Deep Learning APIs

Metacog enables you to build data-rich learning products

automated open-ended scoring

The multiple-choice approach does not teach learners to think critically. It trains teachers to teach to the test and learners how to select an answer.

adaptive online curriculum

Better data can drive your chosen recommendation engines to pick a better “next step.”

dynamic interactive visualization

Metacog enables the data flood to be transformed into manageable abstractions. Inform learning guardians in real-time where learners are and what should they do next.

richer learning objects

Capture everything you think matters without risking overbuilding (or missing something).

edugames that prove their worth

Are you an educational game developer? How would you like direct proof of efficacy so you can quickly overcome the “prove it works and I’ll buy it” sales challenge?

detectors that go beyond standards

Do you know if the learners are bored or enthralled, gaming your system or pushing it to its limits, finishing tasks dutifully or achieving deep learning? Do you want to know?
The only way to know is to try it!

Envision creating intelligent products, more closely aligned with desired learning outcomes. All available via an easy-to-use API

sensors can be created with metacog simulations, eeg brain wave headsets, mobile tablets game controller and web pages

Metacog enables you to build products that escape Bloom’s basement

diagram showing cognitiv process dimension and knowledge dimension bloom

IMAGE CREDIT: Rex Heer, Iowa State University

Build Smart Learning Products

Invisible Integrated Assessment: Score open-ended performance tasks automatically

matrix of cognitive and affective IRR training set machine learning
Are your clients asking you for something better than multiple choice? Are they unsure how to teach and assess without stopping and testing?

PARCC, Smarter Balanced and Common Core standards enable much deeper assessment challenges to help college and career readiness and to develop complex 21st century problem-solving skills. NGSS also emphasizes evidence-based demonstration of procedural knowledge and cannot be practicably assessed utilizing multiple-choice automated scoring.

The problem is that there has been no ability (other than time-consuming hand-scoring) to score open ended challenges at scale to assess competencies from authentic learning challenges. Metacog’s technology tracks complex problem solving processes to learn what they can actually do—automatically.

Use data to inform better products: learn from your learners

diagram of virtuous feedback loop improving products through continuous data informing instruction
The learning industry has been very good at pushing finely crafted content at students but until now there has been no way of hearing back from learners at scale.

Learners generate rich data streams as they solve challenges. This data can be used to understand not only how they came to their final answer, but to assess the quality of the learning objects themselves and to more deeply understand where learners are and what they should be doing next.

As students solve learning challenges, metacog can get incredibly fine grained information about student performances: how they think, how they approach difficult problems, whether they try very hard, or give up easily. We believe this type of information can replace the activity of stopping the learning process for testing.

Build smart + connected products: Compete in the new data age

cluster diagram
Improve your data strategy, product development roadmap and competitiveness in the coming data-driven years. In the Internet of Things (IoT) era, companies who are able to harness data, connect their products together and create systems of systems to optimize their customer experience will be the clear winners.

With metacog you have a platform that can easily data-enable your products to both create and deliver optimal customer value.

Envision learning that much better prepares students for success in college, and beyond. 21st century jobs require 21st century learners and technologies. Metacog makes it easy to integrate advanced data capabilities into your new and existing products.
The only way to know is to try it!

Unleash the power of predictive analytics

Stacks Image 1370
Understand how your products are being used at the metacognitive level. Know what products will be in demand next year, and where can you focus building your next-generation products to maximize efficacy. Sell more into an increasingly challenging market.

Advanced Machine Learning

Stacks Image 1409
Advanced machine learning analytics let you understand how learners tackle challenges, not simply their final answers.

Utilize your data to uncover patterns in learner behavior to provide new services to inform instruction through dynamic visualizations and help learning guardians to target instruction and intervention in real-time.

Uncover if learners are lost, bored, gaming the system or challenged and engaged,
The only way to know is to try it!

Use Case: Next Generation Invisible Integrated Assessment

Stacks Image 2109
Testing and Learning Converge Invisibly

Metacog, an analytics platform that can integrate seamlessly and invisibly into any existing Interactive Learning Object (ILO), provides a powerful solution. The metacog platform enables the capture and analysis of data generated by student interaction with digital learning objects so that student’s metacognitive processes become measurable.

Metacog goes beyond grading a student’s final answers; it captures their trajectory of problem-solving throughout the assessment. Metacog integrates with ILOs via a series of APIs, enabling the deployment of different functionality in different circumstances. These APIs range from data ingest and retrieval, to an automated scoring engine, data visualization, and diagnostic and recommendation preprocessing engines.

The metacog scoring API is an integral part of the metacog system. It is designed to be used by content experts to develop automated evaluations of student performance in a digital learning object.

Serving as a student-evaluation tool, the scoring API processes data generated by a metacog-enabled ILO’s (Interactive Learning Objects) and captured by the Instrumentation Ingest API. These identify salient data in the base learning object, set “hooks” into the object for collecting real-time clickstream (telemetry) data, and then aggregate and output the student interaction data in a form that can be utilized and analyzed.
“Multiple choice wires your brain to pick an answer rather than have an original thought”
— Neil DeGrasse Tyson, Astrophysicist
The scoring engine (configured and trained by the scoring API) then uses the data collected to provide automated evaluation of student performance based on a scoring rubric that the content expert defines.

The resulting scoring model, informed by research and developed using educational data mining, is adaptable, automatic, and scalable to millions of students (and more). It also is built for rigor with human-human and human-machine inter-rater reliability. The scoring model can then be used within visualizations through the visualization API. The visualization can show important patterns in ways that are friendly to researchers as well as teachers and students.

This critical scoring information also feeds into the next services provided by the platform: diagnostic information and recommendations preprocessing that is sent to your favored recommendation engines for students. Now summative and formative testing can converge invisibly and go deeper than ever before.

Use Case: Competency Portfolios

To learn more about how to integrate metacog into your products, please visit our DEVELOPER SITE