devopsy.kinsta.cloud

Category: AI

  • Moogsoft Advances AIOps Workflow

    Moogsoft Advances AIOps Workflow

    Moogsoft is moving to make it easier to create custom workflows around its AIOps platform while expanding the number of integrations provided to include connectors to New Relic Insights, Microsoft Teams and proxy support for polling integrations with IT platforms Zenoss, Zabbix, vCenter, vSphere, Solarwinds, Spectrum and SevOne.

    Gabe Coelho-Kostolny, senior product management and product marketing leader at Moogsoft, said a Workflow Engine that has been added to version 7.2 of the Moogsoft AIOps platform makes it possible to trigger actions such as notifications, ticket creation and other automated tasks within Moogsoft and external systems. Those actions can be employed within incident management workflows ad other automated remediation tools, he said.

    Moogsoft has also added tools that make it possible to visualize alert clustering algorithms in real time and, if needed, fine-tune them. That capability is intended to provide IT operations teams with more confidence in the recommendations being generated by the AIOps platform, Coelho-Kostolny said.

    Finally, Moogsoft is making available an analytics dashboard dubbed Individual Statistics, which enables managers to drill down from the team level to better understand the workload and key performance indicators of each team member.

    Coelho-Kostolny noted there is already a high correlation between organizations that have adopted best DevOps practices and those moving to implement AIOps. The complexity of many DevOps environments is in many cases forcing the AIOps issue because all the alerts being generated creates a level of fatigue among human IT administrators that results in them not being able to keep pace with the overall rate of change, he said. The concern now is no longer so much whether human IT administrators will be replaced by artificial intelligence (AI) as much as it is enabling IT teams to stay on top of all the tasks required. The opportunity now, he said, is to make it easier for organizations that have embraced DevOps to also customize the AIOps platform.

    Of course, AIOps platforms presuppose that IT organizations have access to massive amounts of accurate data about the IT environment. Collecting that data often winds up being a bigger challenge than most organizations appreciate. In general, Coelho-Kostolny said it’s clear AIOps platform adoption is reaching a tipping point in the enterprise. Organizations, however, need to be careful distinguishing between what is truly artificial intelligence and what may be simply a rules-based engine that has been added to an existing IT service management (ITSM) platform, Coelho-Kostolny said.

    Regardless of the AI path chosen, the ongoing shortage of IT expertise with critical skills means organizations will continue to experiment with AI. The real issue is determining to what degree of confidence IT teams will have in those AI platforms. After all, if the assumptions an AI system is relying on to make a recommendation prove faulty, the potential for harm tends to be a lot greater than any mistake a single human IT administrator might ever make on their own.

    — Mike Vizard

  • How AI is Transforming Testing Today

    How AI is Transforming Testing Today

    Currently, there is much debate about how artificial intelligence (AI) will transform our lives. One of the technology’s key benefits lies in its power to transform traditional approaches to app development and software testing. Gartner identified AI-driven development as one of its top three strategic technology trends for 2019.

    Digital transformation is dependent on better-quality software, and software testing matters more than ever. However, the growing complexity of our software-driven world means that new, smarter and quicker forms of testing are critical in the race to build better software faster. But in a hyper-connected and complex environment, this is no simple task, and this is where AI-driven testing can help.


    AI Transforming Testing

    Like any new technology, there are exaggerated claims. However, when it comes to testing, AI is already delivering real improvements and making intelligent automation a reality. AI has moved far beyond an academic concept and is driving business benefits as opposed to hints to those organizations that have adopted it. Below are four of the key ways that it’s already transforming testing.

    Auto-Generation of Test Scripts

    There are undoubtedly complex test cases for which AI can’t auto-generate all the code. However, AI can easily auto-generate the code that clicks the buttons on the screen, fills out a form, logs into the app and other similar functions. Organizations that have deployed AI-driven testing are able to auto-generate 80% of code reliably, delivering substantial efficiency improvements. Along with auto-generation, AI achieves auto-maintenance: Tests are rebuilt when something in the digital product changes, ensuring continuous quality is maintained and further reducing the resource burden on human testers.

    Test Optimization

    If you’ve ever used Amazon, then you know that recommendation engine technologies are not that futuristic. AI can be applied to testing in the same way to help companies determine which test is most likely to find a defect based on the risk information gathered. Rather than taking a haphazard approach to testing, AI enables you to home in on the areas at risk—not only delivering efficiency gains but also ensuring a quality experience that delights.

    Release Impact

    Neural networks, combined with test history and data from current test runs, can predict how an upcoming release will impact users. For example, will customer satisfaction go up or down? Armed with this information, companies can make any adjustments necessary to ensure their users continue to be delighted by their digital experience.

    Customer Impact

    This is a key benefit of deploying AI-driven testing as you can determine the impact on the business from release updates. Neural networks, machine learning and other AI techniques can be used to understand how technical factors impact business performance. For example, are load times negatively impacting conversion rates? By understanding how releases will affect the business, you can make changes that will quickly have a positive impact.

    What Does the Future Hold for AI-Powered Testing?

    In addition to the core benefits that AI-driven testing is already delivering, organizations can expect to see additional benefits within the next 12 months that will further accelerate their ability to fully automate the entire testing life cycle.

    Failure Determination

    Current testing compares the result of a test with a pre-defined known good result. That methodology simply will not work with dynamic systems, such as an internet of things (IoT) device that interacts with multiple other devices which are constantly changing. AI is a key part of the solution: A well-trained AI system can determine whether a real user would consider the actual result correct or not.

