Categories
Data, Automation & AI Featured

AI and the Classroom: Machine Learning in Education

Situation

For years schooling has been typified by its aspect of the physical grind on the part of both students and their teachers: teachers cull and prepare educational materials, manually grade students’ homework, and provide feedback to the students (and the students’ parents) on their learning progress. They may be burdened with an unmanageable number of students, or a wide gulf of varying student learning levels and capabilities in one classroom. Students, on the other hand, have generally been pushed through a “one-size-fits-all” gauntlet of learning, not personalized to their abilities, needs, or learning context. I’m always reminded by this quote by world-renowned education and creativity expert Sir Ken Robinson:

“Why is there this assumption that we should educate children simply according to how old they are? It’s almost as if the most important thing that children have in common is their date of manufacture.”

But as the contemporary classroom has become more and more digitized, we’ve seen recent advances in AI and machine learning that are closing in on being able to finally address historical “hand-wrought” challenges – by not only collecting and analyzing data that students generate (such as e-learning log files) when they interact with digital learning systems, but by pulling in large swaths of data from other areas including demographic data of students, educator demographic and performance data, admissions and registration info, human resources information, and so forth.

Quick Review: What is Machine Learning?

Machine learning is a method of data analysis that automates analytical model building. Using algorithms that iteratively learn from data, machine learning allows computers to find hidden insights without being explicitly programmed where to look Machine learning works especially well prediction and estimation when the following are true:

-The inputs are well understood. (You have a pretty good idea of what is important but not how to combine them.)
-The output is well understood. (You know what you are trying to model.)
-Experience is available. (You have plenty of examples to train the data.)

The crucible of machine learning consists of capturing and maintaining a rich set of data, and bringing about the serendipitous state of knowledge discovery: the process of parsing through the deluge of Big Data, identifying meaningful patterns within it, and transforming it into a structured knowledge base for future use. As long as the data flows, its application is endless, and we already see it everywhere, from Facebook algorithms to self-driving cars. Today, let’s examine machine learning and its implementation in the field of Education.

Application of Machine Learning in Education

Prediction

A few years ago, Sotiris Kotsiantis, mathematics professor at the University of Patras, Greece presented a novel case study describing the emerging field of educational data mining, where he explored using students’ key demographic characteristic data and grading data in a small number of written assignments as the data set for a machine learning regression method that can be used to predict a student’s future performance.

In a similar vein, GovHack, Australia’s largest open government and open data hackathon included several projects in the education space, including a project that aims to develop a prediction model that can be used by educators, schools, and policy makers to predict the risk of a student to drop out of school.

Springboarding from these two examples, IBM’s Chalapathy Neti recently shared IBM’s vision of Smart Classrooms: cloud-based learning systems that can help teachers identify students who are most at risk of dropping out, why they are struggling, as well as provide insight into the interventions needed to overcome their learning challenges:

The system could also couple a student’s goals and interests with data on their learning styles so that teachers can determine what type of content to give the student, and the best way to present it. Imagine an eighth grader who dreams of working in finance but struggles with quadratic and linear equations. The teacher would use this cognitive system to find out the students learning style and develop a plan that addresses their knowledge gaps.

Process efficiency: Scheduling, grading, organization

Elsewhere, several Machine Learning for Education ICML (international machine learning conference) workshops have explored novel machine learning applications designed to benefit the education community, such as:

-Learning analytics that build statistical models of student knowledge to provide computerized and personalized feedback on learning the students’ progress and their instructors
-Content analytics that organize and optimize content items like assessments, textbook sections, lecture videos, etc.
-Scheduling algorithms that search for an optimal and adapted teaching policy that helps students learn more efficiently
-Grading systems that assess and score student responses to assessments and computer assignments at large scale, either automatically or via peer grading
-Cognitive psychology, where data mining is becoming a powerful tool to validate the theories developed in cognitive science and facilitate the development of new theories to improve the learning process and knowledge retention
-Active learning and experimental design, which adaptively select assessments and other learning resources for each student individually to enhance learning efficiency

