we mean by that uh. Second, there may be a chance. Mallorys model was actually biased right for

Instance, what if somebody took really good care of their car, but just failed to hold on to service records? Or maybe you know they did everything this out? They did all the servicing themselves. They have no documentation on bringing it to the shop to get it repaired because they wanted to save money on doing something they could do on their own right and then so. Based on that, the model would rate that persons uh automobile negatively uh, basically because it has incomplete service records, so its very potential uh very possible

That mallorys team would actually miss out on a great investment. Also is there actually enough data present in mallorys model to produce reliable results? Thats, like i said before, data is a foundation to machine learning the more data the better uh. So shes worked on a lot of cars several dozen, but it may not be enough for her to get reliable results. Uh lots

Of machine learning models build on hundreds or thousands or even millions of examples in training data to make sure the machine learns the process properly. Mallory will definitely want to provide it with a lot more data than, of course, our example spreadsheet um as well, discuss later its also important to think about what kind of data mallory provides. So, for example, if she plans to invest in multiple kinds,

Of cars right so like hondas, chevys and chryslers, then shell need to make sure her training model includes. You know a representative range of data, not just simply one automotive brand. She needs to make sure shes providing data across all three of those brands or every brand shes. Looking something else to consider is that the automotive manufacturers discover new ways to actually make their cars more fuel efficient.

More reliable over time, so just because a certain brand was profitable in the past uh, you know, or maybe it wasnt problem pass. It doesnt mean that they can improve or even have a bad year uh for a particular brand uh. You know im sure youve had that conversation before people say: oh yeah, i just it was just a bad year for that. That particular

Brand or that particular model uh that can negatively impact the end results of that data, so as powerful as machine learning can be. A computer cannot replicate the two advantages that your brain already has. You can think critically and ask questions. For example, ask yourself: do i have enough data, and also is my data biased all right, so at this point, weve talked about several examples of how machine

Learning works in real life situations. Now let us talk to you about how you might actually use machine learning in your own lives. So were going to give three specific examples of goals you might want to set for machine learning all right. So maybe you’d like to better understand your customers, so you can provide a more personalized experience and heighten customer service or machine learning can help you optimize inventory to ensure you dont, run out of items or have too much of a certain item on the shelf. You

Might even want to enhance security by identifying potential forward or even suspicious account behavior, all right so im going to give you a second to now think of your own objectives. What problems do you have in your business that you think could be solved with machine learning and now understand? Not all issues can be solved with machine learning, but brainstorming. Some ideas will help you to begin working with a machine learning.

Solution or maybe a provider can actually to help you really figure out what problems can be solved and can help you to advance your business. So as you’re brainstorming, remember the term problem like i said that before uh is like a math problem, we used earlier uh. This is a logical question. You want the computer to solve it. Doesnt have to be a huge problem for your business. It just might be a specific goal you want to accomplish with machine

Learning so, if you’re having any trouble thinking of your own objectives, so here are a few examples to get. You started all right so, like anticipating the needs of your customers right, so a training model for lets say an electronics business. Might let you know that people who buy a tv right on the new flat screen tv

They might need a mount to be able to mount it on the wall right, so um, monitoring your digital security right so create a machine learning model to protect your company against fraud and suspicious activity. Uh talk about delivering targeted, personalized marketing that’s a big thing. Google is really good at doing that. Uh automate, repetitive tasks, very simple thing like invoicing or

You know bill whatever it is. You know invoices billing, you know time whatever it is uh, just something, that’s repetitive, what you know, let the computer take care of it or even addressing customers. Uh questions when we im sure youve all experienced this before when you talk about the um, the ai chat bots on a lot of different websites.

You go in uh. Youre is answering all your questions. You havent actually spoke to a real person, but that means that that bot knows the answer to question. Based on what you typed in and using all that data, you can use those those thoughts to alleviate um. The amount of time your your representatives are spending on the time asks answering simple questions right. You can have an ai

Chatbot, take care of that and then get someone on the line, a live personal line uh when they definitely need that one-on-one with a human being right so now ill. Give you a few moments to think about um on your own right, so im sure youve been thinking right now. I definitely encourage you to write them down, as you think of them, so im going to give you a minute or two to write down.

Some problems that you have big or small and then well move on to the next slide. Okay, so once youve thought of your problem, decide on the specific outcomes you hope to achieve. So what are your goals right? And you look at this? So what problem you’re facing be very specific and then, when you talk about your goals, you need to be specific about that as well great. Now you have a few ideas about what kinds of problems you might be.

