A i *°

TensorFlow was originally a deep learning research project of the Google Brain Team that has since become—by way of collaboration with 50 teams at Google—a new, open source library deployed across the Google ecosystem, including Google Assistant, Google Photos, Gmail, search, and more. With TensorFlow in place, Google is able to apply deep learning across numerous areas using perceptual and language-understanding tasks.

This cheat sheet is an easy way to get up to speed on TensorFlow. We'll update this guide periodically when news and updates about TensorFlow are released.

SEE: How to become a developer: A cheat sheet (TechRepublic)

SEE: How to build a successful developer career (free PDF) (TechRepublic)

Executive summary

What is TensorFlow? Google has the single greatest machine learning infrastructure in the world, and with TensorFlow, Google now has the ability to share that. TensorFlow is an open source library of tools that enable software developers to apply deep learning to their products.
Why does TensorFlow matter? AI has become crucial to the evolution of how users interact with services and devices. Having such a powerful set of libraries available can enable developers to include this powerful deep learning evolution to their products.
Who does TensorFlow affect? TensorFlow will have a lasting effect on developers and users. Since the library was made open source, it is available to all developers, which means their products can be significantly enhanced to bring a higher level of intelligence and accuracy to their products.
When was TensorFlow released? TensorFlow was originally released November 9, 2015, and the stable release was made available on February 15, 2017. Google is now rolling out the alpha of TensorFlow 2.0, which, according to Google, is simpler and more intuitive to use.
How do I start using TensorFlow? Developers can download the source code from the TensorFlow GitHub repository. Users are already seeing its effects in the Google ecosystem.

SEE: How to implement AI and machine learning (ZDNet special feature) | Download the free PDF version (TechRepublic)

What is TensorFlow?

When you have a photo of the Eiffel Tower, Google Photos can identify the image. This is possible thanks to deep learning and developments like TensorFlow. Prior to TensorFlow there was a division between the researchers of machine learning and those developing real products; that division made it challenging for developers to include deep learning in their software. With TensorFlow, that division is gone.

TensorFlow delivers a set of modules (providing for both Python and C/C++ APIs) that enable constructing and executing TensorFlow computations, which are then expressed in stateful data flow graphs. These graphs make it possible for applications like Google Photos to become incredibly accurate at recognizing locations in images based on popular landmarks.

SEE: All of TechRepublic's cheat sheets and smart person's guides

In 2011, Google developed a product called DistBelief that worked on the positive reinforcement model. The machine would be given a picture of a cat and asked if it was a picture of a cat. If the machine guessed correctly, it was told so. An incorrect guess would lead to an adjustment so that it could better recognize the image.

TensorFlow improves on this concept by sorting through layers of data called Nodes. Diving deeper into the layers would allow for more and complex questions about an image. For example, a first-layer question might simply require the machine to recognize a round shape. In deeper layers, the machine might be asked to recognize a cat's eye. The flow process (from input, through the layers of data, to output) is called a tensor...hence the name TensorFlow.

What is TensorFlow 2.0?

Google is in the process of rolling out TensorFlow 2.0, which includes the following improvements:

Helps make API components integrate better with tf.keras (a high-level interface for neural networks that runs on top of multiple backends).
Includes TensorFlow.js version 1.0, which allows the use of off-the-shelf JavaScript models, can retrain existing JS models, and enables the building and training of models directly in JavaScript.
Includes TensorFlow Federated, which is an open source framework for experimenting with machine learning (and other computations) using decentralized data.
Includes TF Privacy, a library for training machine learning models with a focus on privacy for training data.
Features eager execution, which is an imperative programming environment that evaluates operations immediately, without building graphs before returning concrete values.
Uses tf.function, which allows you to transform a subset of Python syntax into portable, high-performance graphs, and improves performance and deployability of eager execution.
Advanced experimentation will be made possible with new extensions Ragged Tensors (the TensorFlow equivalent of nested variable-length lists), TensorFlow Probability (a Python library built on TensorFlow that makes it easy to combine probabilistic models and deep learning), and Tensor2Tensor (a library of deep learning models and datasets).
A conversion tool that automatically updates TensorFlow 1.x Python code so that it can be used with TensorFlow 2.0 compatible APIs (and flags cases where said code cannot be automatically converted).

