What Is a Machine Learning Algorithm?

What are machine learning algorithms?

A machine learning algorithm is the method by which the AI ​​system conducts its task, generally predicting output values ​​from given input data. The two main processes involved with machine learning (ML) algorithms are classification and regression.

An ML algorithm is a set of mathematical processes or techniques by which an artificial intelligence (AI) system conducts its tasks. These tasks include gleaning important insights, patterns and predictions about the future from input data the algorithm is trained on. A data science professional feeds an ML algorithm training data so it can learn from that data to enhance its decision-making capabilities and produce desired outputs.

ML is a subset of AI and computer science. Its use has expanded in recent years along with other areas of AI, such as deep learning algorithms used for big data and natural language processing for speech recognition. What makes ML algorithms important is their ability to shift through thousands of data points to produce data analysis outputs more efficiently than humans.

A list of machine learning's business benefits.
Machine learning algorithms provide several benefits to businesses, including retaining customers, boosting efficiency and detecting fraud.

How ML algorithms work

A data scientist or analyst feeds data sets to an ML algorithm and directs it to examine specific variables within them to identify patterns or make predictions. The idea is for the algorithm to learn over time and on its own. The more data it analyzes, the better it becomes at making accurate predictions without being explicitly programmed to do so, just like humans would.

This training data is also known as data input. The data classification or predictions produced by the algorithm are called outputs. Developers and data experts who build ML models must select the right algorithms depending on what tasks they want to achieve. For example, certain algorithms lend themselves to classification tasks that would be suitable for disease diagnoses in the medical field. Others are ideal for predictions required in stock trading and financial forecasting.

Supervised vs. unsupervised algorithms

Most ML algorithms are broadly categorized as being either supervised or unsupervised. The fundamental difference between supervised and unsupervised learning algorithms is how they deal with data. Two other categories are semi-supervised and reinforcement algorithms.

Supervised algorithms

These algorithms deal with clearly labeled data, with direct oversight by a data scientist. They have both input data and desired output data provided for them through labeling.

Supervised algorithms typically serve two purposes: classification or regression. In classification problems, an algorithm can accurately assign different data into specific categories – such as dogs and cats — which becomes feasible with labeled data.

There are many real-world use cases for supervised algorithms, including healthcare and medical diagnoses, as well as image recognition. In both cases, classification of data is needed.

In regression problems, an algorithm is used to predict the probability of an event taking place – known as the dependent variable — based on prior insights and observations from training data — the independent variables. A use case for regression algorithms might include time series forecasting used in sales.

Unsupervised data

Unsupervised algorithms deal with unclassified and unlabeled data. As a result, they operate differently from supervised algorithms. For example, clustering algorithms are a type of unsupervised algorithm used to group unsorted data according to similarities and differences, given the lack of labels.

Unsupervised algorithms can also be used to identify associations, or interesting connections and relationships, among elements in a data set. For example, these algorithms can infer that one group of individuals who buy a certain product also buy certain other products.

Semi-supervised algorithms

However, many machine learning techniques can be more accurately described as semi-supervised, where both labeled and unlabeled data are used.

Reinforcement algorithms

Reinforcement algorithms – which use reinforcement learning techniques– are considered a fourth category. Their unique approach is based on rewarding desired behaviors and punishing undesired ones to direct the entity to being trained using rewards and penalties.

Types of machine learning algorithms

There are several types of machine learning algorithms, including the following:

