You are here:

Request a Demo Contact Us

A Guide to Machine Learning and L&D

Bob Little 

February 5, 2016

Machine learning has nothing to do with “learning” in a human sense, although its proponents have high hopes of, one day, being able to replicate “human learning” in machines. The concept originated with the pioneers of computing in the 1940s, such as Alan Turing, who began investigating the possibilities for artificial intelligence (AI).

While AI has been the theme of many thriller and science fiction films and television plays since, there have been relatively few advances in it in the real world. Yet what machines can do, a great deal faster and more efficiently than people, is sorting and analyzing vast amounts of data – including today’s explosion of “big data.”

Machine learning is based on algorithms that enable computers to “learn” from data without relying on rules-based programming. Statistical inference is a key foundation, derived from classical statistics theory, and developed between the 18th and early 20th centuries for much smaller data sets than the ones now available. With more data at its disposal, machine learning can make highly accurate predictions, as well as discover insights that human analysts may fail to spot. In recent years, for example, such analysis has added greatly to professional sport, and skill and performance levels have risen accordingly.

Increasingly, companies are using computer analysis of large amounts of data to identify and maintain their competitive advantage. At a consumer level, analyses of collective and individual consumer preferences lead to online shoppers being prompted to buy more by suggestions along the lines of, “People who bought this, also bought… ”

Description, Prediction and Prescription

There are three key stages to machine learning: description, prediction and prescription. The description stage involves collecting data in databases. Prediction involves forecasting future behavior or other outcomes by reviewing this data to see historical trends. (The quality of these predictions therefore depends on the quality of the data.)

The real value of machine learning comes from the third stage, prescription. Understanding why people might do something enables organizations to put in place strategies to encourage or deter that behavior.

The ability to prescribe is bringing in a new era of human–machine collaboration and demands a major change in the way organizations work. Machines identify patterns from data, while human beings interpret these patterns and recommend action. Machine learning is already an important tool for C-level executives, in terms of developing corporate strategy. But it is increasingly going to affect every aspect of an organization’s operations – including, or perhaps especially, L&D. As an L&D professional, you’ll recognize the need to “translate” the machine learning’s impersonal conclusions and predictions into “actionable insights” that people at all levels of an organization can support and implement.

L&D and Data Strategy

A “best practice” data strategy begins by identifying gaps in the data, determining the time and money required to fill those gaps, and breaking down internal silos to eradicate those gaps.

Unfortunately, it’s not uncommon for departments to hoard information and restrict access to it. So, some organizations have created the role of chief data officer (CDO) to overcome these internal obstacles. Other solutions include placing responsibility for generating data in the hands of frontline managers.

In the interests of promoting the well-being of your organization – rather than perpetuating divisive inter-departmental feuds that handicap its health and vitality – it’s important to foster worker buy-in to corporate strategy that’s driven by machine learning.

The L&D department can help a CDO by identifying small but easy-to-find successes – and making these successes known throughout the organization. Examples might include how sales training programs are helping to produce greater sales in the light of improved data about what customers are searching for on your organization’s website.

One of the key things that senior management look to L&D for is its role in generating appropriate behavioral change among the workforce. Increasingly, frontline managers, armed with data-analyzed insights from ever more powerful computers, will make decisions on their own, without the luxury of referring these decisions to more senior executives. Top management will set the overall direction and focus on the exceptions that arise.

This trend will involve L&D professionals in providing their organization’s frontline managers with the necessary analysis, problem-solving and decision-making skills, while senior executives set appropriate incentives to encourage the greater degree of data sharing that this policy necessitates.

For the foreseeable future at least, there’s no question that organizations will still need people – notably to decide the key questions that machine learning will address via data collection and analysis. Human beings will also have to learn how best to formulate corporate strategy in the light of this data’s predictions.

Moreover, just as people need regular assessments, appraisals and performance reviews related to their work, so they will need to regularly evaluate and refine the software related to machine learning. So, while L&D professionals are an established part of the appraisal and review regime for human beings in their organizations, they’ll likely also help to develop the knowledge and skills of those of their colleagues who appraise and review the organization’s machine learning.

The greatest threat to corporate – and even consumer – life as we know it is going to be when, backed by enormous statistical data banks, organizations substitute mere processes for the ability to think, act and respond intelligently. But, at present, that’s an issue that’s the preserve of science fiction – isn’t it?

Leave a Comment

Your email address will not be published. All fields required.

View our Privacy Policy.

One comment on “A Guide to Machine Learning and L&D”:

  1. jack paul wrote:

    There are three key stages to machine learning: description, prediction and prescription. The description stage involves collecting data in databases.