Additional resources

Google launches TensorFlow 2.0 Alpha (ZDNet)
Your life in AI's hands: The battle to understand deep learning (TechRepublic)
Understanding the differences between AI, machine learning, and deep learning(TechRepublic)
Google Cloud adds NVIDIA Tesla K80 GPU support to boost deep learning performance(TechRepublic)
Didi Chuxing to build 'intelligent transport ecosystems' using AI and deep learning (ZDNet)

Why does TensorFlow matter?

Machine learning is the secret sauce for tomorrow's innovation. Machine learning, also called deep learning, is considered a class of algorithms that:

Use many layers of nonlinear processing units for feature extraction and transformation; and
are based on the learning of multiple levels of features or representations of the data; and
learn multiple levels of representation corresponding to different levels of abstraction.

More about Open Source

Thanks to machine learning, software and devices continue to become smarter. With today's demanding consumers and the rise of big data, this evolution has become tantamount to the success of a developer and their product. And because TensorFlow was made open source, it means anyone can make use of this incredible leap forward brought to life by Google. In fact, TensorFlow is the first serious framework for deep learning to be made available through the Apache 2.0 license.

With developers and companies able to use the TensorFlow libraries, more and more applications and devices will become smarter, faster, and more reliable. TensorFlow will be able to sort through vast numbers of images at an unprecedented rate.

Because Google made TensorFlow open source, the libraries can be both improved upon and expanded into other languages such as Java, Lua, and R. This move brings machine learning (something heretofore only available to research institutes) to every developer, so they can teach their systems and software to recognize images or translate speech. That's big.

Additional resources

Google eliminates more spam from Gmail with TensorFlow (ZDNet)
Video: Google on how TensorFlow Lite will help build better AI for smartphones (TechRepublic)
Why machine learning benefits the rich, and everyone else is toast (TechRepublic)
Deep learning for customer service is overhyped: How one company did it faster and cheaper(TechRepublic)
7 companies that used machine learning to solve real business problems (TechRepublic)
Baidu trained an AI agent to navigate the world like a parent teaches a baby (TechRepublic)
Facebook unveils Big Basin, new server geared for deep learning (ZDNet)

Who does TensorFlow affect?

TensorFlow not only makes it possible for developers to include the spoils of deep learning into their products, but it makes devices and software significantly more intelligent and easier to use. In our modern, mobile, and 24/7 connected world, that means everyone is affected. Software designers, developers, small businesses, enterprises, and consumers are all affected by the end result of deep learning. The fact that Google created a software library that dramatically improves deep learning is a big win for all.

Additional resources

How to become a developer: A cheat sheet (TechRepublic)
The future of AI: 10 scenarios IBM is already working on (TechRepublic)
Here's Google's open secret to cloud competition (TechRepublic)
Samsung announces Bixby, new AI assistant for Galaxy S8 (TechRepublic)
Microsoft's AI system for training autonomous cars and drones goes open source(TechRepublic)
Fujitsu to build Riken's new deep learning supercomputer (ZDNet)
Research: Companies lack skills to implement and support AI and machine learning (Tech Pro Research)

When was TensorFlow released?

TensorFlow was originally released November 9, 2015, and the stable release was made available on February 15, 2017. TensorFlow 2.0 alpha is available now, with the public preview coming soon. You can learn more about the TensorFlow 2.0 alpha in the official Get Started with TensorFlow guide.

The libraries, APIs, and development guides are available now, so developers can begin to include TensorFlow into their products. Users are already seeing the results of TensorFlow in the likes of Google Photos, Gmail, Google Search, Google Assistant, and more.

Additional resources

Why IBM's speech recognition breakthrough matters for AI and IoT (TechRepublic)
Could new G Suite features entice enterprise users to switch from Office 365? (TechRepublic)
How AI and next-generation genomic sequencing is helping cancer patients (TechRepublic)
How to become an IoT developer: 6 tips (TechRepublic)
Microsoft buys deep-learning startup Maluuba (ZDNet)

Competitors to TensorFlow

TensorFlow isn't alone in the deep learning field; in fact, there are a number of other companies with machine learning frameworks, including the following.

Additional resources

Caffe2: Deep learning with flexibility and scalability (ZDNet)
IBM, NVIDIA partner for 'fastest deep learning enterprise solution' in the world (TechRepublic)
The cloud war moves to machine learning: Does Google have an edge? (TechRepublic)
5 big data trends that will shape AI in 2017 (TechRepublic)
Artificial intelligence: The 3 big trends to watch in 2017 (TechRepublic)
IBM, Nvidia target deep learning, AI workloads (ZDNet)

How do I start using TensorFlow?

