The main differences between the AI-based rules and machine learning

0
107


Companies from all industries to explore and implement artificial intelligence (I) projects, the Big Data of robotics, automate business processes, improve the customer experience and innovate in the development of products. After McKinsey, ” to adopt the AI promises considerable benefits to companies and economies, thanks to its contributions to productivity and growth “. But with this promise comes challenges.Computers and machines do not come in this world with an inherent knowledge or understanding of how things work. Like humans, they need to learn that a red light means stop and green go. So, how do these machines get really the intelligence they need to perform tasks such as driving a car or diagnose a disease ?

Data or bust

There are many ways of achieving AI, and existential for all of them is data. Without data quality, artificial intelligence is a pipe am. There are two ways to manipulate the data , either by using rules or machine learning, to reach the IA, and some best practices to help you choose between the two methods.

Rules-based systems

Well before the AI and machine learning (ML) will become common terms outside of the realm of high technology, the developers encodaient human knowledge in computer systems rules that are stored in a knowledge base. These rules define all aspects of a task, usually in the form of statements “” ( “if A, then do B, else if X, then do Y“).

Although the number of rules that must be written is dependent on the number of shares that you want a system manages (for example, 20 shares mean write and encode manually at least 20 rules), the rules-based systems are generally less effective, more profitable and less risky, since these rules will not change or will not be updated by themselves. However, the rules may limit the capacities of AI with an intelligence rigid, which may not do what they were written to do.

Automatic learning systems

Although a rules-based system might be seen as having an intelligence ” fixed “, system, machine learning is adaptive and attempts to simulate human intelligence. There is still a layer of underlying rules, but instead of a human writing a fixed set, the machine has the ability to learn new rules on its own, and discard those that no longer work.

In practice, there are several ways that a machine can learn, but the supervised training , when the machine receives data to train, is usually the first step of a program of machine-learning. Finally, the machine will be able to interpret, categorize, and perform other tasks with data that are not labelled or unknown information on its own.

Where to start the strategy of the IA organization :

The anticipated benefits to the AI are high, so that the decisions that a company makes at the beginning of its execution can be critical to success. Foundational aligns your technology choices on the commercial objectives underlying that the AI has been set to achieve. What problems are you trying to solve or what challenges are you trying to address?

The decision to implement a system of machine-learning or rules-based will be a long-term impact on the evolution and scale of the program IA a business. Here are a few best practices to consider when assessing the approach that best suits your organization :

When you choose a rules-based approach, it is logical :

  • Results fixed: When there is a low number or fixed results. For example, there are only two states for which a button “Add to cart” can be set pressure or not. Although it is possible to use machine learning to detect whether a user has pressed the button, it would not be logical to apply this type of method.
  • Risk of error : The penalty for error is too high to risk a false positive and, therefore, only rules, which will be 100% accurate, should be implemented.
  • Do not plan for ML: If those who maintain the system have no knowledge of machine learning and the company has no plans for the source to move forward.

When to apply machine learning :

  • The simple rules do not apply to : When there is no easily definable to solve a task using simple rules
  • Speed of change : When the situations, the scenarios and the data is changing faster than the ability to write continuously to new rules.
  • Natural language processing : Tasks that require an understanding of the language, or natural language processing. Since there are an infinite number of ways to say something, it is unrealistic, if not downright impossible, to write rules for the language normal. The innate intelligence and adaptive machine learning is optimized for the scale.

The promises of the AI are real, but for many organizations, the challenge is to start. If you are in this category, you must first determine if a method based on rules or ml will work best for your organization.

This article was originally published by Elana Krasner On TechTalks, a publication that examines the technology trends, how they affect the way we live and do business, and the problems that they solve. But we also discuss the bad side of the technology, and the implications more dark of new technologies and what we need to watch out for. You can read the original article here.

Published June 13, 2020 — 13:00 UTC



LEAVE A REPLY

Please enter your comment!
Please enter your name here