The pace of change in the information industry is impressive and exhausting. Information continues to grow at an exponential rate and is increasingly variable, diverse and noisy.
The cost of storing and processing information continues to decline – also at an exponential rate. The number of tools and technology platforms that promises to organize, connect and extract value from information is disorienting.
As a technologist, these are exciting times. But, in an increasingly connected world with information traveling instantaneously and the rules of the game always changing, no one can afford to “wing it.”
Enter the world of cognitive computing
Cognitive computing is a suite of technologies and design principles that aim to develop the next generation of smart applications. These are applications users can talk to and ask questions. Applications designed to help users understand the complex nuances of knowledge tasks and make better decisions. And applications that are task-focused, user-centric and able to learn from experience.
Underneath the hood of cognitive computing are familiar, yet rapidly evolving, technologies; things like natural language processing, machine learning, text and data analytics, knowledge representation and reasoning.
Cognitive computing is a critical part of the answer
To be sure, cognitive computing is not the whole answer; but it is a critical part of the answer and is set to transform how knowledge work is done. McKinsey estimates the economic impact of cognitive computing on knowledge work at $5.2 – 6.7 trillion annually by 2025. This is massive, even if McKinsey is off by an order of magnitude.
Gartner advised their clients: “… the risk of investing too late in smart machines is likely greater than the risk of investing too soon.” Other organizations including KPMG, Deloitte and IBM are presenting themselves as providers of holistic cognitive and analytic solutions.
The objective of cognitive computing is to simplify and transform how knowledge work gets done. At a high level, knowledge workers:
- Find information.
- Analyze and try to understand such information, and
- Make decisions based on this understanding.
The process is often iterative and is certainly more nuanced, but such an abstraction provides a directional guide for the kind of capabilities needed.
Cognitive computing brings to the picture important new capabilities. For example, where today the search process is transactional (subsequent queries are treated as independent transactions), the intelligent machine will interact with the user in a stateful dialogue using more natural language. On top of the content and metadata is a knowledge base that supports reasoning. Continuous learning, personalization and self-adaptation are also critical characteristics of cognitive systems.
|Today||Powered by Cognitive Computing|
|Find||Find the information that will satisfy a need. A combination of pull, push and navigate. Pull is primarily through search. Push consists of alerts, recommendations and explicitly designed information widgets. Navigate is achieved via links and relationships. Uses content and metadata.||Expand the range of acceptable input, produce more focused output. Contextual, stateful dialogue with user to explore information need. Close thegap between the information need in its native form and acceptable system input. Leverage knowledge plus content and metadata.|
|Analyze||Make sense of the found information in the context of the task. Primarily done by user based on information and content found.-||Machine reasons over knowledge base, generates and tests hypotheses, collects evidence. User interacts with machine, guides analysis and explores the reasoning. Interactions are fed back as new knowledge. The system learns.|
|Decide||Form conclusions, take action. Today decision making is primarily a user task.||Machine recommends answers, not just documents. Case-based reasoning capabilities to allow users to utilize their own experiences and those of their colleagues. Workers perform scenario simulation and analysis. Decisions are fed back as new knowledge.|
Keys to success
Cognitive recipes contain three key ingredients: domain expertise, content and technology.
Domain expertise: The collective human expertise in an industry or domain and intimate familiarity with how people get their jobs done. This knowledge is critical to ensure the right problems are solved and compelling experiences for users are designed.
Content: Cognitive solutions often depend on huge amounts of content and data. Machine learning algorithms require content to be trained, but once trained, they will amplify the ability to extract value and insights from content and deliver it to users.
Technology expertise is clearly key to building cognitive solutions, but success is not just about deploying cognitive tools. Three types of technology capabilities are needed:
- Foundational: A mature and robust technology stack that can be leveraged and extended.
- Vertical: Expertise in relevant and emerging cognitive technologies
- Practical: Skills and know-how in applying these technologies on our content and domains.
Without all the ingredients, effective use of cognitive computing is severely challenged.
The role of user experience
No post on cognitive computing is complete without a nod to user experience.
People don’t change how they do things because we ask them to. The new user experience must make perfect sense.
If the experience is right, users will be more tolerant of some technical gaps. Otherwise, nothing short of technology magic will convince professional customers to change how they do things now. Technologists are not in the magic business, so user experience it is.
Thomson Reuters Center for Cognitive Computing
The above sounds ambitious and it is. But it is also not a narrative in the future tense. In fact, this is what we are working on at our new Thomson Reuters Center for Cognitive Computing, a team of scientists, engineers and designers with specialized skills in cognitive technologies, based out of our Toronto Technology Center.
The center’s mission is to accelerate and drive the development of cognitive capabilities and solutions working closely with customers and technology partners.
We are developing capabilities that accept a wider range of input (e.g., queries, questions and documents as input); yet provide focused output (e.g., search results, answers and analysis).
We are also working on conversational capabilities that allow knowledge workers to ask follow up questions (to elaborate on or refine on previous questions). We are developing capabilities that will change how users interact with machines.
For the industries Thomson Reuters serves, we find ourselves in a unique position to deliver cognitive solutions. Our people have the deep domain expertise, especially when combined with the knowledge of our customers and partners. Our content is deep, authoritative and often unique. We are leaders in the development of intelligent solutions. This is our sweet spot: providing the intelligence, technology and human expertise our customers need to find trusted answers.
View the story as it appeared on Techvibes.