How Machine Learning and Data Lead to Predictive Intelligent Automation
Featuring Ashok Reddy and Ali Siddiqui
Staying ahead of industry trends and innovations in the market is key to making sure we’re developing products that address our customers’ most pressing needs. To give you the latest insights from our leadership, our GM’s will be sharing their thoughts on what they’re seeing and how we can stay ahead.
Focusing on Machine Learning & Data
Data is the new currency. Access to massive data (both structured and unstructured) and immense computing power for analysis has increased the opportunity for machine learning at a pace that was never possible before. Now, technology like self-driving cars, robotic assistants, and the internet of things are part of everyone’s everyday life. And faster machine learning, analytics and automation are setting the stage for even more advancements.
Ashok Reddy, General Manager of the Mainframe BU, and Ali Siddiqui, SVP and General Manager of the Agile Operations BU, share their insights on machine learning and automation.
Q: What are the most interesting trends you are following and excited about when it comes to machine learning?
Ashok: Developers used to code software based on a business outcome with existing knowledge, processes and rules, but machine learning has turned this upside down. Machine Learning is defined as improving with experience at some task. Now, machines can learn from historical data with series of examples, so in theory, you can use a modern software factory that automatically develops the next application to automate the tasks. For companies going through a digital transformation, it’s all about applying machine learning to create new business models to disrupt their industry. One application of this is taking and applying data to make faster, better decisions. Another is improving customer engagement through natural language processing, chat bots, etc. to personalize and improve the user experience. The third is through intelligent automation — like self-driving cars, you can have a self-driving cloud or mainframe based datacenter. This way, before performance declines or storage gets full, machine learning will help remediate the problem. AI and Machine Learning are changing the game — it is a once in a lifetime general purpose technology, just like electricity. I think it will transform every industry — and in my opinion, it’s the most interesting thing in our lifetime.
Ali: I’m excited that machine learning is making IT fun again! We now have the ability to do root cause analysis to predict and detect issues quickly and solve them before customers experience gets impacted. In today’s app economy, your application is your brand and your marketing platform — so the ability to make predictions and deliver the right services to customers before they can even decide what they need is critical.
Q: As companies become more automated, what opportunities do you see opening up?
Ashok: If machines are able to automate and handle the mundane tasks, then people can focus on higher value things. For example, radiologists won’t have to review X-rays because they’d be able to use Machine Learning to compare and identify anomalies in images and get the results — and spend their time focusing on coming up with the right diagnosis and discussing options with their patients — and reducing the need to get a second opinion.
Another thing I’ve noticed today is that automation is generic for everyone — it’s not personalized, so if you need something unique for a user, it’s not an option. We can help enterprises create a customized experience using data, previous usage, preferences, etc. — so it’s personalizing using intelligent automation. We can take it to the next level by applying the agile mindset and the notion of scrum, so you can learn from everything. We have so much data and we’re not using it to learn. Automation and machine learning will provide a learning based approach to help in many different outcomes.
Ali: Companies across many industries are focusing on and asking for intelligent automation tools since they give businesses the ability to use their resources better, know their environments better and make better decisions. For example, in the health care industry, compliance is a top priority — so when security vulnerabilities happen, monitoring tools detect whether a patch is missing and apply it across the board. This really is intelligent automation at work! Intelligent automation is also key to monitoring cloud workloads for retail customers. For example, Macy’s might run a marketing campaign over the weekend, rolling out a new clothing line — causing traffic to surge suddenly because the campaign really resonated and people started buying online. In today’s world, a company like Macy’s can’t operate without intelligent automation to make predictions and respond immediately to these kind of issues.
Our acquisition of Automic plays a key role in this — they are really blazing the trail of next generation automation tools. The combination of DevOps and monitoring capabilities is really unique in the industry.
Q: What do you see as the next frontier for machine learning and automation?
Ashok: I didn’t think we’d make as much progress by now on voice and facial recognition, but we have –it’s all about having more data and examples to train and refine the algorithms. For me, this is an area where machine learning has surpassed human performance — and it has spread quickly. High speed machines with high recognition cameras can look at different things — people, puppies, etc. — and identify who or what it is. It’s not perfect, but it is better than humans.
Also, today, most AI systems are based on supervised learning — so they have to have many examples with classified and labeled data. In the future though, I think we’ll advance to a point where machine learning will be based on unsupervised learning, so we can just feed unclassified and unlabeled data and see what patterns they can uncover.
Ali: Driver-less cars is definitely one of them. I think this would fundamentally change and reconfigure our cities. In the future, cites might even exist without roads — and without roads, our buildings, offices and communities could be completely different. I am excited about that! This would help protect our environment too. All of this could happen sooner than we think — with the right support in the private and public sector who believe in the benefits of science and technology and with leaders who understand that we all share the same planet, science and technology can be applied to solve really important global problems.
Q: Are there any new developments you’d like to share from a CA perspective?
Ashok: CA is in the perfect spot to apply machine learning — it comes down to data and we have access to operational data for our customers — the world’s most critical companies. We can take this data and do predictive analytics and apply machine learning. For example, with our Mainframe Operational Intelligence (MOI) offering, we can help large banks predict when an ATM outage will happen before it happens. Customers want this on the mainframe, distributed and the cloud — pulling data on CA and non-CA products — and we just launched this over the past month. Also, IBM just introduced next generation mainframe (z14), which is a platform for machine learning built into it — using Apache Spark, it supports Google Tensorflow and Watson. Clients can apply machine learning on top of the mainframe data (70% of enterprise data is in Mainframes) and do real-time machine learning on the most current data. This is important because, for example, if your family shares a credit card and they’re in different locations, it can check for fraud by doing an analysis of regular patterns in real time and within milliseconds. Lastly, I’d like to encourage everyone to get out there and learn about machine learning. I’m personally getting my Master’s degree from Georgia Tech in AI and Machine Learning. You can learn about it online and CA has access to lot of great courses through Tech Aid.
Ali: We are really breaking the barriers between ideas and outcomes for our customers because we bring a full portfolio to solve their most pressing problems. Currently, my team is working with Ashok and his team on MOI, as he explained — we’re focusing on the distributed side (from ATM to mobile phone you can see performance issues end-to-end). We’re also working with Rahim Bhatia and his team on Precision API monitoring, which is exciting since the next-generation architecture will be driven by microservices and APIs. Rahim is providing the foundation for that next-generation architecture — the microservices gateway and APIM — and our team will provide the Lifecycle Management. Together, we’ll be able to provide real solutions for business ideas. We’re also working with Jeff Scheaffer and his team to provide close integration with testing and monitoring solutions sets. So, lots of collaboration across the BU’s.