Welcome to the third installment of ‘Understanding key automation technologies.’ – this week our focus is Cognitive Computing (CC).
As with installments one and two, ‘4 things to know about Robotics Process Automation’ and ‘Machine Learning’ respectively, the aim of this article is to scratch the surface of complicated automation technology (both can be found on our Roots Automation blog). The information is but an introduction to each technology, key uses and advantages.
Without further ado, let’s dive in and see what Cognitive Computing is about.
1. Definition: Cognitive Computing uses Machine Learning and other sophisticated software to stimulate human behaviors. The aim is for a computer-based system to mimic the human thought process to help humans make better decisions, not to make decisions for them.
2. What’s the difference between ML and CC?
Let’s recap Machine Learning from our previous article: Machine Learning is an Artificial Intelligence (AI) function which uses algorithms to enable computer programs to improve and learn over time, automatically.
- Cognitive Computing is also in the AI family and is often described as being a merger of computer science and cognitive science, whereas ML would sit in the computer science category.
- Computer science = a broad term encompassing all things related to digital computers – principles, hardware and software, algorithms, data processing.
- Cognitive science = is the study of the human mind and includes linguistics, anthropology, philosophy, psychology, artificial intelligence, neuroscience
- In a nutshell, CC is a step closer to replicating human behaviors than ML.
- CC is closer to human behavior because it uses the computer science elements that ML does (such as Natural Language Processing (NLP) pattern recognition and speech recognition) as well as further advanced self-learning capabilities such as deep learning, and Human and Computer Interaction (HCI).
3. Cognitive Computing criteria
Here are several common Cognitive Computing features.
- Adaptability
- A Cognitive Computing system must be able to learn from its environment and evolve over time.
- Interactive
- If the aim of CC is to mimic the human thought process and to support human’s decision making, the system itself must allow interaction between human and computer. The interaction experience should appear to be natural and a similar experience to human and human interaction.
- Iterative and stateful
- If the human request is unclear, the system should ask questions to help define the need.
- Contextual
- Human thought process takes a lot of context into account. Computers must also consider several contextual elements to begin replicating this process, such as: sensor data – a system noticing and responding to something in the physical word, e.g. speech, visuals, touch; syntax (language rules and principles, how sentences and phrases are constructed in certain orders); time, goals and so on.
4. Advantages of CC
We’ve mentioned that CC aids better decision making using the methods above, and we’ll illustrate how it does this with a handful of examples.
CC aims to gather all necessary information to support the human and help them make an informed decision, not to resolve a problem or task without human input.
- One example is the Assisted Driving feature from Tesla, this allows a system to drive a car, but the overall responsibility remains with the driver who can override the system if they would like to take control.
- Another example is in the medical world. CC is becoming a valuable assistant for understanding patterns and diagnosing patients - it needs to understand the full context regarding the patient, their age, medical history, prognosis. It can also assess vast amounts of data to help decide on a course of treatment, if required.
- Finally, customer service is another area seemingly reaping the benefits of CC capabilities, CC ‘agents’ use contextual information about the customer to provide genuinely useful information, support and suggestions.
There are countless more business areas CC already supports, from Insurance, Finance, and HR to boosting employee satisfaction by providing new information and patterns that may have gone undetected by humans alone.
To continue the conversation, book a free demo with our team or see how we can train the perfect Digital Coworker for your business, get in touch today: info@rootsautomation.com
Next and last in this series, 4 things to know about Optical Character Recognition.
Automation, Cognitive Computing, CC