    Persona Testing

    One size doesn’t fit all, and the ability to understand how your app reacts under different personas will allow organizations to understand what parts of the app are particularly good or bad for different personas. Armed with this information, you can optimize your app for your target audiences, ensuring an experience that delights.

    As every organization races toward digital transformation, delivering a high-quality digital experience is essential for survival. To create compelling digital experiences, companies need tools that can navigate applications intelligently, learn where quality issues are most likely to occur and identify the critical data correlations that will help developers resolve issues quickly. This is where AI is already delivering significant advantages to organizations that have adopted it.

    It’s clear that AI-driven testing is not a passing fad. Rather than taking a wait-and-see approach, it’s time to jump in and experience how AI can help transform app development and software. With AI-driven testing making a real and measurable impact on the business, you can’t afford to be late to this revolution.

    — Antony Edwards

  • Are Self-Service Machine Learning Models the Future of AI Integration?

    Are Self-Service Machine Learning Models the Future of AI Integration?

    DevOps teams seeking to step up their mojo in developing cutting-edge artificial intelligence (AI) features are facing a big skills bottleneck when it comes to data analytics and machine learning modeling. As a result, the market is seeing an influx of self-service machine learning models and machine learning-as-a-service offerings designed to help development teams more easily integrate AI capabilities into their software.

    This is coming in direct response to an explosion in demand for AI capabilities in the enterprise. According to Gartner analysts, AI adoption in the enterprise tripled in the past year. A report last fall from MIT Sloan Management Review and Boston Consulting Group found that 91% of enterprises believe that AI will deliver new business growth to them by 2023.

    The trouble is that folding AI functions and predictive analytics into software requires a whole new level of expertise in data science and machine learning from cross-functional DevOps teams. They need added skills to choose the right algorithmic approaches, acquire and manage data, train the models and integrate them into the code base and underlying infrastructure so that everything works properly under the hood.

    “We have heard from customers everywhere that they want to adopt machine learning but struggle to actually get models into production,” said Eric Boyd, vice president of cognitive and AI for Microsoft.

    Ali Farhadi Xnor
    Ali Farhadi, Xnor.ai CEO

    DevOps leaders today already face an uphill battle to keep their developer ranks staffed with well-trained software engineers. Piling on additional requirements for very specialized machine learning and data science expertise further strains those recruitment efforts. According to the most recent Harvey Nash/KPMG CIO Survey, the top one and two technical realms that suffer the biggest skills shortages today are in data science and AI, with 46% and 38% of CIOs respectively reporting recruiting pain in those areas.

    “There’s a huge imbalance between the demand in the market and the supply of the very best AI experience to actually train these models,” said Ali Farhadi, CEO of Xnor.ai, a Seattle firm that just last week released a self-service platform for developers. Called AI2Go, the platform provides pre-trained deep learning models to quickly integrate AI features such as facial recognition and object classification directly into their software.

    “We want to enable anyone who can code to benefit from AI,” Farhadi said.

    The Xnor release comes just a few weeks on the heels of a big launch from Microsoft of several self-service machine learning model tools aimed firmly at MLOps teams—those DevOps teams implementing machine learning capabilities. The new services from Microsoft roll out capabilities to manage the code, data and environments used by MLOps teams throughout the machine learning deployment life cycle.

    “Azure Machine Learning service’s MLOps capabilities provide customers with asset management and orchestration services, enabling effective ML lifecycle management,” explained Jordan Edwards, senior program manager for Azure MLOps, in a blog post on the topic. “With this announcement, Azure is reaffirming its commitment to help customers safely bring their machine learning models to production and solve their business’s key problems faster and more accurately than ever before.”

    Tools such as these are not always going to obviate the need for data scientists and machine learning/AI specialists on DevOps teams. As Farhadi noted, even a simple service like AI2Go requires development teams to choose the right models for the right situations. Nevertheless, this burgeoning market could take the pressure off teams by minimizing the scale of specialized recruiting, while reducing bottlenecks and dreaded re-work.

    “Machine learning projects usually involve several people—data engineers, data scientists, DevOps and others. Conventionally, at each stage of the machine learning life cycle, most, if not all, of these people end up doing an excess of custom coding, redundant work and manual trial and error,” said Craig Stewart, senior vice president of product management and product marketing for SnapLogic, which last fall introduced a self-service machine learning add-on for its integration platform. He noted that one of the primary goals is speed, to help cross-functional teams “build, train, validate and deploy high-performing models faster.”

    — Ericka Chickowski

  • DevOps Chat: We Are Devo, With Walter Scott

    DevOps Chat: We Are Devo, With Walter Scott

    If you have massive amounts of data, Devo wants to help. The Madrid-based company specializes in the handling and processing of massive amounts of data.

    In this DevOps Chat we spoke with Devo CEO Walter Scott about the use cases for Devo in a data-crazy world. Walter is a seasoned CEO in the tech sector. He knows a problem when he sees it and he thinks Devo is the answer. Have a listen to why and check out Devo at devo.com.

    As usual, the streaming audio is immediately below, followed by the transcript of our conversation. Enjoy!

