Automated Educational Environments

Take a moment to immerse yourself in a simulated Automated Educational Environment (AEE).

Mock-up of the user experience withAugmented Educational Environments. Bonus points if you catch my inclusion geek-joke.

I hope you noticed that there was nothing particularly special about this environment, an empty lot in a suburban Chinese landscape. AEEs are available anywhere with a data connection. Not a pre-recorded set of learning outcomes, the AEE will rely on a combination of user input and mobile sensors to deliver information in the moment it is most relevant.


Technology Overview

Much of the necessary technology for this educational future is available in today’s mobile devices. Below is a brief overview of the necessary technology as it exists today with a focus on readily available consumer technology.

GPS — most mobile phones have this technology which enables the device to know where you are around the world and provide relevant information. Works best when paired with other technologies such as wifi networks or bluetooth beacons. Everyday uses: automated timezone changes.

Text to Speech — allows access to all text on the internet without taking your eyes off the world around you. Everyday uses: visually impaired web-browsing, turning any ebook into an audiobook.

Natural Language Interface — Apple popularized this with Siri and Google and Microsoft have responded with Now and Cortana, respectively. Everyday uses: voice commands and search on mobile OS.

Machine Learning — a rapidly growing branch of computer science. This is the technology that will make AEE a reliable educational experience, specifically computer vision.

Example of machine vision with Baidu Translate. Note the killer music-themed quilted placemat made by my mom.

Vision — in the intro video, the computer was able to identify the nest on top of the telephone pole. Computer vision is a branch of machine learning that enables computers to see and recognize objects in the real world. A great example of this is Baidu’s image search or translation services. Below is an image of my TV remote control in Baidu Translate. The first attempt came back with “an old mobile phone” but the second was more successful. Computer vision will make digital visual media searchable even when humans did not include relevant tags or titles for the image.

A current example of computer vision enabling more advanced search and interaction within previously unsearchable media is NoteVideo. NoteVideo makes blackboard lecture-style videos navigable through processing the video frames and picking out visual elements to enable searching via transcript or the elements themselves.

AR Glasses — the most famous example being the Google Glass project. A small computer and display mounted on your head enabling a heads up display (HUD) of information based on the environment around you. The 2015 CES trade show had little in the way of consumer-level headsets but they are set to invade the industrial sector with cheaper consumer models on the way. This technology would enable a portable AR experience with data overlayed on the environment. Microsoft’s HoloLens is a good, non-mobile example.

Automated Educational Environments take the analytical power in the worlds data centres and make it available to you through data connections to your mobile device. Many of you can try this right now by asking your phone to call a friend. Your voice command is processed outside of your phone with the resulting actions sent back to the device to be carried out. Currently Siri uses WolframAlpha to provide contextual answers to queries but these responses are limited to curated databases and like most computational knowledge programs today limited to fact-based answers.

Compare a WolframAlpha search to a Google search. The former tries to provide answers based on an understanding of the question and the latter provides relevant pages through keyword and popularity analysis. The necessary technology for AEE requires deep understanding of the search query combined with the users location and interests and the ability to return relevant information.

Educational Overview

In conventional education the student is required to adjust himself to an established curriculum; in adult education the curriculum is built around the student’s needs and interests. (Lindeman, 1926)

The most important part of all this technology is how it will influence the users’ educational experience. I predict initial uptake will be through adult education. Work place training may quickly resemble this style of learning which could be similar to the learning experience at the National Museum of Natural Science. Through use of indoor location technologies, the NMNS created a programmed learning space in which users could freely explore while receiving prompts and information on the objects around them. (Huang, T. et al. 2014) Since these sorts of programmed environments do not rely on advanced computer intelligence rather human labour, they can be created now. The difference between this and the proposed AEE is programmed by humans vs automated by computer. The experience within the programmed environment cannot be replicated outside of that designated space and these educational tools become dated as new developments occur.

Automated environments will rely on knowledge scrubbed from the ever-growing databases within the internet. This knowledge, once interpreted by computers will enable AEE to take advantage of the “teachable moment,” the moment in which you are best able to learn something. Learners exposed to a video on chopstick manufacture in the classroom, may remember it but if exposed while at a restaurant while pondering over the bin of bamboo sticks in front of them are more likely to retain the information. In order for AEE to function properly, the ability to access learning material within the moment without interruption is necessary. This is the missing link, the user interface.

Though much of the technology needed to provide an augmented learning reality exists, the information remains in our pocket. We are only able to obtain the answers we need with concentrated effort that interrupts the learning flow. Once the learner is able to access information through a method that does not interrupt current courses of action, the information will flow freely. The missing link is within the primitive stage of natural language understanding and generation. Computers are not yet able to deliver the conversational experience necessary.

AEE will perfectly fit the model of individual life-long learning most often associated with adult education. The “curriculum” is based around the events and experience within the learner’s daily life. Learners are able to learn what they want in the moment they need it. Mobile devices act as personal tutors, serving the needs of the learner. This individualized learning platform will enable users to explore their interests without interruption.

The mobile device as personal tutor is currently being developed for children under the CogniToys line. These toys use a data connection to IBM’s machine learning system Watson in order to provide feedback to vocalized questions. As the child ages and their level of questioning increases, so do the complexity of the answers. A database of the child’s learning will enable the toy to build upon prior knowledge and test understanding. The toys also provide a parent-oriented display of learner profiles for their children. As data mining techniques improve, more advanced learner profiles will arise. Teachers and parents will be able to see a child’s progress and compare it to any number of standards. The potential for educational advancement and hyper-analysis are huge.

The overall approach to AEEs is simply an advancement of the ease at which we are able to access data on the internet. It’s a natural evolution of our interactions with technology. From orator to print to digital, ideas and information have become easier to access by greater numbers of people. An AEE is the result of a more fluid integration of the world’s data with an interested party. In that sense an Automated Educational Environment will never exist, it will only be anywhere your mind wants to wander.


Works Cited

Huang, T. C., Chou, Y. W., Shu, Y., & Yeh, T. C. (2014). Activating natural science learning by augmented reality and indoor positioning technology. In Advanced technologies, embedded and multimedia for human-centric computing (pp. 229–238). Springer Netherlands.

Lindeman, E. C. (1926) The Meaning of Adult Education, New York: New Republic.