  • Linear regression. A linear regression algorithm is a supervised algorithm used to predict continuous numerical values ​​that fluctuate or change over time. It can learn to accurately predict variables like age or sales numbers over a period of time.
  • Logistic regression. In predictive analytics, a machine learning algorithm is typically part of a predictive modeling that uses previous insights and observations to predict the probability of future events. Logistic regressions are also supervised algorithms that focus on binary classifications as outcomes, such as “yes” or “no.”
  • Decision trees. This is a supervised learning algorithm used for both classification and regression problems. Decision trees divide data sets into different subsets using a series of questions or conditions that determine which subset each data element belongs to. When mapped out, data appears to be divided into branches, hence the use of the word tree.
  • Support vector machines. SVMs are used for classification, regression and anomaly detection in data. An SVM is best applied to binary classifications, where elements of a data set are classified into two distinct groups.
  • Naïve Bayes. This algorithm performs classifications and makes predictions. However, it’s one of the simplest supervised learning algorithms and assumes that all features in the input data are independent of one another; one data point won’t affect another when making predictions.
  • Random forests. These algorithms combine multiple unrelated decision trees of data, organizing and labeling data using regression and classification methods.
  • K-means. This unsupervised learning algorithm identifies groups of data within unlabeled data sets. It groups the unlabeled data into different clusters; it’s one of the most popular clustering algorithms.
  • K-nearest neighbors. KNNs classify data elements through proximity or similarity. An existing data group that most closely resembles a new data element is the one that element will be grouped with.
  • Artificial neural networks. ANNs, or simply neural networks, are groups of algorithms that recognize patterns in input data using building blocks called neurons. These neurons loosely resemble neurons in the human brain. They’re trained and modified over time through supervised training methods.
  • Dimensionality reduction. When a data set has a high number of features, it’s said to have high dimensionality. Dimensionality reduction refers to stripping down the number of features so that only the most meaningful insights or information remain. An example of this method is principal component analysis.
  • Gradient boosting. This optimization algorithm reduces a neural network’s cost function, which is a measure of the size of the error the network produces when its actual output deviates from its intended output.
  • AdaBoost. Also called adaptive boostingthis supervised learning technique improves the performance of an underperforming ML classification or regression algorithm by combining it with weaker ones to form a stronger algorithm that produces fewer errors.
Diagram showing how machine learning boosting works.
The technique of boosting a machine learning algorithm can improve its overall performance.

Data scientists must understand data preparation as a precursor to feeding data sets to machine learning models for analysis. Learn the six steps involved in the data preparation process.

What Is Digital Literacy?

While the word “literacy” alone generally refers to reading and writing skills, when you tack on the word “digital” before it, the term encompasses much, much more.

Sure, reading and writing are still very much at the heart of digital literacy. But given the new and ever-changing ways we use technology to receive and communicate information, digital literacy also encompasses a broader range of skills—everything from reading on a Kindle to gauging the validity of a website or creating and sharing YouTube videos.

The term is so broad that some experts even stay away from it, preferring to speak more specifically about particular skills at the intersection of technology and literacy.

The American Library Association’s digital-literacy task force offers this definition: “Digital literacy is the ability to use information and communication technologies to find, evaluate, create, and communicate information, requiring both cognitive and technical skills.”

More simply, Hiller Spiers, a professor of literacy and technology at North Carolina State University, views digital literacy as having three buckets: 1) finding and consuming digital content; 2) creating digital content; and 3) communicating or sharing it.

Finding and Consuming

In some formats, “consuming” digital content looks pretty much the same as reading print. Reading a novel on a basic e-reader requires knowing how to turn the device on and flip pages back and forth, but other than that, it isn’t so different from reading a book. A PDF of a New York Times article looks a lot like the page of a printed newspaper, except that it appears on a screen.

Donald Leu, an education professor at the University of Connecticut and a recognized authority on literacy and technology, describes this kind of digital reading as “offline reading.”

“It’s not interactive, … there’s one screen, and you just have to read it,” he explained. “It’s the same as reading a [paper] page.”

"I read on my iPad when I'm in a car, or when I'm on a plane when I'm going on a trip.  When I'm at home, I read regular books," says Shota, a 3rd grade student at Indian Run Elementary School in Dublin, Ohio.

The added skills needed for this kind of reading take just a few minutes to teach.

In comparison, what Leu calls “online reading,” in which a digital text is read through the internet, requires a host of additional skills. For example, a New York Times piece viewed on the web may contain hyperlinks, videos, audio clips, images, interactive graphics, share buttons, or a comments section—features that force the reader to stop and make decisions rather than simply reading from top to the bottom.

