![](https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1/files/2020/05/Sriram-Raghavan.png)
![](https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1/files/2020/05/Sriram-Raghavan.png)
In the 1894 mystery “Adventure of the Crooked Man” by Arthur Conan Doyle, Sherlock Holmes analyzes a crime scene by explaining to his trusty sidekick: “You know my methods, Watson.” More than a century later, it appears that Watson has learned a whole lot more.
This month’s first IBM Think event under the new leadership of Chief Executive Officer Arvind Krishna supported his mantra upon assuming the helm: Artificial intelligence and the hybrid cloud will drive IBM’s future.
The company’s major announcement for Watson AIOps showcased the technology’s ability to detect anomalies in IT infrastructure using Red Hat OpenShift run across any cloud. On the COVID-19 front, Watson Assistant can now be accessed on a medical research site in India to help with the virus testing process, and Watson Tone Analyzer is being deployed to help small businesses forecast risk during the crisis.
In April, Watson Machine Learning was used to migrate data and applications from London to Frankfurt in advance of the U.K.’s exit from the European Union. And the IBM Think conference itself was powered last week by Watson Media.
As the use cases for Watson AI become clearer, so too is IBM’s strategic approach for the role it expects the technology to play in its business and for customers as well.
“The story around having a solid enterprise information architecture as the base on which to drive AI is starting to happen,” said Sriram Raghavan (pictured), vice president of IBM Research AI at IBM. “Advancing, trusting and scaling are the three big mantras around which we think of AI.”
Raghavan spoke with Dave Vellante, host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, during the IBM Think Digital Event Experience. They discussed the use of AI to improve operational efficiency, the challenge of monitoring data models to ensure proper performance, building trust and explainability into intelligence tools, and advances in natural language processing. (* Disclosure below.)
This week, theCUBE features Sriram Raghavan as its Guest of the Week.
IBM’s focus on advancing and scaling AI can be seen in the release of Watson AIOps this month. AIOps is a significant step because its IT focus is likely to appeal to an important constituency — chief information officers — who currently play a central role in the overall success of the enterprise.
IBM drove this point home with a video example in which a network problem is detected and an army of engineers and system administrators chew up hours of time and tens of thousands of dollars to find and fix the issue. With Watson AIOps, according to IBM, one person handles the problem in 14 minutes.
“It isn’t sufficient that I tell you that AI ‘can’ do this,” Raghavan said. “How do I make AI do this so you get the right ROI, the investment relative to the return makes sense?”
For that investment to make sense, CIOs need confidence that AI is making progress and it can resolve the thorniest operational issues in their organizations. Building predictive models for back-end analysis? Works great. Deploying those models into production? Not so great.
IBM has clearly been thinking about this problem for some time. In an interview with theCUBE over a year ago, Inderpal Bhandari, global chief data officer at IBM, described the challenge of applying AI and dealing with data in a hybrid and multicloud world.
“The average enterprise makes use of nine different clouds,” Bhandari said. “To manage across all those environments is very tough. So, from a data standpoint, you have that same complexity extending into the data space. That’s an example of the complexities that you have to solve.”
While the release of Watson AIOps received a great deal of attention during IBM Think, additional enhancements were designed to move IBM closer to a central role in resolving the challenges described by Bhandari. There were upgrades to Cloud Pak for Data, new tools for application modernization, and a Workflow Orchestrator that applies AI-based reasoning tailored for specific apps.
IBM is focused on leveraging AI as DevOps for data. After a data model is put into an application, how can an enterprise monitor and trust its behavior in real time?
This is where Watson OpenScale comes into focus. Launched in 2018, OpenScale helps enterprises manage production AI wherever the data lives and alerts developers to bias in machine-learning models.
IBM’s own surveys have shown that trust is the bedrock for AI’s successful adoption. Being able to trust AI’s output was rated highest by 78% of respondents in a recent sampling, and 83% wanted an explanation for how AI arrived at a decision.
With OpenScale, IBM is seeking to solve the “black-box” problem by providing insight into AI health and methods for understanding the reasoning behind risk models.
“AI needs to be less data and resource hungry and more trustworthy and explainable,” Raghavan said. “OpenScale came out of research work in trusted AI and is now a centerpiece of Watson.”
In addition to allowing AI to explain itself, IBM has been working on letting AI also explain the problems of the world. Project Debater represents a step forward for IBM’s research in NLP.
The AI project faced off against the world’s leading debate champions in 2018 and 2019. While IBM’s machine lost to its human adversaries, it held its own in being able to digest and respond to nuanced information.
More significantly, Project Debater served as a key development platform for Watson Assistant and Watson Discovery, IBM’s AI conversational and search solutions.
“Advances in NLP are all making their way into Assistant and Discovery,” Raghavan noted. “They are born out of research and innovation in solving a grand problem like building the Debater project.”
Raghavan has spent his entire corporate career at IBM, starting out as a staff member in the company’s Almaden Research Center in 2004 after completing his computer science degree at Stanford University. He worked his way through a variety of roles, which included leading research efforts for IBM’s operations in India.
Now Raghavan finds himself in charge of the company’s global research initiatives that includes developing and executing its crucial AI strategy.
“It’s a dream job and a lot of fun,” Raghavan said. “I’ve always been fascinated. You can’t find a technology person who hasn’t dreamt of building an intelligent machine.”
Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the IBM Think Digital Event Experience. (* Disclosure: TheCUBE is a paid media partner for the IBM Think Digital Event Experience. Neither IBM, the sponsor for theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)
THANK YOU