    Transcript

    Alan Shimel: Hey everyone, this is Alan Shimel, devopsy.kinsta.cloud. You’re listening to another DevOps Chat. Really happy to have a new company, a new friend, joining us on DevOps Chat today. I have Walter Scott, CEO of Devo. Walter, welcome to DevOps Chat.

    Walter Scott: Thanks, Alan.

    Shimel: So, Walter, you and I met recently in Iceland, of all places, at a Insight IGNITE event that really kinda focused on DevOps, cloud, cybersecurity, if you will, the topics, but it was my first exposure to Devo. I don’t know if our audience is familiar with Devo. Why don’t you–give us the foundation, man. “We are Devo.” What are you guys? What do you do?

    Scott: Well, Devo is a data analytics platform that’s the next gen of technology, that provides a data fabric to allow our customers to take all of the noise that exists in the machines and petabytes, and petabytes of data coming in, for them to really start getting signals so that they can start their automation or their AI or their machine learning process, to make their companies more efficient; to be able to have them act more quickly; as well as to be able to get the people who use our platform, to get them the time back that they spend in managing the underlying data platform and allowing them to add the value to drive IT, to drive that value back to the business.

    Shimel: Excellent. And, you know, Walter, one of the things that impressed me with Devo was this wasn’t your sorta typical Silicon Valley story—a couple of wiz kids come up with a PowerPoint business plan and get funded via VCs, pre-revenue, pre-product even. Devo—the underlying technology here, the company’s been around a while and it’s not a Silicon Valley company, right?

    Scott: That is correct.

    Shimel: Give us a little background on that.

    Scott: Well, you know, over my six stints of being a CEO, I’ve never really run an American-based company. There’s great IT being built all over the world. I’ve done them from Malta to Scotland, Germany to Armenia to now with us here in Madrid, is where the company started. And we had a not Silicon Valley startup, but it was five guys—or four guys and a girl or woman—and started off the product in a research key, to really build something special and change the market because the bank that they worked for had to completely change the way they do business.

    And over the course of three years in research and we finally launched a product in 2014 and have been three-figure type growth every single year, since the date of the company being founded, ’cause it’s such a large market. And it’s really accelerating because of what we’re seeing happening with businesses and trying to get ahead of everything from getting arbitrage out of supply chain to solving their security problems, just to start with basic log management and monitoring to consolidate tools and increase agility.

    Shimel: Sure. So, Walter, you were—as you alluded to, this is your sixth or seventh startup where you’ve come in as CEO and had some tremendous success. And, look, not everyone is a success, right? That’s the mantra in business, but you keep plugging. When we look at Devo—and you and I spoke a little bit about this in Iceland, but so much of what we see as, quote unquote, “AI” today is really machine learning. And so much of machine learning is based, really, on data, right? You’re only as good as your data. In order to find patterns, you have to have enough data to analyze it and make actionable kinda calls out of it.

    But this is not a new problem and there’ve been other solutions. What makes Devo’s solution—I don’t know if you use the term “next-gen” or “new-gen” or something like that, but what makes Devo’s solution different?

    Scott: Well, you know, this has been going around a long, long time and you’ve heard me say, in the absence of data, the machine can’t learn. And I go back, all the way back, to the Walmart days in the ’80s, when they built—I would argue … them as the first person to really take the competitive advantage, coming from their data warehouse, right? And —

    Shimel: I agree.

    Scott: —They actually had the line, they could tell your daughter’s pregnant before you could by looking at the data. Well, today, that’s become necessary, but, over the last 10 years, it’s really—what’s happened with the cloud is being able to get that data manageable. I don’t like the term “AI.” I like more “augmented intelligence” because we’re not at the place—or even narrow AI ’cause I think it can solve very specific business problems that people might have. We have one customer that needed to actually build some ML and machine learning to detect what a bot was, were hitting their site, because, when they launched a new product, the bots would come in and sell it all in the secondary market. They wanted that brand loyalty to go to the end users.

    Or we have another customer that has a bunch of IoT devices integrated into their sporting equipment, where they’re trying to figure out if somebody’s gonna get hurt and using machine learning to be able to define what is that case, in terms of repetitive hits, but they need that data over long scales. And just the data science is actually making that possible. It could be done in the past; the problem was it just wasn’t cost-efficient enough to do it.

    We have customers that are taking literally over a trillion events a day to try to—and building machine learning models to understand which customers may leave their environment, based upon dropped calls or poor support integration. And it’s not just having that one set of data; it’s about having hundreds of different sets of data to help prove those hypotheses, to get those more and more efficient. And the data scientists are spending 70% to 80% of their time getting their arms around the data. We take that problem away from them and allow them to have—to solve that other 80% now, they can put in training their models and getting data to make them more and more efficient.

    Shimel: Excellent. Great answer. Hey, you know, Walter, you and I, this is neither one of our first rodeos. And, when I look at your job, your mission over there, at Devo, one of the things that I guess it’s a double-edged sword ’cause it presents so many opportunities, but, at the same time, there’s an axiom in business—“Put all your wood behind one arrow,” right? There are so many use cases for what you can do here, right? There’s the cybersecurity case, right? Where, “Look, man, we need this data. We need to be logging everything. We need to find if we’re under attack, where the attack’s coming from, what’s happening?,” right? So there’s that. And one can spin up an entire company around just using this kind of technology in cyber.

