Machine learning refers to a computer system's ability to use data to learn how to perform a given task instead of being explicitly programmed to execute that task. This ability is especially important when there is a need for computers to perform functions that are hard to specify. The specification can be difficult because the task parameters are hard to formally describe. For example, it's challenging to explain in words what the difference between cats and dogs is, but it's easy to provide images that explain the difference. Another issue is that sometimes human knowledge is incomplete (e.g., specifying what a fraudulent transaction is). Machine learning algorithms make use of data to build models that can make decisions and predictions, such as whether a transaction is fraudulent or not.

People often use the terms “machine learning” and “AI” interchangeably. However, machine learning is only one of the many fields of AI.

The Fields of AI

John McCarthy, who coined the term in 1955, defined AI as “the science and engineering of making intelligent machines.” This is intentionally a broad definition because AI is a broad field with many specializations. 

Automated Reasoning

This subfield focuses on logical deduction. It defines problem domains by facts, and sets of rules and reasoning are applied to derive new rules and facts. We can use automated reasoning systems to build theorem provers or obtain solutions that fit a set of constraints, a classic example is creating a school’s class schedules.

Robotics: Motion And Manipulation

This subfield relates to control systems that let machines manage physical tasks like movement while adjusting to changes in the environment. 

Knowledge Representation

This field relates to the ability to represent abstract concepts about the world and the relationships between them. The idea is to represent information about the world in a form that a computer system can utilize to solve complex tasks such as deriving new relations between concepts or question-answering.

Perception

Machine perception is the capability of a computer system to interpret data similar to the way humans use their senses (like eyes or ears) to relate to the world around them. Perception includes computer vision (image recognition) and audio processing (speech recognition). 

Natural Language Processing (NLP)

This subfield is concerned with the interactions between computers and human (natural) languages. In particular, how to allow machines to read and understand human language. A sufficiently robust natural language processing system would enable the acquisition of knowledge directly from human-written sources. Some applications of NLP include information retrieval, text mining, question answering, and machine translation.

Several Other Smaller Fields of AI

These fields of AI include planning and creativity, artificial general intelligence, and expert systems.

Economic Benefits of Machine Learning

Of all the technologies in the AI family, machine learning is particularly versatile. Many contemporary business problems involve prediction based on a pool of complex data, often including multiple dependent variables. Machine learning already delivers real value in a wide variety of business environments. It detects changes in customer sentiment, alerts analysts teams to potential fraud patterns, and even saves lives by detecting heart attacks faster and more accurately than human operators at emergency call centers. Machine learning can even re-engineer business processes themselves.

The sector is poised for explosive growth. Analysts project artificial intelligence to create $36.8 billion in revenues across “almost every conceivable industry sector” by 2025. Forbes valued the global machine learning market at $1.58 billion in 2017, and expects it to reach $20.83 billion in 2024. That’s a CAGR growth rate of 44.06% between 2017 and 2024.

Machine learning also brings many predicted overall economic benefits. According to Forrester, machine learning is beginning to “replace manual data wrangling and data governance dirty work,” leading to embedded data analytics software providing U.S. companies with savings of over $60 billion. In general, they expect AI to add up to an additional 4.6% to the U.S. gross value added (GVA) by 2035, representing an additional $8.3 trillion in economic activity.

Why AI Is the Future of Growth

The banking sector recognizes machine learning as an efficient and complementary method of implementing regulatory requirements, such as fraud and money laundering detection, compared to manual human monitoring. Payday lenders such as LendUp and Avant provide automated lending decisions using machine learning techniques, and companies like Feedzai provide sophisticated automatic and intelligent financial crime-detection services.

In the broader financial sector, machine learning is used by a leading Japanese insurance firm to supplant human analysis of claims, and algorithmic stock trading and portfolio management are becoming the norm.

Key Takeaway

Machine learning is a key artificial intelligence technology to address many contemporary business problems that relate to insights from large-scale data. The marketplace recognizes its ability to transform how organizations make decisions, predictions, and reach insights.

AI and machine learning are no longer secret weapons reserved for larger banks with deep pockets. Download our eBook, Democratizing Machine Learning for Community Banks, to learn how community banks and credit unions are finding opportunities to level the innovation playing field.