How Does Machine Learning Work? Definitions & Examples

Machine learning (ML) is one of the most impactful technological advances of the past decade, affecting almost every single industry and discipline. From helping businesses provide more advanced, personalized customer service, to processing huge amounts of data in seconds, ML is revolutionizing the way we do things every day.

So, what is machine learning — and how is it used across industries? In this article, we’ll discuss the applications of machine learning, how the technology works across various sectors and why you should consider enhancing your own professional repertoire with machine learning skills.

What Is Machine Learning?

According to IBM, machine learning is a type of artificial intelligence (AI) that can improve how software systems process and categorize data. The term itself describes the process — ML algorithms imitate human learning and gradually improve over time as they take in larger data sets. Machine learning is a complex topic with a lot of variables, but our guide, What Is Machine Learning, can help you learn more about ML and its many uses.

Generally, there are four types of machine learning: supervised, unsupervised, semi-supervised and reinforcement. Each type of machine learning excels at different tasks, so understanding the differences between them can help you know when to deploy each type.

An image that compares four types of machine learning, supervised, unsupervised, semi-supervised and reinforcement.

Supervised machine learning, also referred to as supervised learning, uses weighted data sets to help an algorithm learn. In practice, a supervised learning algorithm is “taught by example.” Training data is fed into an algorithm in a pre-sorted and tagged manner, which teaches the algorithm how to categorize based on pre-set characteristics. After enough training, the algorithm can similarly sort raw data that hasn’t been pre-sorted or tagged. Supervised machine learning algorithms are great at classifying data into pre-set categories, like identifying spam emails or telling different images apart from one another.

Unsupervised machine learning uses algorithms to cluster and sort unlabeled and unstructured data. These algorithms are good at discovering patterns in data without human intervention. This is useful in situations without clear delineations between different data categories. For example, unsupervised machine learning is often used to recommend similar items to a customer after they’ve added an item to their shopping cart online. There are many situations where data won’t fit into tidy categories, and this is where unsupervised machine learning algorithms shine.

Semi-supervised machine learning combines supervised and unsupervised machine learning techniques and methods in order to sort or identify data. Semi-supervised learning involves labeling some data and providing some rules and structure for the algorithm to use as a starting point for sorting and identifying data. Using a small amount of tagged data in this way can significantly improve an algorithm’s accuracy. A common application of semi-supervised learning is to classify content in scanned documents — both typed and handwritten. Generally, semi-supervised learning algorithms use features found in both structured and unstructured algorithms in order to achieve this objective.

Reinforcement learning is used when an algorithm needs to make a series of decisions in a complex, uncertain environment. Reinforcement learning algorithms are designed with a game-like rules system to reinforce or discourage certain choices, with the algorithm attempting to maximize its “score.” When building reinforcement learning algorithms, programmers set up the game’s “rules” but do not provide suggestions as to where the computer should begin. The computer then uses trial and error to develop the optimal solution to the issue at hand. Reinforcement learning algorithms are used for language processing, self-driving vehicles and game-playing AIs like Google’s AlphaGo.

Machine learning is different from data science. Machine learning trains algorithms to identify and categorize different data types, while data science helps professionals check, clean and transform data for this use. Understanding the differences between these processes is important for anyone interested in machine learning.

Machine learning is applicable in multiple professional fields. You can learn more about machine learning in various ways, including self-study, traditional college degree programs and online boot camps. Machine learning is part of the Berkeley Data Analytics Boot Camp curriculum, which gives students insights into how machine learning works. Another field where machine learning shows up frequently is in finance. Berkeley FinTech Boot Camp can help demonstrate how machine learning works specifically in the finance sector.

The best option for you depends on your personal interests, goals and the field you want to pursue. Below, we dive deeper into eight industries and how ML is used in each.

An image that showcases the eight industries that can benefit from machine learning, as listed in the article.

How Does Machine Learning Work in Finance?

Finance and machine learning are a natural pair. Finance is a very data-heavy profession, and machine learning focuses on processing and categorizing vast amounts of that data efficiently. Machine learning in finance can help organizations process raw data, find trends and create data models surrounding financial products.

Finance Machine Learning Examples

Security and Fraud Protection

Security is essential for finance, banking and financial services companies. As companies get larger and work with more customers, reviewing all transactions and process documents for new user signups can quickly become daunting. Machine learning algorithms can help finance companies efficiently process all of this data. For example: Coinbase, a major cryptocurrency exchange, uses machine learning in the user signup process to detect fraud. If a user uploads documents that don’t match the information provided, their documents will be flagged or rejected. This helps prevent fraud on the platform, reducing staffing costs and fraud-related expenditures.