    You alluded to a few other use cases, right? That had nothing to do with cybersecurity. Better business intelligence. Understanding your customers better. Understanding your product better. There’s only so many targets, though, that you can go after, right? Without a tremendous—you know, you wanna build an IBM kind of force or something or a Dell force, that’s a whole different story. So how do you—Walter, as the CEO, how do you prioritize? What are the best use cases for this technology? What are the right markets for us to pursue?

    Scott: Well, what you say no to is just as important as what you say yes to, in a business.

    Shimel: Yep.

    Scott: And, right now, we are very, very focused in the Americas, as well as in western Europe, right? So we’ve been able to say no to Asia for another year or two. Our value proposition is really about the biggest of the big—the telcos, the banks, the large retailers, people who’re looking at real scale that they’re trying to solve problems with. But the fact is, is where I see the market going over the next three years is more about consolidating the DevOps, the ITOps, the Network Ops, the SecOps. All of those are coming into one. And I’ve heard a couple of people start using the term “fusion center,” which is really helping them drive the business and stop pointing fingers and reducing that mean time to response.

    Well, every security product starts with the logs, right? And every IT monitoring product starts with the logs and the data. And the advantage of us being able to load in all of that relational data, which is about 20% of our data that’s in our systems, across our customer base, is things that is coming out of Xdata or data warehouses or relational or customer application type or weather data or threat hunting data that comes in, and being able to combine all of that. So, for us, it’s the larger customers or companies that have data at scales that we’d see in the global 1,000 and really looking at the people who understand that what Google, and Apple, and Facebook, and Netflix have done is taken a data-first approach to their business.

    And, when we look at customers that are forward seeing—that see forward into the market, they realize there are a whole new set of competitors and being able to ask what their real competitive advantage is in their business and the data that it takes for them to be able to do that, whether that’s a telco to figure out how to launch a 5G network because they have all the 4G, 3G and 2G data, whether it’s in their customer retention rates or net retention of “Can they get more dollars out of their customers each year?” to whether it be their manufacturing process and figuring out that arbitrage in the supply chain.

    So I think we are—because of the ease of use of our product, because it’s so easy to get data in, because we keep data in raw format. So you can have one set of data that’s very easy to get in and use it 100 times, not having to index it or reduce that data because it just makes it more efficient.

    But I see us, in the next couple of years, really focused on the security side of the equation. Looking at people who’re doing the transformation and the weight that it’s gonna take on their DevOps and CloudOps teams and being able to work closely with the network and the DevOps teams to be able to grow their business because the cost of becoming so great on having so many different tools, at different layers of your network, I say that the best and the brightest companies will take control of their own data, consolidate those tools so they have a better competitive advantage.

    Shimel: Good. Good. Good. Walter, I’m gonna throw a term at you and I want your opinion on it, “AIOps.” What do you—this is a term that we, here at devopsy.kinsta.cloud anyway, we see it rising with a bullet. And, like you—to me, there’s AIOps with a capital “A” and a capital “I” and it’s like Robby the Robot stuff that’s gonna make us live like The Jetsons, finally. And then there’s AI with a small “a” and a small “i,” like the real, functional, kinda down in the dirt–maybe it’s not as sexy, but it provides real change for us. AIOps is—man, everyone—what’s your story on AIOps? It reminds me of when cloud came out in 2005. What do you think?

    Scott: So let’s be very clear. AIOps is a term that’s been coined by Gartner. Out of the last 100 conversations I’ve had with CIOs, I’ve only heard it referenced once or twice. I think it’s more the vendors that are getting on it. And, when you look at it, log management is a big part of AIOps. You know, user behavior analytics is part of AIOps. The anomaly detection is part of AIOps, right? There’s seven or eight different components to that. I don’t think any vendor right now, in the world, really has defined themselves as an AIOps vendor. I think we’ve probably got three years, through help of M&A and deciding what that is, ’cause incident response or the service management is also there, so I think it’s gotta evolve.

    But, of the seven or eight different components, the monitoring, the dealing with events, the user behavior, the anomaly detection and the log management is there, which is where we focus. And I think the enterprise log management, I thought, five, six years ago, was solved, when we solved who all those vendors are, but now, with the complexity of the cloud, it isn’t solved at the scale customers need to do with it. We had solved router management; now, with SD-WAN, it’s a whole new market.

    Shimel: Yeah. You’re right. I mean, and it’s funny that things that–it’s kinda like the measles, right? We thought we cured measles and here we are back with measles. Walter, we’re both CEOs of companies and I don’t know how many people listening out there are CEOs—many of our listeners are kinda getting their fingernails dirty, practitioners as well as managers at large enterprise—what keeps you up at night about Devo? What do you worry most about, Walter?

    Scott: Well, I think part is we have to manage tens or hundreds of petabytes of data for our customers, right? And I think we have a very large team internally for security, to make sure that the data that we have is protected and we work very closely with Amazon and Google and Azure to make sure that we address that. And that’s something that you’re never gonna sleep well ’cause you have to be passionate about making sure that you protect your customers.

    We were on the phone literally an hour ago with a manufacturer that makes toys for children, and that was the question that’s keeping him up–he says, “I have 10-year-old boys and girls logging into my website. I have to protect that data, making sure in the things that we do with being able to obfuscate data at the row, in the column level, without having to duplicate it and our algorithms that are in there.”

