top of page
  • alicia5593

Customized Learning at Scale

Optimizing for that cognitive load that is germane to long-term memory and enhanced understanding, by continuously monitoring their cognitive load at any given time.


Every individual learns differently. Learning medium, pace and even stress all contribute to how well a topic is understood and retained. This is no less true when it comes to first responders training for high-consequence/low-occurrence scenarios. In fact, its for these situations where it is most critical that individuals receive the training that maximizes their expertise and ability. In an ideal world, each first responder would receive personalized training, customized not just to their general learning preferences but tailored to their specific cognitive load at any given time in an exercise.

Our latest technology, Pan, does just that. Leveraging biometrics and realtime, spatial computing Pan is able to not only measure an individual’s cognitive load, but use that to dynamically change a training scenario such that the training experience is optimized for that person. This is hyper-personalization at scale.


When, where and how learning content is presented all plays a part in learning, as various factors compete for finite cognitive resources. Anyone who’s learned a new skill can likely relate to this (perhaps overly) simplified description. When a task is too hard, or a topic presented too far above your current level, it can be impossible to understand even the basics — the excessive cognitive load is detrimental to your learning. Similarly, topics present at a level in which you’re already well-versed, presented in too-easy a fashion, don’t elevate your understanding and might leave you susceptible to boredom or distraction — the lack of cognitive load is detrimental to learning. But just the right amount of cognitive load, on the right things, can lead to long-term retention and understanding.

“Pan is about optimizing for that cognitive load that is germane to long-term memory and enhanced understanding, by continuously monitoring their cognitive load at any given time,” describes Chief Scientist, Mik Bertolli. Whether its moving extraneous content from scenarios that are too difficult, or injecting complexity in those that are already mastered, Pan uses content generation to develop a training scenario in real-time, each and every session. “Our machine learning algorithms determine specific cognitive load and associated stressors with very high accuracy, almost 90%. And we’re able to do this on the timescale of human reactions, giving seamless experiences,” explains Bertolli.

Pan goes even farther, not just tailoring a given scenario in real time, but identifying and suggesting topics for training to enhance the overall capabilities of our responders. Consuming a broad spectrum of data, from macro contextual information such as population-level trends to data specific to a location or group of responders, Pan sees a fuller picture of training. This allows algorithms to identify not just where individuals or groups need more training, but what training scenarios might be most impactful for the community the responders serve. “This is an exciting application of AI that allows our first responders to anticipate training needs. For example, a fire department might get a jump-start on training for solar panel fires in areas where consumer adoption is picking up,” according to CEO Alicia Caputo. Pan delivers these insights to instructors as recommendations for training along with relevant content assets, where the instructors can then build exercises with Avrio’s unique content-authoring platform.

Importantly, instructors can override Pan’s choices, or enforce a pre-determined scenario trajectory at any time, whether in real time or when authoring the exercises. “We realize that trainers are the ultimate experts, and nothing will understand a department’s training needs more than the trainers. Our technology can help them and the training participants, but they remain in full control,” explains Caputo.

15 views0 comments

Comments


bottom of page