The first thing any developer should do is read the TensorFlow Getting Started guide, which includes a TensorFlow Core Tutorial. If you're new to machine learning, make sure to check out the following guides:

Developers can install TensorFlow on Linux, Mac, and Windows (or even install from source), or check out their various tools from the official TensorFlow GitHub page.

Finally, developers can take advantage of all the TensorFlow guides:

Additional resources

Google and Udacity launch free course to help you master machine learning (TechRepublic)
Free machine learning courses from Google, Amazon, and Microsoft: What do they offer? (Tech Pro Research)
How to find the real value in operational big data (TechRepublic)
How machine learning helps the United Nations monitor global events (TechRepublic)
How to ease your big data experts' fears about AI automating their jobs (TechRepublic)
How to use Swarm AI instead of polls for market research (TechRepublic)
How deep learning will transform the future of the auto industry (ZDNet)

Artificial intelligence, which has been around since the 1950s, has seen ebbs and flows in popularity over the last 60+ years. But today, with the recent explosion of big data, high-powered parallel processing, and advanced neural algorithms, we are seeing a renaissance in AI—and companies from Amazon to Facebook to Google are scrambling to take the lead. According to AI expert Roman Yampolskiy, 2016 was the year of "AI on steroids," and its explosive growth hasn't stopped.

SURVEY: Take this artificial intelligence and machine learning survey, and get free copy of the research report

While there are different forms of AI, machine learning represents today's most widely valued mechanism for reaching intelligence. Here's what it means.

SEE: Research: Companies lack skills to implement and support AI and machine learning (Tech Pro Research)

Executive summary

What is machine learning? Machine learning is a subfield of artificial intelligence. Instead of relying on explicit programming, it is a system through which computers use a massive set of data and apply algorithms to "train" on—to teach themselves—and make predictions.
Why does machine learning matter? Machine learning systems are able to quickly apply knowledge and training from large data sets to excel at facial recognition, speech recognition, object recognition, translation, and many other tasks.
Who is affected by machine learning? Machine learning touches industries spanning from government to education to health care. It can be used by businesses focused on marketing, social media, customer service, driverless cars, and many more. It is now widely regarded as a core tool for decision making.
When did machine learning become popular? The term "artificial intelligence" was coined in the 1950s by Alan Turing. Machine learning became popular in the 1990s, and recently came into the public eye when Google's DeepMind beat the world champion of Go in 2016.
How are businesses using machine learning? Business applications of machine learning are numerous, but all boil down to one type of use: Processing, sorting, and finding patterns in huge amounts of data that would be impractical for humans to make sense of.
What are the security and ethical concerns about machine learning? AI has already been trained to bypass advanced antimalware software, and it has the potential to be a huge security risk in the future. Ethical concerns also abound, especially in relation to the loss of jobs and the practicality of allowing machines to make moral decisions like those that would be necessary in self-driving vehicles.
How can our business use machine learning? Businesses like IBM, Amazon, Microsoft, Google, and others offer tools for machine learning. There are free platforms as well.

SEE: Sensor'd enterprise: IoT, ML, and big data (ZDNet special report) | Download the report as a PDF (TechRepublic)

What is machine learning?

Machine learning is a branch of AI. Other tools for reaching AI include rule-based engines, evolutionary algorithms, and Bayesian statistics. While many early AI programs, like IBM's Deep Blue, which defeated Garry Kasparov in chess in 1997, were rule-based and dependent on human programming, machine learning is a tool through which computers have the ability to teach themselves, and set their own rules. In 2016, Google's DeepMind, beat the world champion in Go by using machine learning—training itself on a large data set of expert moves.

Here are several kinds of machine learning:

In supervised learning, the "trainer" will present the computer with certain rules that connect an input (an object's feature, like "smooth," for example) with an output (the object itself, like a marble).

In unsupervised learning, the computer is given inputs and is left alone to discover patterns.

In reinforcement learning, a computer system receives input continuously (in the case of a driverless car receiving input about the road, for example) and constantly is improving.

A massive amount of data is required to train algorithms for machine learning. First, the "training data" must be labeled (for instance: a GPS location attached to a photo). Then it is "classified." This happens when features of the object in question are labeled and put into the system with a set of rules that lead to a prediction. For example, "red" and "round" are inputs into the system that leads to the output: Apple. Similarly, a learning algorithm could also be left alone to create its own rules that will apply when it is provided with a large set of the object—like a group of apples, and the machine figures out that they have properties like "round" and "red" in common.