Existing Platforms

Recently, digital education venture capitalist Tom Vander Ark shared 8 different areas where leading-edge platforms are already leveraging machine learning in education:

1. Content analytics that organize and optimize content modules:
a. Gooru , IBM Watson Content Analytics

2. Learning analytics that track student knowledge and recommend next steps:
a. Adaptive learning systems: DreamBox, ALEKS, Reasoning Mind, Knewton
b. Game-based learning: ST Math, Mangahigh

3. Dynamic scheduling matches students that need help with teachers that have time:
a. NewClassrooms uses learning analytics to schedule personalized math learning experiences.

Grading systems that assess and score student responses to assessments and computer assignments at large scale, either automatically or via peer grading:
a. Pearson’s WriteToLearn and Turnitin’s Lightside can score essays and detect plagiarism.

5. Process intelligence tools analyze large amounts of structured and unstructured data, visualize workflows and identifying new opportunities:
a. BrightBytes Clarity reviews research and best practices, creates evidence-based frameworks, and provides a strength gap analysis.
b. Enterprise Resource Planning (ERP) systems like Jenzabar and IBM SPSS helps HigherEd institutions predict enrollment, improve financial aid, boost retention, and enhancing campus security.

6. Matching teachers and schools:
a. MyEdMatch and TeacherMatch are eHarmony for schools.

7. Predictive analytics and data mining to learn from expertise to:
a. Map patterns of expert teachers
b. Improve learning, retention, and application.

8. Lots of back office stuff:
a. EDULOG does school bus scheduling
b. Evolution , DietMaster.

Reflection

As the modern classroom becomes more and more digitized, we are able to gather myriad sets of data. The trick is, of course, being able to purpose it. The prize at heart of machine learning is knowledge discovery, the process of parsing through the deluge of Big Data, identifying meaningful patterns within it, and transforming it into a structured knowledge base for future use. In this article, we’ve seen examples utilizing machine learning in the education sector for prediction, scheduling, grading, and organization. We’ve also listed existing education-related platforms that use a machine learning component.

What does it mean to me?

Big Data have swept into every industry and business function and are now an important factor in production, alongside labor and capital. In a decision making system, the bigger the data, the higher the likelihood is of making good decisions. The time is now for organizations, in education or otherwise, to research how a cost-efficient machine learning component can transform your operational output. For more information, Check out this detailed guide by Jesse Miller on the amazing benefits of technology in the classroom and suggestions on ways to incorporate technology in the classroom.

“Parents are continually exposed to new technology via their children. Whether it be iPad App usage tricks, to the advent of robotics competitions, and perhaps now “new ways of thinking” as a result of interaction with Machine Learning based educational environments. Siloed educational content may give way to a topology of learning experiences.” O. Liam Wright – CEO, True Interaction

True Interaction produces custom full-stack end-to-end technology solutions across web, desktop and mobile, integrating multiple data sources to create a customized data solution. True Interaction can determine the most optimal means to achieve operational perfection, devising and implementing the right tech stack to fit the specific school and or district need. True Interaction pulls together disparate data sources, fuses together disconnected silos, and does exactly what it takes for school data systems to operate with high levels of efficiency and efficacy, ultimately leading to improved student achievement outcomes.

Categories
Insights

Big Data: Trends in the Education Sector

Our recent blog article highlighted 6 things to keep in mind when optimizing Big Data for your company. An essential component highlighted by Michael Davison, co-founder and Editor-in-chief of True Interaction, is the understanding that data analytics provides AN answer, rather than THE answer.

This concept resonates in the education sector, with the U.S. Department of Education calling the use of student data systems to improve education a “national priority.” Teachers are inundated with data points that quantify formative assessment results, parent call logs, absences, time-on-task, observations and more. But as with any sector, what matters most in education is how you use data, rather than that you have it.

