Clients Use Cognitive Computing to Solve Business Challenges
Illustration by Kai and Sunny
Artificial intelligence (AI) has been the stuff of science fiction for decades. But now it’s reality, centered on computer systems that aggregate and analyze massive amounts of data, using what they’ve “learned” to provide actionable information.
“AI is a concept that has locked in popular consciousness—this idea that we create machine intelligence,” says Scott Soutter, senior offering manager, high-performance computing (HPC) and data analytics, IBM. “We’re teaching computer systems to develop high levels of insight. And they’re looking more and more like living entities that can respond to a series of tasks.”
The S822LC for HPC server is designed to manage big data and deliver the performance and throughput of POWER8* technology by offering insights from that data faster, and at a low cost.
“We’ve made this available to all of our clients because we believe it’s one of the most exciting technologies shaping computing today.”
—Scott Soutter, senior offering manager, high-performance computing and data analytics, IBM
In November 2016, IBM launched PowerAI, a platform for the S822LC for HPC, designed to help organizations use machine intelligence to develop practical solutions to real-world enterprise and organizational challenges.
Businesses can use cognitive computing to turn zettabytes of data into meaningful information, helping to solve critical issues such as online language translation between the business and international customers. With PowerAI, S822LC clients can put AI to work today.
AI at Work
Several terms in the AI world overlap. Cognitive computing, for example, is often used interchangeably with AI.
Cognitive computing embraces machine learning—the capability of computer systems to observe and learn through artificial neural networks (ANNs) modeled after the multiple layers of neurons in the human brain. Deep learning focuses on developing ANNs to analyze massive quantities of data. It allows computer systems to learn over time so they can make quicker, more accurate data observations in the future, and create their own algorithms, rather than relying solely on human programmers.
“Deep learning and machine learning are expressions of the same sort of technical, problem-solving tool,” Soutter explains. “Deep learning is an evolution of traditional machine learning, where the programmers create the algorithm. We use the computer to start to develop an algorithm. Then we allow it to train itself with data.”
Organizations in many industries detect patterns in data sets, but until now, that’s typically been done through traditional analytics—a process that can be tedious and time-consuming. “Deep learning provides another way to get answers,” Soutter says.
By reprocessing data through a deep-learning system, organizations are developing algorithms and models that are approaching or eclipsing the level of precision of algorithms and models used in traditional machine learning.