Applications of AI in Learning and Development (Adaptive Learning) Part II

What is adaptive learning? It is learning that adapts to the learner’s learning level at every step. It is continuous and not discrete. It creates a unique learning path at each point. It reduces the actual learning time while increasing the quality of learning. The aim is to make learning a fun filled experience based on what an employee:

• Needs to learn

• Wants to learn

• And the most preferred medium to learn

  
  

The core foundation of an AI adaptive learning platform is flexibility. And that means, the possibility for learners to get personalized course materials and learning paths based on their learning ability. This is the major shift that advancement in AI and Machine learning algorithms have been able to bring. The Traditional approaches were less flexible as they offered the ability to modify course materials using a set of pre-defined rules.

Now before we dive deeper into the features and components of adaptive learning lets step back and see the Global training landscape to answer questions a) How much are we learning? b) What are we learning?

Global training spends in 2019 stood at $370 billion. ~85% of that goes towards technical or hard skills training and 15% towards Soft Skills. 

The hard skills would vary from industry to industry but the top 5 are not surprisingly technology related:

• Blockchain expertise knowing how to wrangle cryptocurrency (mining it, validating it, storing it, or moving it) not a super common but immensely valuable skill

  
  

• Cloud computing with everything is backed up in “the cloud” and with so many companies running their entire business it is one of the hottest skills going. 

  
  

• Analytical reasoning given the amount of data with companies today compared to a decade back there is a huge need of analysts who can take look at this data and help turn it into predictive analytics, or insights.

  
  

• Artificial intelligence according to Forbes there are 13 Industries that are going to be revolutionized by AI. https://www.forbes.com/sites/forbestechcouncil/2019/01/16/13-industries-soon-to-be- revolutionized-by-artificial-intelligence/#50ad14713dc1. Whether it’s business analysis, predictive algorithms, and metrics, or interacting with customers, AI is really the future of business. 

  
  

• UX design with so much of commerce and business conducted via online platforms. UX designers who can create good user experience are a hot commodity. 

Busy, Busy, Busy – is an apt way to describe an average worker today. So, when do you take the time out to learn? Where do you learn?  

With most of the employees preferring to learn at work, at their own pace and at the time of the need (68% of employees prefer to learn at work, 58% of employees prefer to learn at their own pace, 49% of employees prefer to learn at the point of need), Learning in the flow is the new mantra. It recognizes that for learning to really happen, it must fit and align itself to working days and lives. Rather than think of corporate learning as a destination, you go to, it’s something that comes to you. Using design thinking and AI & Cloud, you can build solutions and experiences which weave into your day jobs in such a seamless way that make learning almost invisible. Haven’t Google and YouTube done exactly that? They pioneered the space of flow learning and adaptability. Any time during the day you can reach out to Google to have your question answered, understand a concept or find what is the latest trend. Fifty-nine percent of Gen Z prefer You Tube as a learning tool. This tells you the popularity of Videos as a learning medium. The rich content that you Tube has makes it highly adaptive learning platform as you can pick and choose from multiple videos on the same topic. 

Here in the picture you see an instructor or coach training an employee on a new system. Based on her reactions, she will alter her instructions. The reactions could come in the form of a non-verbal cue like a facial expression or a hand movement, or could be verbal. At each point both her and her student are in sync. While this is the best model to upskill but not scalable or affordable. Adaptive Learning algorithms try to mimic the same experience for the student using machine learning, cloud computing and data analytics. As the system adapts to your style more and more the sense of being in the flow of learning increases.

Another interesting area that ALS (adaptive learning system) draws a lot of inspiration from is the research conducted in the field of Brain Science. It is involved in understanding how the mind works, from perception to learning, language, attention, memory, problem-solving, decision-making and judgment. Based on this research we know that every person’s brain has to assimilate, apply, retail and recall information. What differs is the timing of forgetting or retaining from one person to the other. If too much time passes between learning something and trying to remember it, chances are it will be forgotten.  

To maximize learning, spacing it out with bringing in practice exercises of applying the tools is the answer. The practice sessions get you to retrieve the learning. Retrieval practice involves recreating what you learnt a while back from your memory. This combination has been proven to increase retention by as high as 146% compared to the simple study method, as you can see in the graph. A good example of practice tool is uSpeek & kWurd for communication skills (www.uspeeknow.com & www.kwurd.com)

One of the other key components of an Adaptive Learning Platform is Recommender Systems. The biggest and most widely used example of the same is the Netflix algorithm. As most of you must know each viewer has a unique Netflix home screen. Based on your preferences, viewing history, time and whether you watch Netflix on mobile, TV or laptop, all these and many more factors are all pulled together by the algorithm to create your very own personalized home screen. 