“The text is designed so that no two readers experience it in the exact same way,” said Troy Hicks, a professor of literacy and technology at Central Michigan University.

The reader determines, among other things, when to click on videos or hyperlinks, how long to stray from the initial text, and whether and how to pass the information along to others.

The process of finding digital content to read also requires different skills than finding printed texts. In seeking print materials, students might flip through magazines or head to the library and search through stacks of books. They learn to use a table of contents and an index to locate information within a book.

But part of digital literacy is learning to search for content in an online space. Students have to query a search engine using keywords and navigate those results, including assessing the reliability of particular authors and websites.

Creating Content

Digital literacy also refers to content creation. That includes writing in digital formats such as email, blogs, and Tweets, as well as creating other forms of media, such as videos and podcasts.

Renee Hobbs, a professor of communication studies at the University of Rhode Island, talks about digital authorship as “a form of social power.” At a weeklong professional-development institute on digital literacy held at URI this past summer, she showed examples of student activists sharing their messages about the Black Lives Matter movement through YouTube videos.

Creating digital content is a “creative and collaborative process that involves experimentation and risk-taking,” she said. There’s more risk-taking than in print writing because digital writing is so often meant to be shared.

Sharing and Communicating

While traditional writing can be a personal endeavor, digital writing is generally intended to be communicated with others. And digital-writing tools are designed to make that easy to do.

As North Carolina State’s Spiers and her co-author, Melissa Bartlett, wrote in a 2012 white paper about digital literacy and learning, “Web 2.0 tools are social, participatory, collaborative, easy to use, and are facilitative in creating online communities.”

"It's on a book, on a paperback book because I've been reading like that since I was kid," says Hareem, a 10th grade student at Mineola High School, Mineola, NY

That makes digital writing a potentially powerful lever for social good, allowing students to “actively participate in civic society and contribute to a vibrant, informed, and engaged community,” as the ALA notes.

It also makes digital writing a potentially dangerous tool—decisions about when and what to share online can have repercussions for a student’s safety, privacy, and reputation.

For that reason, learning about appropriate internet behavior is also a part of digital literacy, many say.

“We need to help kids see they can use digital tools to create things and put things out into the world, but there’s a responsibility that comes with that,” said Lisa Maucione, who attended the URI institute and who is a reading specialist for the Dartmouth public schools in Massachusetts.

Evolving Technology

Because the term “digital literacy” is so wide-ranging, it can cause confusion. What exactly is someone talking about when he or she refers to digital literacy? Is it the consumption, creation, or communication of digital materials? Or is that person discussing a particular digital tool? Do technology skills like computer coding fall under the digital-literacy umbrella as well?

Some experts prefer the term “digital literacies,” to convey the many facets of what reading and writing in the modern era entails.

“The concept should instead be considered plural—digital literacies—because the term implies multiple opportunities to leverage digital texts, tools, and multimodal representations for design, creation, play, and problem solving,” Jill Castek, a research assistant professor with the Literacy , Language, and Technology Research Group at Portland State University, wrote in an email.

Leu of UConn avoids the term altogether.

“Is someone who is ‘digitally literate’ equally literate when searching for information, when critically evaluating information, when using Snapchat, when using email, when using text messaging, when using Facebook, or when using any one of many different technologies for literacy and learning?” asked Leu in an email. “I don’t think so.”

He prefers the term “new literacies,” which he says better conveys how rapidly technology is changing. Other experts have used terms like “literacy and technology,” “multiliteracies,” and “21st century literacies.”

But for now, digital literacy seems to be the prevailing term among educators. “I understand this is the term that is popular today,” Leu said, “just as I understand a newer term will appear in the future that will replace it.”

Opening the definition dialogue: Personalization, individualization and differentiation

A recent eSchool News contributed article, Differentiation, individualization and personalization: What they mean, and where they’re headedwhich helped define personalized learning in relation to both differentiation and individualization, is a wonderful reminder that while there is broad agreement on many aspects of the definition of personalized learning, there remains an open dialogue on other parts of this complex definition.

