What is Learning Architecture? 

    In a Bersin (2016) study, p43, ‘Learning Architecture’ was outlined as a key component employed by the Top 10% of ‘High-Performing Learning Organizations’. Amongst the organizations described by Bersin as ‘High Impact’, they all shared a strong focus on their ‘Learning Architecture’. Other areas they appeared to emphasise were ‘Knowledge Management’, ‘Business Analytics’, ‘Rich Media, New Media’ and ‘Performance Consulting’.

    So, what is ‘learning architecture’ and how can it be used for advantage? Let’s take a look. The term was originally coined by the Advanced Distributed Learning Initiative (ADL). ‘Learning Architecture’ was later defined by Bersin as “an organization’s unique map of agreed-upon learning needs, learning strategies and delivery strategies for all of its training.”.

     Why is Learning Architecture important

    In short, an organization needs to carefully put in place their ‘learning architecture’ to optimise the learning impact on teams.

    To look at it another way, what is the danger of not having the correct ‘learning architecture’ in place? Training Industry warned, in 2019, of digital transformation actually being impeded by legacy technology tools and learning architecture that met past needs. The article further suggested that learning architecture ought to be focused on needs of the future. This stood for needs of both the organization and employees.

    Deloitte, 2016 recommended organizations “Adopt a learning architecture that supports continuous learning”.

    So, let’s take a closer look at how you might stay ahead of your organizational elearning technology needs.

    Learning architecture’s role in the 70/20/10 model

     In other discussions around what ‘learning architecture’ should deliver, Lea, et al refer to learning architecture’s role in supporting the design surrounding the ‘70/20/10’ model. What does this mean? Essentially, it is a suggestion that you have the correct tools, policies, support and resources in place. These should all be focused on enabling your audience to obtain required learning through each aspect of their 70/20/10 journey.

    70% experiential learning

    For example, L&D teams can support the learning that happens ‘on the job’ – the 70% component  – (otherwise known as ‘experiential’ or ‘informal’ learning) by ensuring that employees have easily accessible references. These might comprise ‘step-by-step’ guides or reference material to support individuals in their day-to-day roles and tasks. Your technology needs to match the needs of employees to access appropriate content. So, for example, your LCMS or LMS needs to be available for individuals to search for relevant learning resources at the point of need. In the instance of a LCMS, then this might be through the digital repository that it features.

    20% Social learning

    The 20% can also be thought of as ‘learning through others’. For example, such as in ‘communities of practice’ or colleagues (and/or a manager). L&D teams (or ‘learning architects’) can ensure they play an important part in facilitating this aspect through a variety of means. Elearning technology should be an ideal solution. Examples might include learning technology that provides a content sharing element, such as subject-specific (or general) online forums, etc. eXact learning has a knowledge-based platform called ‘con-X’ that provides a variety of functionality for knowledge-based informal learning. It should be possible with good learning technology platforms to integrate a 3rd party platform to enable ‘social learning’.

    10% Formal learning

    Obviously, this is where L&D professionals provide formal training to your audience through courses, modules and content chunks. Increasingly – particularly in 2020 – learning will have been delivered remotely via a technology platform. Owing to the pandemic, gathering for training in person in  workplace settings has been prevented. The use of remote learning technology has accelerated, to provide any training gaps where in-person training has been impossible. There has been much discussion and study of the vast expansion in remote learning tools in 2020 and how far they have met aims on a grand scale, globally. The ‘Learning Architect’ aims to ensure that the learning technology in place will be an integral part of  learning ‘design’. Naturally, learning technology will have formed a core part of the organization’s fundamental learning strategy and learning delivery.

    Does your learning architecture design meet future or past needs?

    In 2020, L&D teams may have quickly arrived at an answer to the earlier question of whether or not legacy technology met or impeded digital transformation needs. This is why a focus on learning architecture is relevant now.

    To analyse your learning architecture another way, you might examine it from a different perspective to the 70/20/10 model.

    Lea et al suggest that the ‘learning architect’ should consider three aspects of learning:

    I. “Live learning & tools”

     This is the equivalent of the ‘20%’ of ‘informal learning’, referring to social tools and learning from others. Are your ‘learning architects’, providing the learner with sufficient support via platforms and networks (online and/or offline) to enrich a learner’s knowledge? Does it enable your audience’s growth?