Trying to solve using a machine learning model next, you need to identify the data that will help you to address your problem, so you need to start thinking about what kind of data you want to collect to address this issue, and so let me just give you A couple of examples here on on that all right lets just use mallory right so lets say: mallory was a restaurant investor who had to decide where to open new restaurants in different parts of the city? Can you think of

The type of data that she might want to include in our model to make a better decision ill, give you a couple seconds to think about it and then well give you some of our suggestions, all right, so these, like i said these, are just some examples Here, but for this project you know you can include the current restaurants in each neighborhood success rates of the restaurants and the different neighborhoods or even proximity of locations to mass

Transit see some things that can help uh, really that machine learning model to make better decisions right, uh, so lets take this other example here so and lets say if mallory was a hospital administrator who wanted uh to help healthcare providers to make better evidence-based decisions. Uh more quickly right what kind of data would be most useful to her ill? Give you a couple seconds to think of your

Examples and then well show you some of the ones that we have all right. So for those of you who are hospital administrators things like um, patient history, right results of medical decisions or even the effect of of drugs or treatments – and these are all examples of data that um mallory could have used for machine learning as she builds these different Businesses and remember that mallory would need to make sure her data represents a large enough portion of the

Population in order for her decisions or fed up machine learning model decisions to be accurate all right. So the usefulness of machine learning model is entirely dependent on the data right that’s given by the human right and its only accurate as the data used to make the model. In other words, if you change your data, you will change your outcome right and that’s. A perfect example here lets take a look here at an example. So imagine a business that makes

Winter sports care right: they want to predict peoples buying habits, so they want to know how many skis to make for the upcoming season. So what do you think would happen if the computer model is made only using data about people who live in vermont? Now, understanding that this manufacturer is looking to sell skis all across the united states or maybe even all across the world right if they give data specifically about only the people who live in vermont, uh its going to be very biased, uh, simply because uh, the data Of people in vermont is, is

Snowing it has lots of hills and mountains for skiing right and so the machine learning model after going through all this data will conclude that a large percentage of the overall population wants to purchase, skis right, they all purchase keys, but what, if that computer model same Exact model was made using data about people who live all over the country or all over the world right you’re going to get different um in results right, so the outcomes for these models would likely prevent uh, very different percentages of interested buyers. Based on that data set uh versus the vermont

Data set so its very important understand that your data really will drive the end results and then, when we talk about this example, its very biased towards skiers, or at least people who live in snowy mountainous areas. Now, if i was a small ski shop in vermont and of course, i know that my main um target group is going to be in vermont

Then that may work out well right, but just really understand who your overall audience is uh, what you’re looking to do with that and then make sure you have a data set that supports that all right. So when you’re creating a model, its important that you use data that’s appropriate to all kinds of the problems you want to solve, if the data you provide, the computer is too limited like a vermont example or

It includes too much of one kind of thing in the training data. The machine learning model will not be able to reliably predict outcomes right. So if this sports memorabilia company wants to predict how many hats to make for its whole business, which serves the whole country, the the whole country uh, it will need a similarly broad set of

Data uh: this principle applies to any kind of training model uh. If mallory wants to predict whether chevys or foreign cars are good investments, then she has to make sure a model includes more data than just information about ford. Sometimes your data set should be specific. Like i said before in that vermont example, if the ski shop is in vermont, then that would work really well for determining in that vermont population. How many skis should maybe a

Store purchase for the upcoming season um, but so like another example, is like. If a healthcare provider wants to learn, uh what you know what kinds of risk factors contributes to womens breast cancer. Then the data given to it should be about women and, probably not about men all right so that’s why its important to use, data that’s. Relevant, unbiased and fair and representative of lots of diverse perspectives and experiences right. A primary driver of machine learning technology is

data in the end your data strategy will be the biggest factor contributing to your potential for success with machine learning all right so now lets think about your own situations what kind of data do you think you will actually need to be able to tackle this problem right again machine uh you know giving machines the best possible data to learn from is essential in machine learning so now.

think about your training data uh and even think about like you know examples we gave you uh like marilyns training data for her her um computer for her car investment initiative and then take a few moments and think about what kind of data that you might use uh and so if you’re having issues like once again so here are some common training data.