SEE: What is machine learning? Everything you need to know (ZDNet)

Many cases of machine learning involve "deep learning," a subset of machine learning that uses algorithms that are layered, and form a network to process information and reach predictions. What distinguishes deep learning is the fact that the system can learn on its own, without human training.

Additional resources

Top 5: Things to know about machine learning (TechRepublic)
Understanding the differences between AI, machine learning, and deep learning(TechRepublic)
Video: Is machine learning right for your business? (TechRepublic)
Microsoft says AI and machine learning driven by open source and the cloud (ZDNet)
IBM Watson: The inside story of how the Jeopardy-winning supercomputer was born, and what it wants to do next (TechRepublic cover story)
Google AI gets better at 'seeing' the world by learning what to focus on (TechRepublic)
IT leader's guide to deep learning (Tech Pro Research)

When did machine learning become popular?

Machine learning was popular in the 1990s, and has seen a recent resurgence. Here are some timeline highlights.

2011: Google Brain was created, which was a deep neural network that could identify and categorize objects.
2014: Facebook's DeepFace algorithm was introduced. The algorithm could recognize people from a set of photos.
2015: Amazon launched its machine learning platform, and Microsoft offered a Distributed Machine Learning Toolkit.
2016: Google's DeepMind program "AlphaGo" beat the world champion, Lee Sedol, at the complex game of Go.
2017: Google announced that its machine learning tools can recognize objects in photos and understand speech better than humans.

SEE: A guide to data center automation (ZDNet special report) | Download the report as a PDF(TechRepublic)

Additional resources

Linux Foundation pushes open-source AI forward with the Deep Learning Foundation (ZDNet)
Alibaba neural network defeats human in global reading test (ZDNet)
Facebook drives 2.5x performance boost in machine translation using neural networks(TechRepublic)
Google DeepMind: A cheat sheet (TechRepublic)
How Google's DeepMind beat the game of Go, which is even more complex than chess(TechRepublic)

Why does machine learning matter?

Aside from the tremendous power machine learning has to beat humans at games like Jeopardy, chess, and Go, machine learning has many practical applications. Machine learning tools are used to translate messages on Facebook, spot faces from photos, and find locations around the globethat have certain geographic features. IBM Watson is used to help doctors make cancer treatment decisions. Driverless cars use machine learning to gather information from the environment. Machine learning is also central to fraud prevention. Unsupervised machine learning, combined with human experts, has been proven to be very accurate in detecting cybersecurity threats, for example.

SEE: All of TechRepublic's cheat sheets and smart person's guides

While there are many potential benefits of AI, there are also concerns about its usage. Many worry that AI (like automation) will put human jobs at risk. And whether or not AI replaces humans at work, it will definitely shift the kinds of jobs that are necessary. Machine learning's requirement for labeled data, for example, has meant a huge need for humans to manually do the labeling.

There are several institutions dedicated to exploring the impact of artificial intelligence. Here are a few (culled from our Twitter list of AI insiders).

The Future of Life Institute brings together some of the greatest minds—from the co-founder of Skype to professors at Harvard and MIT—to explore some of the big questions about our future with machines. This Cambridge-based institute also has a stellar lineup on its scientific advisory board, from Nick Bostrom to Elon Musk to Morgan Freeman.
The Future of Humanity Institute at Oxford is one of the premier sites for cutting-edge academic research. The FHI Twitter feed is a wonderful place for content on the latest in AI, and the many retweets by the account are also useful in finding other Twitter users who are working on the latest in artificial intelligence.
The Machine Intelligence Research Institute at Berkeley is an excellent resource for the latest academic work in artificial intelligence. MIRI exists, according to Twitter, not only to investigate AI, but also to "ensure that the creation of smarter-than-human intelligence has a positive impact."

Additional resources

IBM Watson CTO: The 3 ethical principles AI needs to embrace (TechRepublic)
Machine learning helps science tackle Alzheimer's (CBS News)
New research shows that Swarm AI makes more ethical decisions than individuals(TechRepublic)
AI gone wrong: Cybersecurity director warns of 'malevolent AI' (TechRepublic)
Russian facial recognition program beats Google, but big privacy questions linger(TechRepublic)
AI stops identity fraud before it occurs (TechRepublic)

Who is affected by machine learning?