     II. “Learning journeys”

    There needs to be a connection and an ability to track growth from the beginning and beyond of a learning journey. In other words, you (and the learner) need to be able to identify a ‘way forward’ for knowledge acquisition. The direction needs to be clearly sign-posted for an individual in their pursuit of growth. For instance, do you have in place:

    – awareness campaigns;

    – formal training;

    – forms of review (such as assessments, feedback loops, review periods);

    – opportunities for refreshers; and

    – supporting networks (including technology and eg, interactive communities of practice)?

    Where you might be in a highly-regulated industry, then you may need to ensure that you have highly-robust review structures in place, to support any regulatory obligations on industry-specific training. For example, technology tools to provide data on completion and pass/fail records, including training dates, and repeat training, etc.

    III. “Learning content”

     This component focuses largely on content that is created to achieve a specific learning outcome. It may comprise many ‘learning objects’ that fit together in a microlearning or module. As a learning architect you need to ensure that the ‘authoring’ support tools (via technology) are in place to enable subject matter experts to easily create the learning message to achieve the end goal. Consider also, what other support structures need to be in place to complete the content’s journey to your audience. IE, where will you store legacy content? How can you ensure that users revise and review it? How will you track the production process of content and, then, deliver it?  Here you will need to think about things such as a digital repository, project management tools,  and delivery mechanisms.

    While considering these aspects, keep in mind the warning about ensuring that your technology is fit for future needs. You need to be sure that your technology is not focused on legacy processes from the past. Ensure that your platform(s) will meet your needs beyond the short-term future.

    Deep Learning Architecture

    Here seems an appropriate place to consider technology needs of the future. Also think about these needs combined with the opportunities that ‘Deep Learning’ can bring to the learning world.

    ‘Deep Learning’ itself has been defined by Opperman, 2019, as:

    a subset of Machine Learning, which…is a subset of Artificial Intelligence.”

    Machine Learning and Artificial Intelligence both have some very familiar uses. They have been employed by some of the ‘Tech giants’, such as Google and Facebook, to provide services like facial recognition and intelligent assistants – ie, ‘Cortana’.

    We have previously explored this topic and have written elsewhere about the emergence in 2011 of Neural Machine Translation (NMT). As we have previously described, Machine Learning or ‘NMT’  “utilises ‘deep-learning’ techniques to constantly learn from the ecosystem created”.  Opperman, 2019further defines the abilities of Artificial Intelligence (AI) to “sense, reason, act and adapt”.

    Ultimately, in ‘Deep Learning’, the more amount of data the system uses, the better the accuracy of the output.

    IBM provide this excellent table of examples of outcomes that can be achieved with Deep Learning Architecture:

    Source: Jones, T. ‘Deep learning architectures, The rise of artificial intelligence’, 2019, IBM

    Some examples of Deep Learning

    If we specifically look at Learning, these transformative applications present us with a great number of possibilities.

    Automatic translations

    eXact learning has worked extensively in integrating Deep Learning Architecture with its LCMS technology to provide automatic translations. This can help organizations roll-out consistent global learning to all of its audience. AI can support delivering learning via multiple languages and character-sets. Advantages include the much-reduced production time, with high-levels of content accuracy.

    Time and cost savings post-integration are significant. As we have previously reported, cost savings can be between 30% and 50%, with time savings of up to 80% in our Case Study, along with impressive improvements in accuracy.

     Biometrics to authenticate learners

     We have recently written about the use that biometrics can have  for L&D teams in providing compliance training. This is particularly pertinent in highly-regulated industries, where training can be a regulatory obligation. Biometrics – that is an extra safeguard in ‘authenticating’ who is actually taking the training – can provide an extra ‘safeguard’. There are a number of ways that you might integrate Biometrics into your learning technology as part of your ‘Deep Learning Architecture’. By way of example, this could be via facial, voice or fingerprint methods.

    Keep looking to the future

    There are many other ‘Deep Learning’ techniques that can form a part of your ‘learning architecture’. However, as discussed, it is vital to keep an eye on how your existing and planned technology, alongside your learning support structures, will need to deliver in the future. You need to keep it aligned to strategic aspirations to ensure that you make your learning architecture fit to that – not the other way around.

    If you have any questions about this topic, or if you would like to know more about how eXact learning supports organizations, including with AI and automatic translations  with our award-winning, scalable, Learning Content Management System (LCMS)  or would like a copy of our E-book on How AI helps learning then do get in touch.