Generations of Machine Learning

00:08 Introduction
00:36 When did the first generation of Machine Learning come about? What did it look like?
02:13 I imagine back then things were very different in the world of computational analysis. What was the main focus at the time?
03:23 How long did it take for the next generation to come about?
06:12 How did Machine Learning come about? What impact did it have?
07:58 What were the big improvements in technology and what were the big breakthroughs that really helped?
09:59 What was the impact of Deep Learning? What was the key difference between Deep Learning and standard Machine Learning?
14:36 What about the technology? What are the key advances which have contributed towards this advance in Machine Learning?
17:13 Certainly Machine Learning has gradually evolved over the last few decades. Can we already know what’s coming up next?
19:14 Often, what we see in the suply chain industry is that it is a few years behind the Big Four in terms of what they are focusing on. Where are they putting their money at the moment in terms of research?
22:43 To conclude, as a supply chain executive, it seems you are entering an environment that is constantly changing. Everything you implement will be outdated by design. How can you possibly keep up?


Machine learning is an umbrella term that includes diverse algorithmic approaches. In supply chain, the historical way of doing machine learning was time-series forecasting. However, this approach has been superseded by a series of superior forecasting approaches.

AI algorithms are being used to drive vehicles, translate languages and even identify cancers. As such they have captured the attention of the world’s media, but these advances in machine learning technology are not the overnight sensation that many are led to believe. In this episode of LokadTV, we go back right to the start; we investigate the gradual development of computational analysis over the last few decades and understand if this gradual evolution of the industry can give us any clues as to what the future may hold.

We discover what the first generation of computational statistical analysis, based on the parametric models, looked like. In addition, we explore how the 80s and 90s welcomed the arrival of non-parametric statistical models - which increased the chance of overfitting -, while the new millennium saw the arrival of Deep Learning, characterised by “hyper-parametric” models.

The increase in spending on research has had a real impact upon the computational capability. We discuss how something designed to support video games was actually revolutionary in terms of processing power. We also discuss the impact of technology such as the Cloud, which allowed us to store more data, more efficiently.

Finally, we talk more in depth about the key problems that the “Big Four” are currently focusing on. We debate why the supply chain industry is often a few years behind the latest advances in technology, and to wrap things up, we try to understand how an executive can keep up in an environment that is constantly changing.