you might use based on the needs of of your business right so lets say a business who wants to use machine learning for personalized marketing might need a data set including customers um their purchase history right an accounting company that wants to use machine learning for auditing would need a data set including the company the companys financial transactions a small business that wants to improve its customers.

service like with chatbots might need transcripts of previous customer service shots in order to teach the computer how a human might respond to those frequent customer requests so these are all examples right so just think about those examples brainstorm for a little bit and then well move on to our next discussion all right so finally lets talk about.

the ethical implications of how we use machines machine learning is improving uh and improving lives around the world right but its also raising new questions about the best way to build fairness privacy and security into these systems and there are some strategies that can help you to do this all right so in order to ensure you have a variety of perspectives in any machine learning project and increase the number of people who benefit from the actual technology be sure to engage with a diverse set of.

users and use case scenarios and make sure to incorporate peoples feedback before and throughout the project development so as we discussed previously data bias is another important consideration its important to understand the limitations of your data um your data in the dataset model always keep in mind that machine learning models are reflection of the training data and lastly laws about privacy and.

security vary from country to country make sure you’re aware of the local laws about privacy and security to ensure you’re keeping your users data safe all right so this video shows how much machine learning platform uh tensorflow is helping farmers in tanzania uh ensure the health of one of the most important food crops tensorflow is an open source machine learning platform that can help you to get started with.

machine learning quickly but this video will give you a sense of how machine learning can be used in all kinds of industries to improve outcomes and to make peoples lives and businesses better cassava is a really important crop it provides for over 500 million africans every day when all other crops fail farmers know that they could rely on their cassava plants to provide them food there are several diseases that affect cassava and these diseases makes the roots unedible it is very.

crucial to actually control and manage these diseases so were trying to use machine learning to respond to those diseases and tensorflow is the best foundation for our solutions the app that weve designed can diagnose multiple diseases its called nuru swahili for light the light that farmers can use to see their problems and find solutions you wave your phone over a specific leaf look at it and if it has.

a symptom the box will pop up saying you have this problem when you get a diagnosis we have an option for you to get advice and learn about the best management practices the object detection that we use through tensorflow relies upon our team annotating images weve collected over 5000 high quality images of different cassava diseases for this project we use a single shot detector model on a mobilenet architecture its able to make predictions in less than one second instead of having to implement thousands of lines of code tensorflow provides a.

library of functions that allow us to build architectures in much less time we need something that can be deployed on a phone without any connection tensorflow is able to shrink these neural networks to be able to fit on your mobile device the human input is absolutely critical were really building something that augments your experience and makes you better at your job so with ai tools and machine learning you can improve the yields you can protect your crops.

and you can have a much more reliable source of food ai offers the prospect to fundamentally transform the life of hundreds of millions of farms around the world you can see a product that can actually make someones life better this is kind of revolutionary in this workshop we covered what machine learning is and some of the basics on how it works by using data to create a model and apply that model to new situations and we talked about how machines learn right and how you might use machine.

learning to solve a problem in your business and also what kinds of data you might need in order to get that problem solved and then we considered some of the ethical implications of working with this kind of data so we talked about your our next steps now weve identified your problem and you start thinking about the data you might need so what do we do now right uh you’re most likely to succeed.

if you start small so choose just one problem and work to finding a solution using machine learning secondly there are lots of resources available to help you along the way and so were going to talk about those resources now so if you’re interested in learning about machine learning you want to check out these resources so learn with google ai includes educational content for.

people who want to get started with machine learning this is a website full of learning resources and it includes you know filters so you can actually search for your particular role and the kind of content you actually want and also you know what stage of process you’re in uh when you talk about ai adventures is a youtube channel hosted by google cloud you can use it to dive deeper into machine learning and also how it works and then google cloud has partnered with.

uh coursera their world-class education platform to design several courses on machine learning if you’re interested you can use these courses to learn more about how google does machine learning and how to structure your own machine learning project uh if you want hands-on uh like you know a hands-on experience and exploring how machine learning works definitely check out googles teachable machine its a web-based tool that creates fast models without any coding and that’s at teachablemachine with google com if you’d like to learn more skills you can apply to your life.

your career google has training resources that you can access remotely there are many different learning paths and options so well talk about these tools and you can determine which ones makes the most sense for you right now so you can continue to grow your skills no matter where you’re located through our online workshops and you can learn the skills that can help you stay connected and productive and grow your skills while of course preparing for the new career you want so if you’re curious about additional workshops you can take virtually from.