    But my job, as a CEO, is to help people build careers. I’ve been very fortunate in my career and have had some very large exits in part of the business, but it’s making sure that the team around us continues to make sure that we handle all of the things, in dealing with our business and our main offices in Madrid. We also have a big office in Cambridge, Massachusetts, because I love the guys and women coming out of MIT, where we do a lot of hiring. But it’s mostly about the team and then making sure that we continue to have what I call a industry-abnormal or an anomaly in terms of net retention. And our goal is to help our customers continue to grow because it benefits us and wanting to make sure that we make it easier for them to be able to get more value because our mission is about making IT so that they can provide data to the business that is actionable.

    And my belief is is the CISO or the CIO that has control of that data, if they push it out to the business and say, “Hey, did you know that, on this particular case or when this happens, we’re losing customers or we’re gaining customers?” That data is 80% more likely to be leveraged than it is if the customer asks the business. So—

    Shimel: Nice.

    Scott:–I think it’s empowering the IT to accelerate value to the business and most of my nights are, well, we see this new customer is bringing in this type of a data, bringing in our data science teams to help think about use cases that we can drive more value from.

    Shimel: Excellent. So, Walter, as I mentioned to you before we started recording today, the good news and bad news about this audio format is it goes so darn quickly. And we blew past our time. So what we’ll do, though, is we’ll—secret’s out: you live down in Florida, too. We’re gonna need to get you down here and into our studio and let’s do a whole video on this. We’ll continue our conversation in a video format. But, for now, we’re gonna call a—have to pull the plug on this one. Walter Scott, CEO of Devo, thanks for being our guest today on DevOps Chat.

    Scott: Great, Alan. Thanks.

    Shimel: All right. Hey, this is Alan Shimel, devopsy.kinsta.cloud, and you’ve just listened to another DevOps Chat.

    — Alan Shimel

  • Technology: Isolation in the Age of Automation

    Technology: Isolation in the Age of Automation

    Dependence on technology makes us more independent overall, but it also makes us more isolated. So, what happens if things come to the point where we depend on machines more than we depend on other human beings?

    Earlier this month I attended a presentation about machine learning on mobile phones at Google IO 2019. Right now Google is spending a lot of time and money developing machine learning models and capabilities that operate solely within the mobile device. No connection to the internet is required. It’s a big deal.

    During the presentation, the company showed a video that featured an illiterate woman in India who uses her phone’s text recognition features to read to her. All she needs to do is point the phone’s camera at a recipe or newspaper and the phone converts the text to speech.

    I found the story inspiring as a technologist and as a human being. But it’s not all peaches and cream. At the end of the video, the woman said something that gave me pause. When talking about using the technology to book a train ticket, she said, “I can book it myself. I don’t need anyone’s help.”

    There’s no doubt the phone liberated her, very much in the same way that the light bulb liberated millions of other people in times past, bringing human activity forward beyond the setting sun. Dependence on a particular technology can create greater independence overall. However, I do wonder what happens when things evolve to the point where we depend on machines more than we depend on others, when we no longer need “anyone’s help.”

    To my thinking, human interdependence is the thread that binds the fabric of a society together. The thousands of little interdependencies that make up the human experience have beneficial side effects. To get along, we learn the basic rules of civility. We learn to say “please” and “thank you.” We learn how to wait in line at the DMV. We pick up after our dog. We open the door for others. It’s all part of the implicit contract that makes living with other people possible.

    But, these days it seems we depend more on machines than the kindness of strangers. The fact is, you don’t have to be nice to a machine, even for a self-serving purpose. You can yell at your cell phone. You stick your tongue out at the ATM as it accepts your deposit. Whether you’re nice to it or not doesn’t matter. Your deposit is not going to be processed any faster or slower. Your cell phone isn’t going to disconnect you because you were rude to it.

    Take away these simple rules of civility and strange things happen. It’s interesting that road rage is common, yet sidewalk rage is rare. Maybe wrapping ourselves up in a piece of silicon-powered transportation hardware predisposes us to a hostile attitude when interacting with others in the same situation.

    As I stated earlier, depending on technology makes us more independent in the big picture. But, it also makes us more isolated. Whereas in the past, talking about the weather with a stranger might be a prelude to another conversation that’s more meaningful, today the stranger next to me usually has his ears plugged up listening to his own interests. And, if I really want to know about the weather, I can ask Alexa.

    Modern technology gives me help on demand. I no longer ask a stranger for directions to the train station, I use my phone. It seems as if I need the help of others less during those small, random moments of vulnerability. A machine is but a voice prompt away. And, thus, the thread that binds me to the world at large gets a little bit looser.

    Now don’t get me wrong—isolation is nothing new. Five hundred years ago most people never traveled far from where they were born except on very special occasions. Village life is a case study in group isolation. But, while villages were isolated from the outside world, inside the village it was a different story. The social fabric was dense; the threads were tight. Being shunned by the group was a matter of life and death.

    Today we’re all over the place, literally. But we move in our own domains in which we end up wishing happy birthdays to people we’ve never met but who somehow have meaning to us by way of LinkedIn, Facebook and the plethora of other social platforms brought to life and sustained by something other than human labor.

    Is this such a bad thing? Dunno. I like it that my phone is now my personal stenographer. It helps me think better. But, I’m not sure that Google Assistant will ever be a good best friend. What’s even stranger is that a time might come when I don’t even need one. It will be just me and my mobile device. Then I won’t need anyone’s help, ever.