Pasi Sahlberg, a Finnish educator, author and scholar, wrote:

“Despite all this new information and benefits that come with it, there are clear handicaps in how big data has been used in education reforms. In fact, pundits and policymakers often forget that Big data, at best, only reveals correlations between variables in education, not causality.”

Despite pronouncements such as this from key education reformers, in the ed-tech industry, big data and analytics are exceedingly prolific. A multitude of companies collect and analyze information on how students interact with digital content.

In the United States, almost all teachers (93 percent) use some form of digital tool to guide instruction. But more than two-thirds of teachers (67 percent) say they are not fully satisfied with the effectiveness of the data or the tools for working with data that they have access to (Gates Foundation, 2015).

The majority of school districts (70 percent) report having had an electronic student information system providing access to enrollment and attendance data for six or more years, and more recently, districts have begun acquiring electronic data systems:

– 79% report having an assessment system that organizes and analyzes benchmark assessment data

– 77% report having a data warehouse that provides access to current and historical data on students

– 64% report having an instructional or curriculum management system

Despite this proliferation in the use of data systems, the Gates Foundation found that there are additional barriers that prevent full implementation of data systems in schools.

Because of these barriers, teachers say data are often “siloed” and difficult to work with, inflexible, and unable to track student progress over time. How can school and district leaders optimize the data systems to impact student performance?

1. Involve teachers in data analysis.

Often, teachers are seen as the data “collectors” while school-based and district leaders are tasked with analysis, synthesis, and recommendations. This might contribute to the slow modification of classroom practices in response to data. With teachers left out of the data analysis, there is a significant barrier to classroom integration. It is important to recognize that in terms of promoting student growth, teachers know best the strategies and methods to employ. The disconnect between teachers and the top-down approach to data use in schools has created a false narrative that teachers are unmotivated and disinterested in employing data-driven instruction in their classrooms. In reality, 78 percent of teachers believe that data can validate where their students are and where they can go. District leaders and product developers can harness this desire to integrate data to provide customized solutions, tailored to grade level, content area, and demographics.

2. Invest in professional development of staff to integrate tools and practice.

Similar to our first recommendation, it is important that education leaders invest resources – financial, human capital, and time – to the development of teachers’ capacity to fully utilize and customize lessons based on student data.

Various research studies have found that those who participate in professional development programs that include coaching/mentoring are more likely to deploy new instructional strategies in the classroom. Effective professional development that truly enables teachers to integrate data systems is continuous and ongoing. Discrete training sessions to show teachers how to use specific hardware/software tools are important, but truly integrative professional development should go further by providing ongoing support and on-the-job training on how to collect/analyze data and how to adjust teaching in response to data analytics. Fishman (2006) noted that learning how to use technology is not the same as learning how to teach with technology.

3. Promote the use of personalized learning.

69 percent of teachers surveyed by the Gates Foundation believe that improving student achievement depends on tailoring instruction to meet individual students’ needs. Connecting data from multiple sources across a student’s academic, social, behavioral, and emotional experiences may help teachers gain a fuller picture of each student. Schools that adopt a learner‐centered pedagogy tend to experience greater integration and more effective use of technology in the classroom.

4. Work with product developers or vendors who can conduct full analysis of teachers’ needs when designing or optimizing data systems.

Over 60 percent of districts reported that lack of interoperability across data systems was a barrier to expanded use of data-driven decision making. True Interaction produces custom full-stack end-to-end technology solutions across web, desktop and mobile, integrating multiple data sources to create a customized data solution. True Interaction can determine the most optimal means to achieve operational perfection, devising and implementing the right tech stack to fit the specific school and or district need. True Interaction pulls together disparate data sources, fuses together disconnected silos, and does exactly what it takes for school data systems to operate with high levels of efficiency and efficacy, ultimately leading to improved student achievement outcomes.
Contact our team to learn more about how we can optimize your school or district data system.

Joe Sticca, Chief Operating Officer of True Interaction, contributed to this post.

by Jessica Beidelman