Building a Platform Requires AI Diversity & Partnership

ML is perhaps the most relevant subset of AI to the average enterprise today. As explained in the Executive’s guide to real-world AI, our recent research report conducted by Harvard Business Review Analytic Services, ML is a mature technology that has been around for years.

ML is a branch of AI that empowers computers to self-learn from data and apply that learning without human intervention. When facing a situation in which a solution is hidden in a large data set, machine learning is a go-to. “ML excels at processing that data, extracting patterns from it in a fraction of the time a human would take, and producing otherwise inaccessible insight,” says Ingo Mierswa, founder and president of the data science platform RapidMiner.

An explainer article by AI software company Pathmind offers a useful analogy: Think of a set of Russian dolls nested within each other. “Deep learning is a subset of machine learning, and machine learning is a subset of AI, which is an umbrella term for any computer program that does something smart.”

In our plain English primer on deep learning, we offer this basic definition: the branch of AI that tries to closely mimic the human mind. With deep learning, CompTIA explains, “computers analyze problems at multiple layers in an attempt to simulate how the human brain analyzes problems. Visual images, natural language, or other inputs can be parsed into various components in order to extract meaning and build context, improving the probability of the computer arriving at the correct conclusion.”

Deep learning uses so-called neural networks, which “learn from processing the labeled data supplied during training, and uses this answer key to learn what characteristics of the input are needed to construct the correct output,” according to one explanation provided by deep AI. “Once a sufficient number of examples have been processed, the neural network can begin to process new, unseen inputs and successfully return accurate results.”

When you take AI and focus it on human linguistics, you get NLP. SAS offers one of the clearest and most basic explanations of the term: “Natural language processing makes it possible for humans to talk to machines.” It’s the branch of AI that enables computers to understand, interpret, and manipulate human language.

NLP itself has a number of subsets, including natural language understanding (NLU), which refers to machine reading comprehension, and natural language generation (NLG), which can transform data into human words.

But, says ISG’s Butterfield, the premise is the same: “Understand language and sew something on the back of that understanding.”

NLP has roots in linguistics, where it emerged to enable computers to literally process natural language, explains Anil Vijayan, vice president at Everest Group. “Over the course of time, it evolved from rule-based to machine-learning infused approaches, thus overlapping with AI,” Vijayan says.

NLP might employ both ML learning and deep learning methodologies in combination with computational linguistics in order to effectively ingest and process unstructured speech and text datasets, says JP Baritugo, director at business transformation and outsourcing consultancy Pace Harmon.

As NLP is to speech, computer vision is to sight. Per SAS, computer vision is the “field of artificial intelligence that trains computers to interpret and understand the visual world. Using digital images and deep learning models, machines can accurately identify and classify objects – and then react to what they ‘see.’”

While early computer vision efforts date back to 1950, the confluence of hardware and software advances, along with an influx of new visual data from mobile devices and other cameras, are setting off a computer vision renaissance. Indeed, 82% of respondents to a 2019 survey by Harvard Business Review Analytic Services said that they would be assessing, piloting, implementing, or would have adopted computer vision capabilities within two years.

Computer vision can learn to view and interpret the visual world in much the same way humans do – and as AI capabilities have advanced, it can enable machines to gauge things people can’t, such as temperature or air quality. Incorporating deep learning, computer vision tools get better at detecting patterns in images or other data over time. Accuracy rates for object identification and classification have gone from 50% to 99% in less than a decade — and today’s systems are more accurate than humans at quickly detecting and reacting to visual inputs. What’s more, computer vision capabilities can process, categorize, and understand images and video at a scale and speed that would otherwise be impossible.

AI can seem like a black box: Input, then output – with little transparency on how the machine got from point A to point B. As PubNub CTO and co-founder Stephen Blum explained in our article, What is Explainable A!?, that may be acceptable for use cases like chatbots or social sentiment analysis, but for others – autonomous vehicles, aerial navigation, and drones, military applications – “being able to understand the decision-making process is mission-critical. As we rely more and more on AI in our everyday lives, we need to be able to understand its ‘thought process’ and make changes and improvements over time.”