    — Bob Reselman

  • Nutanix Eases Development of IoT Apps

    Nutanix Eases Development of IoT Apps

    Nutanix, a vendor of hyperconverged infrastructure solutions, has released a new developer environment that allows developers to try and test IoT and AI apps in the cloud. Starting May 7, developers can sign up for a free 10-day trial of this experience.

    The developer environment is built under Nutanix’s Xi IoT Platform and gives developers access to a rapid development environment for creating internet of things and artificial intelligence apps in the cloud. The Rapid Development Environment for IoT includes features such as Xi IoT Cloud Instance, Xi IoT Sensor App and Xi IoT App Library and supports the ability to easily deploy IoT software in the cloud or at the edge, according to the company. The environment has been tuned specifically for IoT and AI, eliminating many of the configuration chores developers had to perform using other cloud-based development environments, and enables single-click deployment via a management console.

    For developers targeting production-edge environments, the new environment brings forth the ability to develop for the edge or for the cloud, bringing application portability into the development and testing process. By being able to use a common set of APIs, cloud and edge development becomes relatively seamless and deployment friction is vastly reduced, according to the company.

    Beyond ensuring compatibility across platform services and APIs, the environment enables development and production environments to use the same components, so recoding is not needed as part of the transition from dev to ops. That helps to simplify lifecycle management as well as fit agile practices better into the dev process.

    Nutanix also offers an extensive library of applications that can be used to ease development chores. The Xi IoT App Library offers  individual components (message bus services, time series databases, etc.) and full solutions that developers can use to address various business problems.

    Nutanix is framing the new delivery mechanism as a new experience to get more developers—from novice to advanced—creating AI and IoT scenarios.

    — Frank Ohlhorst

  • 3 Steps to Building Data Science Models

    3 Steps to Building Data Science Models

    Have you ever considered supplementing your development skills with some data science? Expand your range and follow this quick tutorial on how to build your own predictive models using open datasets, Jupyter notebooks and Sequential models.

    Pick Your Data Game

    Your first step in learning data science is to find a dataset that fits your interests. A robust platform where you can find inspiration is Kaggle, a web-based data science environment for machine learning enthusiasts. It offers, not only open datasets, but also competitions and even free micro-courses.

    Maybe you want to change the world and help with scientific research. Maybe you’d like to jump in to win $65,000 in a sponsored challenge. Or maybe you’d like to start with some machine learning techniques. There is plenty of data to search through and there always are a couple of open contests to enter.

    For this tutorial we are going to use a dataset from a competition called “Histopathologic Cancer Detection.” It is very important to understand not only the data we are going to work with, but also the main purpose of this challenge. Looking at our example, we read the following: Identify metastatic tissue in histopathologic scans of lymph node sections.

    In layman’s terms, the goal is to determine/recognize/discover if there is a sign of tumor tissue in a given image based on a dataset of lymph node scans. This is enough information to know that our model will have to answer this question in a binary manner: if there is a metastatic tissue in a given image, the answer is yes; if there is not, the answer is no.

    Choose a Game Plan

    There are two main types of machine learning, supervised and unsupervised.

    In our case, we have a training set and a ground truth–train labels. In other words, we have prior knowledge of what the output values for our training images should be. Therefore, if we have a ground truth, then we know we are dealing with supervised machine learning.

    The goal here is to learn a function based on example input-output pairs. Having this function, the task is to map a new given input to a predicted output. Sound reasonable? Let’s jump into the competition sources and see if we have what we need.

    Train set: This set contains the data you will train your model on. It’s the set of plain images for which we already know the correct answer.

    Train labels: As you may have already guessed, this file contains the correct answers to each image in a train set.

    Test set: Set of images for which we do not know the answer.

    Our goal is to create a new file (similar to train labels) containing predicted answers for images from the test set.

    Game On!

    For this opportunity, I decided to explore Google Colaboratory. It is a free Jupyter notebook environment with Python 3 that requires no setup and runs entirely in the cloud. If you haven’t used Jupyter before, now is a great chance to get started. With Jupyter notebooks, you can write and execute code, save and share it, and even access powerful computing resources, all from your browser.

    We start our game by importing the files we need to Google Colaboratory. Here’s how: Remember to copy the imports as we will need them later.

    We will use the Kaggle official API (you can read more about the setup and obtaining needed keys here) to easily retrieve large datasets directly to a server environment.

    Now when you execute ls /content/, you should see:

    We are going to read dataset into the pandas data frame.

    The output should be as shown below:

    Having all data in place we are ready to construct our first model.

    The Sequential model is a linear stack of layers. The model needs to know what input shape it should expect. For this reason, the first layer (and only first) needs to receive information about its input shape. You can read more about the construction of Keras Sequential model here.

    We need to declare some helper functions to transform the data to fit our needs (map train set with train labels).

    Execute the code below and put on your favorite TV show … this will take some time to compile! Remember to change hardware accelerator to GPU. It’s free and can speed up model building time up to five times. Go to Runtime, then change runtime type and change None to GPU.

    Data Science Endgame

    Before we generate predictions, we should probably preserve our model (better be safe than sorry).

    The last step is to get the answers for the images in our test set. When all is computed we can see predictions for the first couple of images.

    To save final results and download them run this:

    It’s time to upload the files to Kaggle and wait to see our final percentage score.

    Stepping Up Your Data Science Game

    There’s nothing better than a little healthy competition. Since we are taking part in a real contest, we have to consider that there will be some winners and some losers. There’s no doubt that we want to be in the first group, but the model that we have trained so far does not guarantee this. We should improve the predictive power as much as possible (although watch out for overfitting).