Process Automation

Machine learning also provides opportunities to automate processes that were once the sole responsibility of human employees. This is a broader example across many industries, but the data-driven financial sector is especially interested in using machine learning to automate processes. For example, the total value of insurance premiums underwritten by artificial intelligence applications is expected to grow to $20 billion by 2024. This is because AI- and ML-assisted processes can onboard customers more quickly and streamline the underwriting process.

Are you interested in learning more about machine learning in finance? Berkeley FinTech Boot Camp can help you learn the skills you need to jump-start your career in finance.

Get Program Info


Step 1 of 6

How Does Machine Learning Work in Healthcare?

Machine learning is a natural match for data-driven fields like healthcare. In the healthcare space, ML assists medical and administrative professionals in analyzing, categorizing and organizing healthcare data. ML systems help hospitals and other medical facilities provide better service to patients regarding scheduling, document access and medical care. AI and ML are helping to drive medical research, and IBM’s guide on AI in medicine can help you learn more about the intersection between healthcare and AI/ML tech.

Healthcare Machine Learning Examples

Helping Pharmacists Dispense Medication

For the consumer, picking up medication at the pharmacy often feels like a simple transaction, however, the situation behind the pharmacy counter is a different story. Pharmacists have to use information from doctors, patients, insurance companies and drug manufacturers in order to prescribe medication effectively. Historically, this process involved many data silos and made it difficult for pharmacists to get a complete picture regarding patient information. Walgreens worked with Microsoft Azure to implement a machine-learning-powered back end system to improve their quality of care.

Every time a new prescription request is processed, Walgreens’ new system intelligently queries hundreds of data points, almost instantly, and provides a dashboard for pharmacists to get a more complete picture of drug stock levels, patient information, insurance information and more. The AI-powered system takes in all of the information for each patient, and provides individualized information for the pharmacist. This system enables Walgreens to provide better care to its customers, ensuring the right medications are delivered at the right time.

Drug Discovery and Development

Bringing a new drug to market can cost around $3 billion and take around 2–14 years of research. Designing new molecules is the main reason for the cost and time — it’s an incredibly labor-intensive and complex process. Unstructured machine learning algorithms can create optimal molecule candidates for testing, which significantly speeds up the process. This can help drug manufacturers develop new medicine more quickly and cost-effectively, ultimately helping patients with new drug therapies.

How Does Machine Learning Work in Entertainment?

There are so many options for entertainment these days, between video streaming services, music, podcasts and more. Many of these services use machine learning for a critical purpose — personalizing recommendations. People can get easily overwhelmed by the amount of choices available. YouTube, for example, states that over 500 hours of content are uploaded to the video hosting platform each minute. Using ML can help people discover the shows, music and platforms best suited to their unique preferences.

Entertainment Machine Learning Examples

One great example of recommendations in entertainment comes from Netflix. As noted on Netflix’s machine learning research page, the company supports 160 million customers across 190 countries. Netflix offers a vast catalog of content across many genres, from documentaries to romantic comedies to everything in between. However, every customer’s tastes vary. Netflix uses machine learning to bridge the gap between their massive content catalog and their users’ differing tastes.

Machine learning can recommend new content to watchers, readers or listeners based on their preferences. Netflix takes data from its users — the kinds of things they’ve watched, how long they’ve watched them and any thumbs up/thumbs down ratings provided by the user — to match users with recommended content from its extensive catalog.

How Does Machine Learning Work in Education?

Education is a field that is historically slow to change. In part, this is due to the fact that the efficacy of methods and tools used in education need to be studied and understood before being deployed more broadly. As machine learning becomes more common, its influence on education has grown. Machine learning in education can help improve student success and make life easier for teachers who use this technology.

Education Machine Learning Examples


Machine learning algorithms can efficiently process and transcribe spoken audio, which can be beneficial to certain students who struggle with note-taking. This is especially true for students who are deaf or hard of hearing, as well as for students with ADHD or dyslexia. is one example of an ML-powered note-taking service designed for professional and educational use. The service allows students to upload audio recordings of class and receive a written transcript of the material from that recording.

Student Success

Many factors contribute to a student’s success, and navigating the education system can be difficult — especially for first-time college students. One use case for machine learning in education is identifying and assisting at-risk students. Schools can use ML algorithms as an early warning system to identify struggling students, gauge their level of risk and offer appropriate resources to help them succeed.

How Does Machine Learning Work in Manufacturing?

Manufacturing is another industry in which machine learning can play a large role. This field thrives on efficiency, and ML’s primary purposes, in this sense, revolve around upholding a reasonable level of fluidity and quality.