Just about any organization that wants to capitalize on its data to gain insights, improve relationships with customers, increase sales, or be competitive at a specific task will rely on machine learning. It has applications in government, business, education—virtually anyone who wants to make predictions, and has a large enough data set, can use machine learning to achieve their goals.

Additional resources

Demand for AI talent exploding: Here are the 10 most in-demand jobs (TechRepublic)
Here are the 10 most in-demand AI skills and how to develop them (TechRepublic)
The 6 most in-demand AI jobs, and how to get them (TechRepublic)
Cheat sheet: How to become a data scientist (TechRepublic)
73% of developers who don't use AI plan to learn how in 2018 (TechRepublic)
MIT's automated machine learning works 100x faster than human data scientists(TechRepublic)
Apple, IBM add machine learning muscle to enterprise iOS pact (ZDNet)

How are businesses using machine learning?

2017 was a huge year for growth in the capabilities of machine learning, and 2018 is poised to be a huge year for business adoption, according to research from Deloitte.

One of the things that may be holding that growth back, Deloitte said, is confusion—just what is machine learning capable of doing for businesses?

There are numerous examples of how businesses are leveraging machine learning, and all of it breaks down to the same basic thing: Processing massive amounts of data to draw conclusions much faster than a team of data scientists ever could.

Some examples of business uses of machine learning include:

Alphabet-owned security firm Chronicle is using machine learning to identify cyberthreats and minimize the damage they can cause.
Airbus Defense & Space is using ML-based image recognition technology to decrease the error rate of cloud recognition in satellite images.
Global Fishing Watch is fighting overfishing by monitoring the GPS coordinates of fishing vessels, which has enabled them to monitor the whole ocean at once.
Insurance firm AXA raised accident prediction accuracy by 78% by using machine learning to build accurate driver risk profiles.
Japanese food safety company Kewpie has automated detection of defective potato cubes so that workers don't have to spend hours watching for them.

SEE: Executive's guide to AI in business (free ebook) (TechRepublic)

Any business that deals with big data analysis can use machine learning technology to speed up the process and put humans to better use, and the particulars can vary greatly from industry to industry.

AI applications don't come first—they're tools used to solve business problems, and should be seen as such. Finding the proper application for machine learning technology involves asking the right questions, or being faced with a massive wall of data that would be impossible for a human to process.

Additional resources

5 tips to overcome machine learning adoption barriers in the enterprise (TechRepublic)
How the NFL and Amazon unleashed 'Next Gen Stats' to grok football games (TechRepublic)
Predictive analytics: The smart person's guide (TechRepublic)
How ML and AI will transform business intelligence and analytics (ZDNet)
IBM Watson: Six lessons from an early adopter on how to do machine learning (TechRepublic)
AI and the Future of Business (ZDNet special feature)
How to launch a successful AI startup (TechRepublic)
The practical applications of AI: 6 videos (TechRepublic)

What are the security and ethical concerns about machine learning?

There are a number of concerns about using machine learning and AI, including the security of cloud-hosted data and the ethical considerations of self-driving cars.

From a security perspective, there are always concerns about the theft of large amounts of data, but security fears go beyond how to lock down data repositories.

Security professionals are nearly universally concerned about the potential of AI to bypass antimalware software and other security measures, and they're right to be worried: Artificial intelligence software has been developed that can modify malware to bypass AI-powered antimalware platforms.

Several tech leaders, like Elon Musk, Stephen Hawking, and Bill Gates, have expressed worries about how AI may be misused, and the importance of creating ethical AI. Evidenced by the disaster of Microsoft's racist chatbot, Tay, AI can go wrong if left unmonitored.

SEE: Machine learning as a service: Can privacy be taught? (ZDNet)

Ethical concerns abound in the machine learning world as well; one example is a self-driving vehicle adaptation of the trolley problem thought experiment. In short, when a self-driving vehicle is presented with a choice between killing its occupants or a pedestrian, which is the right choice to make? There's no clear answer with philosophical problems like this one—no matter how the machine is programmed, it has to make a moral judgement about the value of human lives.

Along with whether giving learning machines the ability to make moral decisions is correct, there are issues of the other major human cost likely to come with machine learning: Job loss.

If the AI revolution is truly the next major shift in the world, there are a lot of jobs that will cease to exist, and it isn't necessarily the ones you'd think. While many low-skilled jobs are definitely at riskof being eliminated, so are jobs that require a high degree of training but are based on simple concepts like pattern recognition.

