In a 2018 report by the United States National Science & Technology Council (NSTC), the Committee on STEM Education shared the nation’s goals for American STEM education—building a stronger foundation for STEM literacy, inclusion, and diversity of STEM students and preparing the STEM workforce of the future.
Today, STEM problem development is a fully manual effort that is challenging to scale. With greater adoption of artificial intelligence (AI) being made in professional arenas, the reasoning for implementation of AI in education is becoming more important than ever and could be a bid to solve developmental challenges in STEM education.
A major challenge in STEM is the limited number of problems, exercises, and tasks it can take for students to master a skill. Because of the manual effort required, education providers are forced to outsource experts. This typically comes at a high cost and requires large amounts of time to produce a sufficient amount of STEM problems that cover the varying levels, diverse teaching styles, and learning objectives.
As this research is done individually, another element educators need to be increasingly aware of is ensuring all problems are similar in quality, which can be difficult to manage due to human factors and error. AI eliminates human error and reduces the time required for repetitive tasks, giving teachers more time to focus on more creative and targeted content to better suit individual student needs.
Two primary elements of AI-based learning
Addressing the challenges of STEM, there are two primary elements of AI-based learning that can enhance students’ and teachers’ curricula learning capabilities.
1. STEM problem development automation
STEM problem development automation is achieved by using deep neural networks and other machine learning methods for natural language processing (NLP). NLP models are able to train an AI-based machine to find and analyze the original STEM problem and automatically generate a similar one using the original exercise as a template.
This provides the ability to develop new, high-quality learning content (e.g., STEM problem sets for exam) quicker, at scale, and less expensively than a human expert-based approach.
When textbook tagging is automated, the process can be used to identify existing problems in textbooks or other source material and tie them to a learning objective associated with Bloom’s Taxonomy, which characterizes higher learning in six stages:
For example, the tagging engine encounters the problem “Find 45% of 120,” then analyzes it and identifies the learning objective as “Compute Basic Percentages.” This tagged problem becomes a model for generating similar problems.
2. Adaptive learning
Adaptive learning, also known as adaptive teaching, is an educational method in which an AI-machine delivers learning content tailored to the student’s objectives, rate of learning, and aptitude.
To adapt STEM content to a learner, the adaptive learning engine needs a variety of STEM problems to choose from by matching those with student’s needs. As previously mentioned, a sufficient variety of STEM problems can be achieved through automated development.
Through adaptive learning and development automation, educational providers can leverage tailored and scalable content to address students’ unique needs. With the incorporation of automated content, AI helps alleviate the manual time, money, and effort that previously fell on educators. This shift benefits both the educator and student with self-paced, adaptive learning pathways available at a lower cost and greater quality.
Challenges to STEM problem development automation
However, technology and challenges go hand in hand. These are some challenges to STEM problem development automation we face today.
- Ingesting Content for Machine Training at Scale – Much STEM content exists, but it has been produced in human-readable rather than machine-sensible formats. Therefore, large-scale content ingestion may be time-consuming and resource-intensive.
- Obtaining High-Quality Content – Content publishers seek to protect their content from use, and public-domain content may be poorly curated or moderated, and therefore may be difficult to trust. The adage of “Garbage In, Garbage Out” holds, and training machines with poor content will lead to poor outcomes.
- Machine Training Time is Long and Costly – Training machines take substantial expert time and significant elapsed time. As time is money, this challenge is one of intellectual and financial resource availability.
- Creating a pedagogical taxonomy – Supervised training uses samples of various pedagogical content knowledge taxonomies. Over time, we will create a universal taxonomy that would cover the learning objectives of the major textbook publishers.
- Developing Appropriate Datasets – For each topic in the taxonomy, we need an associated dataset. This ensures that STEM problems in the dataset will match the parameters in the taxonomy topic.
- Achieving a High Accuracy Rate – A high accuracy rate in STEM problem generation saves both expert training time and the time that experts would need to create problems manually. High error rates consume valuable resources and waste time.
These challenges and others have our focus and will ultimately yield to collegial partnering, ongoing development, and thorough testing.
The pandemic has accelerated the need for the adoption of digital tools in education. With the growing demand for advanced skill sets, educators can provide creative and more targeted learning rather than focusing on the repetitive tasks of creating problem sets. The net result is better STEM learning outcomes for more students.
As stated in the Committee on Stem Education of The National Science & Technology Council report, developing a STEM workforce will require involvement from an entire ecosystem—no longer is it the responsibility of private institutions and entrepreneurs.
AI is increasingly evolving into a community-driven initiative, with significant sharing and innovation through open-source AI software that integrates off-the-shelf tools, working with libraries and machine learning algorithms to improve server-side performance.
With these advances, both teachers and companies focused on STEM education can leverage automation to develop high-quality content more quickly, in greater quantity, and at a lower cost, all of which benefit the student.