The piece cuts right to the heart of the issue by noting something I have often discussed—most of us struggle to clearly delineate differentiation, individualization, and personalization.

This struggle for a definition poses a larger question: If we cannot clearly and successfully define our approaches, what chance do we have for successful implementation?

Much of the article squares with some of the most widely accepted and agreed upon definitions around personalization. I would, however, like to offer some additional points to consider in forming a definition around this complicated topic.

Exploring the Term “Student Agency”

The article does a great job of delineating personalized learning from individualized learning by offering that “in a personalized scenario, the teacher is no longer the sole driver of instruction—each learner now collaborates with the teacher to drive his or her learning, with the students taking a hands-on role in determining their own needs and informing the design of their lessons.”

In breaking down this definition a bit further, I’d recommend applying the term “student agency” to the outcome of students taking ownership of their learning needs. Student agency has become an increasingly popular term and discussion topic during recent education conferences.

Adding to the definition of personalized learning

To augment the discussion, I’d like to point to the US Department of Education’s definition of personalized learning:

Personalized learning refers to instruction in which the pace of learning and the instructional approach are optimized for the needs of each learner. Learning objectives, instructional approaches, and instructional content (and its sequencing) may all vary based on learner needs. In addition, learning activities are meaningful and relevant to learners, driven by their interests, and often self-initiated.

Richard Culatta, CEO of the International Society for Technology in Education, expanded on this definition in his article What are you Talking About?! The Need for Common Language around Personalized Learningexplaining that individualized learning consists of “learning experiences in which the pace of learning is adjusted to meet the needs of individual students, focusing on the ‘when’ of personalized learning,” while differentiated learning is “learning experiences in which the approach or method of learning is adjusted to meet the needs of individual students, focusing on the ‘how’ of personalized learning.”

Personalized learning, then, envelopes both differentiated and individualized learning, and goes ever further with the elements of student agency.

(Next page: Individualization and student pace)

Individualization and students moving at their own pace

“Individualized learning” is an admittedly hard term for which to find a broadly acceptable definition. Culatta asserts that “in individualized learning, all students go through the same experience, but they move at their own pace.” I find that many others do not accept the “go through the same experience” aspect, but what is present here and what must be presented somewhere within the overall dialogue on personalized learning is a discussion of students moving at their own pace.

In my view, the truly new element of personalized learning is individualization, which for our purposes here I will treat as synonymous with competency- or mastery-based learning. Conversations around differentiation have been ongoing for years and although the term “student agency” is relatively new, the overall concept is older and found within other approaches (eg formative assessment). But, in individualization, we see something truly new.

Prior to these discussions, course counts, Carnegie units, graduation requirements, hours of contact and seat time have been at the center of our considerations. These were all remnants of the factory model. Now, conversations are shifting to “mastery,” “outcomes” and “competency-based learning.” School policies mandated seat time, but not mastery.

Adjusting the pace of learning to stress mastery, then, makes tremendous sense—focusing on mastery rather than seat time. But if we accept this, much of the structure of school as we know it unravels. The bell schedules and course counts go out the window and are replaced with mastery models, clear outcomes, flexible schedules and resources and redefined teacher roles.

This element, learning at a flexible pace that is driven by true mastery, must also be a part of our personalized learning conversations. When we get to this point, the depth of personalized learning becomes apparent. This is no minor or even moderate change to the status quo, this is a radical re-thinking of school as we know it.

That all said, despite the root word (“personal”) and references to/requirements of individual plans, we should acknowledge that many lauded examples of personalization also reference extensive group work on broad, multi-disciplinary projects and other cooperative activities. Given this, it is critical to understand that personalizing education certainly does not mean that things have to be completely personalized at every moment.

Additionally, in a school of any significant size, there would surely be groups of students all working on mastering similar skills. So, while the prevalence of lectures and other whole-class activities will decline under personalized learning, a teacher of 125 students would not truly be planning 125 different lessons.

Guided by additional clarity on the multiple elements included within personalized learning, we can begin to turn our attention to the real challenge–implementing at scale.

Latest posts by eSchool Media Contributors (see all)