Radiologists, pathologists, oncologists, and other similar professions are all based on finding and diagnosing irregularities, something that machine learning is particularly suited to do.

There's also the ethical concern of barrier to entry—while machine learning software itself isn't expensive, only the largest enterprises in the world have the vast stores of data necessary to properly train learning machines to provide reliable results.

As time goes on, some experts predict that it's going to become more difficult for smaller firms to make an impact, making machine learning primarily a game for the largest, wealthiest companies.

Additional resources

Facebook's Zuckerberg: Here's how we'll fix our massive data privacy problems (TechRepublic)
Why AI bias could be a good thing (TechRepublic)
How machine learning can prevent factory machine breakdowns and create more jobs(TechRepublic)
Artificial data reduces privacy concerns and helps with big data analysis (TechRepublic)
Can AI really be ethical and unbiased? (ZDNet)
Artificial Intelligence: Legal, ethical, and policy issues (ZDNet)

How can our business use machine learning?

There are many online resources about machine learning. To get an overview of how to create a machine learning system, check out this series of YouTube videos by Google Developer. There are also classes on machine learning from Coursera and many other institutions.

And to integrate machine learning into your organization, you can use resources like Microsoft's Azure, Google Cloud Machine Learning, Amazon Machine Learning, IBM Watson, and free platforms like Scikit.

Additional resources

Google offers free 15-hr machine learning crash course as part of AI resource center(TechRepublic)
Microsoft offers developers pre-built machine learning models for Windows 10 apps(TechRepublic)
DevOps, machine learning dominate technology opportunities this year (ZDNet)
Cloud AutoML: How Google aims to simplify the grunt work behind AI and machine learning models (ZDNet)
Amazon AI: Cheat sheet (TechRepublic)
Facebook's machine learning director shares tips for building a successful AI platform(TechRepublic)
AI helpers aren't just for Facebook's Zuckerberg: Here's how to build your own (TechRepublic)
How developers can take advantage of machine learning on Google Cloud Platform(TechRepublic)
How to prepare your business to benefit from AI (TechRepublic)

In September 2010, a three-person AI startup called DeepMind Technologies launched in London, with the goal of "solving intelligence." Four years later, Google acquired the company for $500 million. And by 2016, it had achieved a major victory in AI: Mastering the complex game of Go.

This story represents the fantasy of many AI researchers, eager to launch their own ventures in the AI startup space. But the field has become saturated, and the terms "AI," "deep learning," and "machine learning" are often overhyped and misunderstood. Companies and VCs often hear these buzzwords but don't know what, exactly, it is that they are investing in.

So how can you start a successful company, grounded in AI, that can rise above the noise?

Prateek Joshi, an immigrant entrepreneur, recently launched his own startup called PlutoAI. Hailing from a small town in India, Joshi realized that water quality is critical to the health of a community. His company was created to address water wastage, predict quality, and lower operating costs at water facilities—by using AI.

Here are four tips from Joshi from his experience getting an AI startup off the ground.

1. Don't sell AI

"Every single company you talk to is doing some kind of AI," said Joshi. "The problem is, it gets a bad rap since many companies don't even know what they mean when they say 'AI.'" In order to build a successful AI company, said Joshi, "you shouldn't sell AI to customers." Instead, he said, "AI is a tool you use to solve problems."

More about artificial intelligence

2. Think outside of the box

A lot of AI research, said Joshi, is focused around image recognition, voice recognition, and robotics. "But what people don't realize is AI is a fantastic tool to solve many other problems," he said. Joshi recommends that entrepreneurs start looking for important problems to address, to see how AI can contribute to solutions.

If you remove AI from the company, said Joshi, and still have a valuable product, you're on the right track. But "if AI is your only thing, then neither the customers nor investors will be excited about it," he said. "AI is too hyped—and once the hype dies down, your company shouldn't die down."

3. Highlight your mission

Since the market is so saturated, said Joshi, many businesses that want to invest in AI are struggling to choose the right solution. "They are going after AI companies, but they have a thousand options to pick from," he said. "In some cases, what happens is they're like, 'There's so much hype. I don't want to get burned, so I'm just not going to do it.' And that's the worst outcome."

SEE: Google's DeepMind 'Lab' opens up source code, joins race to develop artificial general intelligence

"You don't want people to stop believing in data science or AI just because of a few bad apples," he said.