    There are many techniques to do that, including adding more data, algorithm tuning or ensemble methods. It’s easy to list them, but not so easy to pick the ones that actually will improve our score.

    There are endless possibilities to increase the predictive power of this or any model. Pick a new dataset and try for yourself to expand your data science skills.

    — Magdalena Baracz

  • What Can Automation Learn From DevOps?

    What Can Automation Learn From DevOps?

    A recent Capgemini survey revealed that while enterprise-scale automation is still in its infancy, IT automation projects are moving along. IT is starting to view automation less tactically and more strategically.

    IT leads automation implementation (respondents were asked to select all that apply: “In which of the following functions has your organization implemented automation initiatives?”). Capgemini Research Institute, Automation Use Case Survey; July 2018, N=705 organizations that are experimenting with or implementing automation initiatives.

    The Capgemini survey also showed that IT automation can be responsible for several quick wins, including self-healing, event correlation, diagnostics, application releases, cybersecurity monitoring and storage and server management tasks. These projects not only lead to massive IT cost savings but, more importantly, to an increase in reliability and responsiveness to customer demands and business services. That would indicate that while automation is a great solution for manual work, it’s also a part of a high-level, strategic IT plan to innovate the business.

    As DevOps practices such as agile methodology, continuous deployment and optimization start to take hold within the modern enterprise, it stands to question: Can automation be agile as well? This is the promise of artificial intelligence (AI) for IT operations, or AIOps, but if that’s not a possibility for your IT organization today, it’s important to make sure that your automation practices are continuously optimized to fit the task. Setting and forgetting was a practice of the server era, and in a world of on-demand infrastructure, automation ought to be continuously optimized and evaluated for maximum benefit.

    The New Expectations of Automation

    IT automation projects can have serious ramifications if anything goes wrong because when the machines execute a policy, they do it in a big way. This is perhaps the chief argument as to why it’s critical that progressive steps are used to define and evaluate both the process being automated and the automation itself—they mitigate the seriousness of any issue that can arise. This is why it’s important to consider the following:

    1. Is this a good process, and is it worth automating?
    2. How often does this process happen?
    3. When it happens, how much time does it take?
    4. Is there a human element that can’t be replaced by automation?

    Let’s break down these steps and see how it can provide the basis for an iterative approach to automation:

    Is this a Good Process?

    This may seem like a rudimentary question, but in fact, processes and policies are often set and forgotten—even as things change dramatically. Proper continuous optimization or agile automation development will force an IT team to revisit existing policies and identify if it’s still right for the business service goals.

    Some processes are delicate and automation may threaten their integrity, whereas others are high-level and automation neglects the routine tasks that underlie the eventual results. A good automation engineer understands what tasks are best for automation and sets policies accordingly.

    How Often Does this Process Happen?

    Patching, updating, load balancing or orchestration can follow an on-demand or time-series schedule. As workloads become more ephemeral moving to serverless cloud-native infrastructure, this process’s schedules will change as well. An automation schedule ought to be continuously adapted to the workload need, customer demand and infrastructure form. As the business continues the march toward digital transformation, the nature and schedule of particular work may become more dynamic.

    When It Happens, How Much Time It Takes?

    This also depends on the underlying infrastructure. Some legacy systems require updates that may take hours and some orchestration of workloads will be continuous. Automation must be tested to be efficient and effective on the schedule and frequency of the manual task.

    Is There a Human Element that’s Irreplaceable?

    As much as you may want to, it’s difficult for automation to shift left (to more experienced tasks and teams) without the help of AI or machine learning. Many times, there is a human element involved in deriving insights, creating new workflows, program management or architecture that take place. When building an iterative automation practice, make sure you identify where human interaction must occur to evaluate and optimize. In our lifetime, technology has advanced at lightning speed, with robots now completing jobs that were once held by people. However, there are times when a machine just cannot deliver the same quality a human can.

    Automation for All

    Automation is perhaps one of the most defining signatures of the future of IT operations management. It relieves teams of routine work and helps improve overall efficiency, all while driving quick wins that turn an IT team into heroes. But don’t let automation be the end goal. Instead, consider it a tool that can drive action from data. And until AI is an everyday option, it’s inherent on the IT professional to continuously optimize the data that drive that action.

    — Bhanu Singh

  • XebiaLabs Injects Predictive Analytics Into DevOps Platform

    XebiaLabs Injects Predictive Analytics Into DevOps Platform

    XebiaLabs is moving to take the guesswork out of DevOps following the launch of a set of tools that leverage machine learning algorithms to predict when and how the release of an application is most likely to go awry.

    The Risk Prediction Module being added to the XebiaLabs DevOps platform comes on the heels of XebiaLabs’ move to make its XL JetPack software for automating application deployments available on the Amazon Web Services (AWS) Marketplace. XL JetPack is an automation framework based on YAML files and a declarative programming framework that enable DevOps teams to push applications into production on a cloud platform in as little as 15 minutes.

    XebiaLabs CEO Derek Langone said that while DevOps advancements have been made in terms of achieving continuous integration, the goal of achieving continuous delivery has proven to be more elusive. As automation frameworks such as XL JetPack become more accessible, however, IT operations teams will be better able to keep pace with the rate at which organizations want to be to deploy and update applications, he said.

