9 AI terms to know

  • Small
  • Medium
  • Large

The future is robotic; at least that's what some healthcare technology leaders believe.

Take doc.ai, for example. This Palo Alto, Calif.-based startup has engineered a platform where patients can discuss their medical data with an artificially intelligent physician robot.

IBM Watson uses AI to interpret genomic data in just 10 minutes, a task that takes human experts 160 hours.

Even Microsoft has leveraged one application of AI, called machine learning, in its antivirus software.

It's time to face the facts: AI is on its way to powering nearly all aspects of the healthcare industry, from advancing cybersecurity to expanding precision medicine techniques.

Here are nine terms related to artificial intelligence that healthcare executives should know, compiled with the help of Science.

1. Artificial intelligence describes intelligent behaviors commonly associated with humans but exhibited by machines. It often takes the form of tasks like problem-solving. AI might apply an algorithm, or series of algorithms, to an artificial neural network to train itself for various tasks.

2. An algorithm is a set of simple or complex step-by-step instructions a computer or software system uses to complete a specific task, such as to solve a problem, answer a question or deliver an alert.

3. A neural network is an abstracted, simplified model of the human brain used in machine learning. The goal of a neural network approach is to enable a machine to replicate problem-solving processes that take place in the human brain. Through inputted algorithms, a developer trains a neural network to understand tasks based on examples. When given a problem, the neural network will then analyze it using layers of artificial neurons.

4. Machine learning is an application of AI that uses algorithms to find patterns in data without instruction. Machine learning automates a system's ability to learn, so it can improve from experience without being programmed for each individual task it completes. A machine learning model is "trained" on relevant examples from diverse data sources.

5. Backpropagation is the method by which many neural networks learn. Specifically, neural networks find differences between their output and their desired output to adjust the calculations of each artificial neuron. Backpropagation is a type of "supervised learning," because it requires a machine to already know an established desired output. Using a backpropagation algorithm, a neural network that reaches an undesired outcome calculates the error contribution of each artificial neuron to adjust its output and learn from its mistakes.

6. Deep learning, a subset of machine learning, is how a neural network with multiple layers of artificial neurons becomes more sensitive to intricate patterns in a data set. Deep learning builds upon neural networks and machine learning techniques by applying deep networks with unsupervised learning.

7. Natural language processing is a computer's attempt to interpret written or spoken language, often using machine learning. Because language is so complex, computers must carefully parse vocabulary, grammar and intent, while also allowing for variation in word choice when processing language, which is why programmers often take multiple AI approaches to NLP.  

8. Cognitive computing simulates human processes related to thought and learning through data-mining, pattern recognition and natural language processing. Cognitive computing builds upon neural networks and deep learning to build a system that covers multiple disciplines, including machine learning, natural language processing, speech recognition and human-computer interaction.

9. The Turing test is a test of the AI's ability to pass as human. To pass a Turing test, a machine would have to exhibit behavior at least as intelligent as a typical human. This is the original test centered on a machine's ability to successfully converse with a human evaluator.

More articles on artificial intelligence:

Facebook, Microsoft team up for AI interoperability

Sophia Genetics raises $30M for AI, genomics diagnosis efforts

Horizon BCBS New Jersey, Pager partner to help members navigate healthcare landscape

Copyright © 2021 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.


Featured Whitepapers

Featured Webinars