AI shouldn't be a part of your story, said Joshi. The story should be about the mission. "Your customer will use you because you save them something, make their life easier, save them money, save them time," he said."AI is a thing that you use to enable that."

4. Understand your customers

In Silicon Valley, people "get so wrapped up in their own technology that they forget why would the customer would buy it," said Joshi. "They spend like a year building it, and then realize, 'Whoops, you know what? The customer didn't even need it.'" Before you start building, he said, try to understand the needs of the customer. "Silicon Valley is dominated by engineers, and they start writing code before talking to customers," said Joshi. "You should do the opposite of that."

Also, when selling to industries such as water or manufacturing, "learn how to articulate your mission of origin in terms of something they'd understand," said Joshi. "If you say 'Hey we are building this new Deep Learning algorithm that can be parallelized on GPUs,' they'll stop listening. Removing the tech jargon from your story is very important if you want to grow as a business."

Next Big Thing Newsletter

Be in the know about smart cities, AI, Internet of Things, VR, autonomous driving, drones, robotics, and more of the coolest tech innovations. Delivered Wednesdays and Fridays

Sign up today

Also see...

Google DeepMind: The smart person's guide (TechRepublic)
How Google's DeepMind beat the game of Go, which is even more complex than chess(TechRepublic)
Google Deepmind AI tries it hand at creating Hearthstone and Magic: The Gathering cards(TechRepublic)
Google's DeepMind expands NHS partnership to improve eye health (ZDNet)
Google uses DeepMind AI to reduce energy use at data centers and save money(TechRepublic)
OpenAI, DeepMind open source AI training platforms (ZDNet)

With the boom in artificial intelligence (AI) affecting virtually every industry, there has been an explosion in the research and development of machine learning, a subfield of AI. And, perhaps, no company better illustrates what machine learning is capable of than Google's DeepMind.

More about Innovation

Founded in London in 2010 by Demis Hassabis, Shane Legg, and Mustafa Suleyman, DeepMind has developed machine learning systems that uses deep neural networks, reinforcement learning, and systems neuroscience-inspired models. The startup was purchased in January 2014 by Google for a reported 400 million, with Hassabis remaining CEO of DeepMind.

SURVEY: Take this artificial intelligence and machine learning survey, and get free copy of the research report

Instead of relying on explicit programming, DeepMind applies general-purpose learning algorithms to a large data set in order to "train" the system and make predictions.

So, how does DeepMind really work, and what is it capable of? This comprehensive guide explains what the program really does.

SEE: All of TechRepublic's cheat sheets and smart person's guides

Executive summary

What is DeepMind? Google DeepMind is a machine learning system that uses algorithms based on deep neural networks and reinforcement learning to train on massive datasets to be able to predict outcomes.
Why does DeepMind matter? Google DeepMind is a prominent example of machine learning that illustrates what advanced AI is capable of.
Who does DeepMind affect? Anyone from businesses to computer scientists to engineers to end-users will be impacted by machine learning. The principles used by DeepMind can be applied to businesses that want to improve efficiency, gamers who want to learn how to master certain games, like Go.
When is DeepMind happening? Google DeepMind came into the public eye in October 2015 when it beat the European champion of Go, marking a breakthrough in artificial intelligence that came a decade earlier than many experts predicted.
How can I take advantage of DeepMind? While Google isn't sharing all the details of its machine learning secrets, much of its code is open source, and it has also shared the code of its software called TensorFlow, a deep learning engine. DeepMind also publishes many academic papers about its work online.

SEE: Research: Companies lack skills to implement and support AI and machine learning (Tech Pro Research)

Google's AlphaGo beats Lee Sedol at the game of Go

Image: Screenshot by Max Taves/CNET

What is DeepMind?

DeepMind is a subsidiary of Google that focuses on artificial intelligence. More specifically, it uses a branch of AI called machine learning, which can include approaches like deep neural networks and reinforcement learning to make predictions. This can rely on massive data sets, sometimesmanual data labeling—but sometimes not.

Many other AI programs like IBM's DeepBlue, which defeated Garry Kasparov in chess in 1997, have used explicit, rule-based systems that rely on programmers to write the code. However, machine learning enables computers to teach themselves and set their own rules, through which they make predictions.

In March 2016, DeepMind's AlphaGo program beat world champion Lee Sedol in 4 out of 5 games of Go, a complex board game—a huge victory in AI that came much earlier than many experts believed possible. It did this through combining "Monte-Carlo tree search with deep neural networks that have been trained by supervised learning, from human expert games, and by reinforcement learning from games of self-play," according to Google.