Recommender systems are one of the most successful and widespread application of machine learning technologies in business. Machine learning algorithms in recommender systems are typically classified into two categories — content based and collaborative filtering methods although modern recommenders combine both approaches. Content based methods analyse the content of the items in order to devise a measure of similarity between items; collaborative filtering, where the similarity between items depends on user preferences, i.e., items that are liked by the same people are considered to be similar. So, you look for clusters of collaboration. 

Now that we have some semblance of what adaptive learning systems are and what comprises its core engine. The foundational requirements of ALS can be divided into 3 categories:

We can divide it into 3 categories:

• Content: the material to be learned, broken out into interconnected modules. For the recommender system to work well you need a huge variety both horizontally and vertically. This implies you need to cover a vast breadth of skills and within each skill you need to ensure you have material suitable for different proficiency levels.  

• Assessment: Continuous assessment: assessment embedded throughout the curriculum that creates data points about students, gives them feedback on their work, and points them toward new understanding.

Competencies: Competencies defined as units of knowledge, skills, and abilities used to evaluate skill mastery and the demonstrated application of those skills. The sub competences for each competency are mapped out in a parent and child relationship. For example, if the competency is Develop Customer Relationships and the Ability to communicate is a sub competency.

As you can guess by now setting up an Adaptive Learning System is a reasonably expensive task both from the investment needed in setting up the platform as well as the data collection and content creation.  So, why would one bother? To put it simply, because it delivers results. Why? Because personalization pays off. Think about how Netflix, Amazon, or Apple Music recommend content and songs that you’d be interested in based on what you’ve previously watched or listened to. The content is tailored to YOU. This and some other benefits of ALS are:

• Significant reduction in learning times – this has a huge business impact. Shorter learning times implies higher productivity, lesser errors both have a direct tangible $ impact to the bottom-line.

  
  

• Provides Deep Powerful Insights- Once you get the data collection right while setting up ALS the dividends of that effort will be reaped by the organization for decades to come. The reason being that this data has trickling upstream and downstream benefits. Knowing the correlation of skills, competencies and the knowledge level can increase your hiring quality tremendously. Better hiring mean lower attrition which means lesser cost of hiring. 

• Rich Analytics on Learners – The vast amount of learner data that the machine learning algorithm of the recommender system of your ALS will gather is a priceless treasure. You could mine it any which ways to help understand how you need to transform the systems and structure of your organization to produce better business results. For global organizations it will generate vast amount of information on cultural diversity and how different people learn differently across cultures

  
  

• Enhanced Employee Satisfaction: it’s an unfortunate but well-known fact that almost half of the workforce feels stressed at work today. Extensive Automation, digitization, changes in economic landscape, shorter shelve lives of skills and only a few amongst many other factors we are live with. Offering your employees, a truly nurturing fun filled and non-threatening environment to learn via an ALS can go a long way in building strong bonds. You can show them you genuinely care! 

  
  

The thing with adaptive learning systems is that they work for all age groups. Children, adults, you might be college student or a mid-career employee the principles of learning are the same.  Interestingly the advancement and investment in these learning systems has been higher for school and college education compared to corporate learning.  Here are top 5 ALS systems which you might want to look at:

• Smart Sparrow- Recently acquired by Pearson Education Smart Sparrow is known for creating personalized and beautifully stunning learning experiences through their unique content design. 

  
  

• Knewton- has such a powerful engine that they have made 23 partnerships with education companies and schools in just 7 years. Their learning analytics give them tremendous insights into each student which in turn helps them personalize the experience.

  
  

• DreamBox- This Software focusses primarily on Math from Kindergarten to 8th Grade. Dreambox dynamically adapts to the learner, providing personalized instruction.

  
  

• Fulcrum Labs- targeted at both students and Corporates they platform is based on adaptive 3.0 technology with the promises that it replicates one-on-one coaching and provides the learner data to predict, verify and increase application-level mastery.

  
  

• ScootPad- has over 2 MM users globally. Targeted at school children Unlike traditional tools, ScootPad's powerful hybrid engine delivers data-driven differentiation through automatic adaptivity and targeted teacher-driven intervention.

  
  

And It is not just the learning experience ALS amplifies. These platforms also provide dramatic efficiencies to learning development and content creation. 

Exciting? We believe so. Do write to us what you think in the comments section below. Like and subscribe. If you liked the video, keep watching this space for more on how AI will transform the way we learn.