Deep Learning APIs

What exactly is metacog?

API Diagram

Whenever I meet anyone at a technology or learning conference - and they ask what exactly metacog is, and what it does - I do have to admit I always hesitate a little bit and am always tempted to ask “What do you want it to do?” because the platform is extremely multifaceted and versatile. Instead of being tongue in cheek, I will summarize the basic metacog API services here, which is virtually impossible to do in brief casual conversation. So here goes:

What is it?
Metacog is a cloud-based SDK toolkit and data storage and analytics platform for adaptive and competency-based learning. The toolkit enables the capture and analysis of data generated by learner interaction with digital learning objects so that their Metacognitive processes become measurable: That is, the system sees not only how the students score, but also how they are thinking about the assessment. The large data storage system is built on a secure and PII free cloud based real-time analytics platform. Access to metacog and our clients data is delivered via a series of APIs to enable the deployment of particular parts of the system in different circumstances. These capabilities enable you to gather whatever data you want from any source learning objects you need it from. You can store as much data as you like, forever in the cloud. You can learn from this data by querying against it, visualizing it, to enable new levels of understanding on how your products are being used so as to continuously improve them for optimal proof of efficacy.

Problems metacog solves:
• Authentic assessment doesn't scale easily due to the need for human scoring.
• At the same time, personalized learning needs authenticity
• It is difficult to provide evidence that educational games are actually teaching.
• Motivational systems award badges based on superficial data, rewarding students who get right answers easily, and therefore fail to have a data-driven mechanism for rewarding struggling students who are working hard.
• Product developers need deeper data in order to refine their products.

Instrumentation Ingest API – This part of the toolkit can be used to “retro-fit” learning objects with metacog capabilities, or to apply metacog capabilities to learning objects as they are developed.
Data Retrieval API – This part of the toolkit determines what data you want to capture and how you want to capture it.
Scoring Service API – This part of the toolkit enables the system to provide scoring based on the data being captured and analyzed.
Visualization Merge API – This part of the toolkit enables the system to be applied to complex visual learning objects such as simulations that entail integration with visual elements.
Diagnostic API – This part of the toolkit enables the system to be customized and applied to make specific diagnoses of student thought processes in real time.
Recommendation API - This part of the toolkit is used in particular to facilitate adaptive learning, using the data generated by student interactions with learning objects in order to offer the student recommendations of other objects or tasks within the assessment system being enhanced by metacog, based on whether the student is struggling with the current task or finding it too easy.
Micro-credentialing/badge issuance API – Unlike traditional courses (an all-or-nothing sequential set of learning experiences), competency-based learning and project-based learning align with a just-in-time/just- what’s needed learning model. It’s still important to issue micro-credits (badges) for mastery at critical junctures in learning. As Metacog accumulates evidence of mastery, this API will manage the issuance of badges – and provide the audit trail should any accreditor want to understand why a badge was issued.

• Enables the collection and analysis of data that has previously been inaccessible or accessible only by human observation and scoring – reducing or eliminating the need for human scoring enables vastly increased scale of deployment of complex learning objects.
• The modular, API-driven system enables the “retro-fit” of existing learning objects, preserving the publisher’s sunk investment in those learning objects.
• Many learning objects such as games are not designed to be, nor should they be, rigorous assessment instruments; however, it is still highly desirable that student interactions with such objects generate evidence that learning is taking place: metacog creates a framework by which that evidence can be gathered, analyzed, and acted on.
• Knowing not just how students score, but how they think as they interact with learning objects will be of tremendous value to product developers as they refine their products and create new ones.
• The modular, flexible system enables publishers, professors, and other to create their own content at a level of technological sophistication that would be inaccessible to almost all subject-area and pedagogical experts without a toolkit like metacog.
• metacog radically enhances the nature and the value of reports about student learning and does so using an intuitive “dashboard” approach.

Hope this helps!

Owen Lawlor