But beyond mastering games, deep learning has other more practical applications. In 2012, it was used to recognize a million images with a 16% error rate—which is now at about 5.5%. Deep learning is also used in text-based searches and speech recognition. According to founder Mustafa Suleyman, it achieved a "30% reduction in error rate against the existing old school system. This was the biggest single improvement in speech recognition in 20 years, again using the same very general deep learning system across all of these."

Deep learning is also used for fraud detection, spam detection, handwriting recognition, image search, speech recognition, Street View detection, and translation. According to Suleyman, deep learning networks have now replaced 60 "handcrafted rule-based systems" at Google.

Additional resources:

Machine learning: The smart person's guide (TechRepublic)
Is 'data labeling' the new blue-collar job of the AI era? (TechRepublic)
How Google's DeepMind beat the game of Go, which is even more complex than chess(TechRepublic)
Google Deepmind AI tries it hand at creating Hearthstone and Magic: The Gathering cards(TechRepublic)
Google AI gets better at 'seeing' the world by learning what to focus on (TechRepublic)

Why does DeepMind matter?

While Google DeepMind's accomplishments in the gaming world are impressive, the implications of its machine learning platform are far-reaching. An announcement that DeepMind was able to slash Google's electricity bill by increasing energy-efficiency has big implications in both economic and environmental realms.

Also, DeepMind's partnership with the National Health Service, part of DeepMind Health, employs machine learning in spotting critical conditions in eye health. It hopes to eventually use algorithms to personalize health care treatments, determining which work best on patients, given their previous medical history.

Additional resources:

Who does DeepMind affect?

DeepMind's machine learning platform has implications for just about any organization that wants to capitalize on its data to gain insights, improve relationships with customers, increase sales, or be competitive at a specific task. It has applications in government, business, education—virtually anyone who wants to make predictions, and has a large enough data set, can use machine learning to achieve their goals. It also has the potential to create many jobs in data labeling as well as disrupt jobs that were traditionally done manually.

Additional resources:

Why AI is the 'agent of the economy': EmTechDIGITAL leaders show global impact of AI(TechRepublic)
Google to take on Slack and Facebook with new AI-powered chat, says report (TechRepublic)
Why Facebook wants to use AI to track your conversations online (TechRepublic)
Workaholics: Let Google's machine learning tool solve your work-life balance problems(TechRepublic)
Facebook embraces AI to get businesses and customers talking (TechRepublic)

What is the timeline for DeepMind?

Machine learning, popular in the 1980s, has seen a recent resurgence. Google DeepMind is one step in the chain of AI platforms that use neural networks to make predictions. Here are some highlights.

2010: DeepMind is founded in London, England.
2011: Google Brain was created, which was a deep neural network that could identify and categorize objects.
2014: Facebook's DeepFace algorithm was introduced, which could recognize people from a set of photos.
2015: Amazon launched its machine learning platform and Microsoft offered a Distributed Machine Learning Toolkit.
2016: Google's DeepMind program "AlphaGo" beat the world champion, Lee Sedol, at the complex game of Go.

Additional resources:

Google tailors A.I.-powered search for enterprise (ZDNet)
How Google's DeepMind beat the game of Go, which is even more complex than chess(TechRepublic)
Microsoft Build: 5 big moves you need to know (TechRepublic)
Google Cloud Platform signs up enterprise giants, how does it compare to AWS?(TechRepublic)
IBM Watson takes on cybercrime with new cloud-based cybersecurity technology(TechRepublic)

How can I take advantage of DeepMind and machine learning?

There are many online resources for machine learning. To get an overview of how to create a machine learning system, a series of YouTube videos by Google Developer has come in handy for me. There are also classes on machine learning from Coursera and many other institutions.

And, to further integrate machine learning into your organization, you can use resources like Microsoft's Azure, Google Cloud Machine Learning, Amazon Machine Learning, IBM Watson, and free platforms like Scikit.

Additional resources:

Facebook's machine learning director shares tips for building a successful AI platform(TechRepublic)
AI helpers aren't just for Facebook's Zuckerberg: Here's how to build your own (TechRepublic)
IBM Watson: What are companies using it for? (ZDNet)
How developers can take advantage of machine learning on Google Cloud Platform(TechRepublic)
How to prepare your business to benefit from AI (TechRepublic)
Executive's guide to AI in business (free ebook)

Google Sites

Report abuse