Data, in all forms, is expanding as a resource to be utilized. Yet in many industries and professions, the data explosion is outstripping the human capacity to understand the meaning hidden within that data. Patterns for Cognitive Computing Systems can unlock the potential in all data, internal, external, structured, unstructured, voice, and visual, and make it work together.

Cognitive Computing is perhaps most unique in that it alters the established IT doctrine that a technology’s value diminishes over time; because Cognitive Systems improve as they learn, they become more valuable. This quality among others makes cognitive technology highly desirable for business, and many early adopters are leveraging the competitive advantage it affords.

Think of a Pattern as a use case on a specific scenario such as Event based Real Time Marketing for Real Time Analytics, Anti-Money Laundering, address Data Oceans by reducing cost of Hadoop, and so on. Those are just a few of the Cognitive Patterns now available. Patterns identify and address pain points for cognitive infrastructures. These entry points then help you understand where you are on the cognitive journey and enables IBM to demonstrate the set of solutions capabilities for each lifecycle stage.

Here are 5 things to know about IBM Patterns for Cognitive Computing Systems:

1.  What are Patterns for Cognitive Computing Systems?

Patterns for Cognitive Computing Systems are solutions addressing a combination of customers’ challenges and pain points utilizing IBM Power Systems Solutions portfolio to deliver a viable recommendation to help solve the problem. Patterns for Cognitive Computing Systems helps to classify (prediction) with accuracy and can determine appropriate products to be recommended to customers. For example, utilize history for different customers, ratings on products and Machine and Deep Learning techniques to make targeted recommendations that ensure better product purchase rates, click rates, customer retention rates, better customer experiences, and others.

2. Patterns for Cognitive Computing Systems helps to eliminates data infrastructure silos and manages information lifecycle.

In the fast-paced environment customers are faced with, information storage often gets optimized to a single application or departmental requirement. This leads to information silos in an organization and creates significant difficulties in synchronizing and sharing this information across the enterprise. These silos can be spread across many different storage and file system technologies, further complicating this unifying process. Utilizing Spectrum scale and IBM Elastic Storage Server, for example, a customer can logically or physically combine these silos into a unified, more manageable and sharable storage environment.

3. Patterns for Cognitive Computing Systems helps to backup, restore and archive big data.

With the massive amounts of data that comes with big data implementations, it is often very difficult to back up business critical sets of data. Utilizing IBM Spectrum family of products (Spectrum Scale, Spectrum Archive, Spectrum Protect) and low cost and high capacity storage capabilities (Cloud Object Storage, Elastic Storage Server, Tape) customers can establish economic and performing backup, restore and archive capabilities for their big data environments.

4.  Patterns for Cognitive Computing Systems helps to meet performance requirements for technical workloads.

As requirements for instantaneous analysis of information grows with high-performance computing (HPC), high-performance data analysis (HPDA) and cognitive workloads, the high-performance access of data becomes critical. Utilizing Elastic Storage Server with Spectrum Scale Native RAID provides high data reliability, high data integrity, consistent performance.

5.  Patterns for Cognitive Computing Systems helps to augment data warehouses with Hadoop.

This cognitive pattern leverages Hadoop to help customers optimize and improve cost effectiveness of their data warehouses. Traditional, proprietary data warehouse systems are very expensive to purchase, upgrade and maintain, making them a highly desirable area for cost reduction. These systems also have difficulty in supporting unstructured data. However, they tend to be very integrated into customer business operations making them difficult to eliminate. Customers often enrich the structured data in a data warehouse with unstructured data in a Hadoop environment to provide a complete data environment.

The idea is to start the conversation with customers, listen to their challenges and pain points, and bring answers to the challenges discussed during these conversations incorporating IBM Power Systems Solutions portfolio.  For more information and examples of some of the solutions already available across industries, refer to the following documentation.

IBM Cognitive Systems PowerAI makes deep learning, machine learning, and AI more accessible and more performant.  Refer to the following webpage for more details:

https://www-01.ibm.com/common/ssi/cgi-bin/ssialias?htmlfid=POD03135USEN

For IBM Watson: AI and Cognitive computing solutions with IBM Power Systems, and for a cognitive computing system designed for complex analytics -- integrating massively parallel POWER processors and DeepQA technology can be read at the following website:  https://www.ibm.com/power/solutions/cognitive-computing

For Enterprise AI, refer to the following website:  https://www.ibm.com/it-infrastructure/us-en/resources/power/enterprise-ai/

For IBM Power Systems solutions information, refer to the following website:  https://ibm.co/2DOzPoS

Cesar Maciel
Marcos Quezada
Dino Quintero