home visit g dot co slash grow on air for our current virtual digital workshops in their descriptions definitely be sure to check out our workshops um that are available on demand like maybe even something like google analytics that gives you a little idea about how machine learning works primer right uh google primer if you want to quickly sharpen your skills you.

can definitely check out uh primers lessons on finding jobs and career advancement primer is a free app that you can download on your phone apple ios or android it doesnt matter and its you know there’s just its an app with that offers five-minute free lessons right focus on whatever your goal is right so if you want a little bit of machine learning about growing your business marketing whatever it is.

definitely check out primer theyre fun easy lessons um you can download today apply digital skills this is uh free once again free training curriculum that helps you with so much more that you want to learn if you talk about things like maybe data and uh data analysis and research communication uh to resume writing there’s lessons for everyone from adult learners all the way down to middle school students its all free you dont have to sign in you can just start taking lessons or if you want to track your progress you can sign in with your.

google account and then keep track of your progress as you go along the google work space certification training right so to get more in-depth training uh training on the collaboration features that google has to offer we offer free or workspace training you can even get an official certification you can see the badge right there in google workspace if you like to add.

it to your credentials your resume to your social whatever it is anyone can complete all of the google workspace training at no cost through the apply digital skills website like i talked about before um even if you dont plan on getting your workspace uh certification uh these lessons will at least help you to get familiar with all of the google workspace applications some of the ones you see right there.

listed like drive gmail docs sheets and slides all right well thank you for joining us today im going to pass it over to anna who will answer some of the questions you have today well see you next time thanks vasim hi everyone im anna im excited to answer a few of the questions that you asked throughout the session and weve received quite a few remember if we dont get to your question well try to go back and answer as many questions as possible.

on the grow with google on-air page shortly after the virtual event concludes all right lets get started my first question comes from bruce bruce asked how can i start setting up machine learning to improve my companys customer service performance great question you can actually visit our contact center ai page for everything you need to know about integrating machine learning into.

your company companys customer service page and you can find that at cloud gov solutions slash contact center again that’s cloud google com solutions slash contact dash center okay my next question comes from nick nick asks where can i learn about different machine learning providers that’s a great question there are many different machine learning providers google cloud and tensorflow are some examples and you can learn more about google cloud at.

cloud google com slash products slash ai okay my next question comes from natasha how much does it cost to set up machine learning another fantastic question once you know the products that you want to use for machine learning you can visit the google cloud pricing site which tells you exactly what you pay for each product and you can find that site at cloud google com pricing slash list all right my next question comes from peter peter asks.

how can i learn about local laws around what data i can collect on my customers very important question for my small business owners in the united states there is no single law about data protection and privacy there are many different ones on a both federal and state level so its important to do a search for the data privacy laws in your individual state countries outside of the united states also have widely varying laws around consumer data make sure that you know the laws before you start collecting any data okay my next question comes.

from tony tony asks do i need coding or programming knowledge to use machine learning tools great question tony you wont need to know how to code to use machine learning tools but you will need to become familiar with how they work in order to use them well luckily there are lots of opportunities to learn online or even consult with a google cloud specialist to set up your machine learning system also dont forget to click into the.

resources at the top of the screen where it says resources and there are some really helpful links there including a course from coursera okay my next question comes from carol carol asks is machine learning the same as artificial intelligence artificial intelligence or what we call ai and ive seen that floating a little bit around on the q a is the field of study that focuses on machines that can mimic some aspects of human intelligence like the ability to reason learn or predict machine learning is a specific branch of.

artificial intelligence its the part of ai that actually teaches machines to learn all right last but certainly not least my next question comes from steve steve asks what kinds of industries currently use machine learning all kinds of industries steve so that could include retail government agencies education financial services and small and medium businesses you probably interact every day.

with organizations or companies that are using machine learning all right those are all of our questions for today thank you so much for joining us for understand machine learning basics for small businesses hopefully you all learned something new you can find our follow-up resources including some links to some youtube videos as well as a course from coursera where you can learn more about machine learning on the page that.

you’re watching where it says resources above the video player you can also access a recording of todays session by clicking on demand at g dot co slash grow on air well also be sending a follow-up email with a survey for you to fill out regarding your experience in the virtual workshop today its so important that we can get feedback so we can quickly improve these sessions and offer the content that’s.

most helpful to all of our learners thanks again for joining us i grow with google dont forget that you can sign up for upcoming workshops on the grow with google on our home page at g dot co slash grow on air see everybody next time you