Manufacturing Machine Learning Examples

Machine learning models can help improve efficiency in the manufacturing process in a number of ways. An article in the International Journal of Production Research details how manufacturing and industrial organizations are using machine learning throughout the manufacturing process. For example, computer vision algorithms can use machine learning to perform automatic quality control functions on a manufacturing line. These algorithms can improve supply chain efficiency, inventory control, loss reduction and delivery rate improvement.

How Does Machine Learning Work in Customer Service?

Customer service is an essential part of any organization, but it’s often time-consuming, requires a large talent expenditure and can have a major impact on a business if implemented poorly. Machine learning can help brands with their customer service efforts, as listed in the examples below.

Customer Service Machine Learning Examples

Machine learning is on track to revolutionize the customer service industry in the coming years. According to Gartner, one in four organizations is currently deploying AI and ML technologies, but 37.5 percent of customer service leaders are investigating or planning to deploy chatbot machine learning solutions by 2023. Machine learning applications equipped with natural language processing (NLP) technology can answer customer questions automatically, allowing customer service employees to focus on more complex and important customer issues. Algorithms can offer superior personalization and provide quick, efficient assistance for customer issues.

How Does Machine Learning Work in Supply Chain?

Supply chain management is incredibly complex. It requires tracking a high number of components and/or products, knowing their current locations and helping them arrive at their final destinations. Machine learning modernizes the supply chain industry in ways we never thought possible.

Supply Chain Machine Learning Examples

Computer Vision

Computer vision is precisely what it sounds like — a machine learning algorithm that gives a computer the ability to “see” and identify objects through a video feed. There are many use cases for this technology across the supply chain industry. For example, computer vision algorithms can enable robots to navigate a warehouse and move products safely and efficiently. This technology is also used for reading barcodes, tracking products as they move through a system and inspecting packages for damage.

Demand and Inventory Optimization

Supply chain management uses data-based predictions to help organizations forecast the amount of inventory to stock and where it should be along the supply chain. Machine learning is now a major tool for supply chain management. ML algorithms can help forecast changing demand and optimize inventory to keep products flowing through a supply chain. Sensors, RFID tags, meters and GPS modules can feed information into the machine learning system, allowing the algorithm to know where items are throughout the supply chain and adjust plans based on changing circumstances or identified obstacles. Machine learning is likely to become an even more important part of the supply chain ecosystem in the future.

How Does Machine Learning Work in Advertising

As the internet becomes a more significant part of our lives, the technologies that support its functionality will become more complex. Many online businesses generate revenue through advertising, and advertising companies use advanced systems to try and provide the most relevant ads for every consumer. Getting involved in the advertising industry can be a great career path for anyone with ML skills.

Advertising Machine Learning Examples

Advertising platforms want to match each potential customer with the best possible ad for their situation, as personalization can help drive better advertising performance. Facebook uses machine learning heavily in its recommendation algorithm: Its model incorporates existing behavioral data from each user and predicts the types of advertisements most beneficial for each individual. This is a great use case for machine learning because of the large amount of data that is processed. The recommendations also continue to improve over time, as more data is fed into the system, allowing for more personalized advertising for every user.

How to Get Started with Machine Learning

Machine learning is an exciting and high-growth sector. As technology advances, organizations will continue to collect more and more data to grow their companies. Being able to process that data effectively will be critical to their success. We’ll discuss how to get started with machine learning in this section.

One great way to learn machine learning is through a boot camp. Online boot camps provide flexibility, innovative instruction and the opportunity to work on real-world problems to help you get hands-on experience. These online programs provide the flexibility needed to learn machine learning in 24 weeks while maintaining your work or college schedule.

Machine learning is an integral part of multiple fields, so there are many opportunities to apply your ML skills. Berkeley Data Analytics Boot Camp offers a market-driven curriculum focusing on statistical modeling, data visualization and machine learning. Another option is Berkeley FinTech Boot Camp, a curriculum teaching marketable skills at the intersection of technology and finance. Topics covered include financial analysis, blockchain and cryptocurrency, programming and a strong focus on machine learning and other AI fundamentals.

Are you interested in machine learning but don’t want to commit to a boot camp or other coursework? There are many free resources available as well. This list of free STEM resources for women and girls who want to work in machine learning is a great place to start. These kinds of resources allow you to get started in exploring machine learning without making a financial or time commitment.

No matter how you get started, ML skills are valuable and can help you progress even in your current career. The importance of data and machine learning will only be more profound in the future, and learning these skills now will help you keep your competitive edge no matter what industry you’re in or plan to transition into down the road.

Get Program Info


Step 1 of 6