    It’s becoming increasingly apparent that a combination of automation frameworks and machine learning algorithms will enable DevOps teams to make a significant leap forward in terms of application release lifecycle management. Because of all the nuances of the different platforms that IT organizations are trying to leverage, it’s become a major challenge to roll out an application manually. As the number of applications that need to be released at roughly the same time increases, managing the release cycle process is moving beyond the capabilities of an IT team made up of mere mortals.

    The Risk Prediction Model from XebiaLabs promises to increase the odds of success by applying predictive analytics infused with machine learning and other proprietary XebiaLabs algorithms to identify DevOps bottlenecks before they are encountered. Capabilities provided by that module include alerts that warn the team when a release is likely to be delayed or to fail before the release pipeline starts running; a Risk Forecast view that summarizes predicted delays and failures for every task in the release process; an overview of statistics for similar releases to provide historical context; and forensics tools that uncover flaky automated tests, problematic build setups, long-running deployments and time-consuming manual tasks. That Risk Prediction Module is meant to complement the risk scoring capability that XebiaLabs already includes in its platform.

    Arguably, the biggest barrier to DevOps adoption today continues to be the level of expertise required to succeed. But as the combination of predictive analytics infused with machine learning algorithms and algorithms that can discern instantly how changes to the IT environment will impact applications become more common, the level of expertise required to master DevOps processes should decline. As that transition occurs, the number of organizations embracing DevOps more broadly should increase.

    In the meantime, DevOps teams would be well-advised to consider to what degree they might next want to rely on predictive analytics to prescriptively automate processes with eye toward eliminating the need for as much human intervention as possible.

    — Mike Vizard

  • Developing Simple and Stable Machine Learning Models

    Developing Simple and Stable Machine Learning Models

    A current challenge and debate in artificial intelligence is building simple and stable machine learning models capable of identifying patterns and even objects. Is it possible to know which models will work best or to simply see the data? It is possible, and the question is figuring out how to get there.

    Simple models are popular because they are easy to understand, trustworthy and tend to be more resilient to change than complex models. Simple models also have the benefit of overcoming transfer learning pains. While everyone wants simple models, they also want accurate models. In practice, there are always transfer learning issues when applying learning from the past, into future practice.

    An issue arises when trying to fit a model from a known family to the available data and hoping it will generalize effectively. It could be any proposed family: linear models, trees or trained deep neural networks.

    There remains a fundamental problem with any overly complicated model—linear, tree or neural network—in the risk of “overfitting” the data resulting in poor generalization and results. At the same time, an overly simple model risks missing the nuances of the data and sacrificing accuracy. There are techniques to handle this problem.

    First is to reduce the size of the set of models. This works well with the PAC model. For example, reduce the number of parameters, reduce the depth of the tree and reduce the nodes in the network. Reducing the size of the set means it’s easier to understand, and with fewer possibilities built into the model, it’s less likely to overfit the data.

    Next, add L1 or L2 regularization. Models with small weight, or many zero weights, are preferable and will help the model generalize. The question is, why? Why do we use trees or neural networks rather than another hypothesis family? Why do we want models with small weights? Why do some network topologies work well?

    These questions bring in the actual world we live in and the types of problems that likely need solving. There is no learning when any function is allowed, and the model family isn’t limited. It is well-known that when the family of allowed hypotheses is too large, the model overfits the data ad-absurdum and becomes useless. Thus, machine learning almost always optimizes along a well-understood family of hypotheses while human feature engineering and representation construction captures our prior knowledge about the world. The question is how to exploit such knowledge when automating machine learning. Arbitrary limiting is better than not limiting at all, but we can do better.

    The world has patterns repeating themselves across many domains. Different applications will use similar building blocks to analyze rocket movement or medical sensors, or to solve a predictive maintenance task. But we are all experts in the world we live in. Even studying a domain where we don’t have specific expertise, we still recognize some patterns and rules.

    For example, take a look at this ring-tailed lemur. Even if you’ve never seen one before, you are now likely to identify this type of primate if you come across one. My 5-year-old son was shown this picture and he could visit the zoo and identify a ring-tailed lemur. It only took one photo.

    Compare that result to modern neural network research which typically requires many examples—often into the millions—when training to identify objects, and still they suffer from embarrassing mistakes when new data differs from the data the network was trained on.

    How did my 5-year-old son learn to identify ring-tailed lemurs from a single picture? He’s an expert. Not a lemur or primate expert, but an expert in the world we live in. He understands identifying an object from its background. What a tail, eyes, ears, legs and fur look like. He knows how to imagine a 3D object from different directions. At 5, he already brought a ton of knowledge with him when learning to identify lemurs.

    This is the essence of learning. We understand the notion of simplicity is not in number of nodes in the network or number of lines of code, but that it relates to simplicity when explaining human language.

    This process can be formalized by codifying human knowledge to solve tough problems. We can formalize the many patterns seen before in various domains and in the various tools data scientists use when constructing features. Then we can look for patterns that make sense and are similar to what we’ve seen before.

    We can also look for more novel patterns represented within the complex combinations of things we’ve seen before. As an industry, we need to do exactly this: Codify human knowledge in curated code libraries, and in structured knowledge and facts, use them to search for patterns in data and build highly accurate resilient models. This will bring simplicity and stability to machine learning models without sacrificing accuracy by leveraging humanity’s existing knowledge.

    — Meir Maor