Learning about Machine Learning

Across a series of blogs, with thanks to research from Apps for Good Volunteer Researcher Federica Luciolli*, we will explore some of the important topics surrounding the artificial intelligence and machine learning discourse. These will include the ML and AI education landscape and how this technology can be used for good.

Later this year Apps for Good, in collaboration with SAP, will implement machine learning resources on our platform. These will be targeted to Fellows, teachers and older students. The resources will enable an understanding of the topic and the potential problems posed by new technologies.


Defining Artificial Intelligence

The phrase artificial intelligence (AI) is generally used to refer to any sort of machine learning program. But some people prefer to reserve the phrase AI for the narrowly defined thing that can replicate many aspects of human intelligence, and become an entity in its own right. We haven’t reached that stage, yet. However, according to predictions by Zuckerberg, Machine learning-powered artificial intelligence will match and exceed human capabilities in the areas of computer vision and speech recognition within five to 10 years.

Defining Machine Learning

At its most basic level, machine learning refers to any type of computer program that can “learn” by itself without having to be explicitly programmed by a human.

 Screen Shot 2017-10-03 at 15.49.55

Today, machine learning is a widely used term which encompasses many types of programmes which are present across big data analytics and data mining. Machine learning algorithms are currently used to power the most predictive programs; spam filters, product recommenders, and fraud detectors.

Data scientists can program machine learning algorithms using a range of technologies and languages, including Java, Python and Scala, amongst others. They can also use pre-built machine learning frameworks to accelerate the process.

Machine learning can be categorised into three types: supervised, unsupervised or reinforcement learning. This is essentially the way the machine learns; whether the data scientist is involved in teaching the algorithm about what conclusions to draw.

The below diagram shows the basic differences between the three:

Screen Shot 2017-10-03 at 15.50.29


The Evolution of AI

Artificial Intelligence has been around since the 1950s, and has gone through many cycles. It is currently going through a new peak and is here to stay.

There are two key factors enabling the present growth of AI:

  • Unlimited access to computing power: Public cloud computing was estimated to reach almost US$70 billion in 2015 worldwide; data storage has also become abundant.
  • Growth in big data: Global data has seen a compound annual growth rate (CAGR) of more than 50 percent since 2010 as more of the devices around us have become connected. This growth of data is constantly contributing to improvements in AI.

Screen Shot 2017-10-03 at 15.50.09 1

Today, AI works in the following ways:

  • Automated intelligence: Automation of manual/cognitive and routine/non routine tasks
  • Assisted intelligence: Helping people to perform tasks faster and better (e.g. in cars the GPS navigation program that offers directions to drivers and adjusts to road conditions)
  • Augmented intelligence: Helping people to make better decisions (e.g. the combination of programs that organise cars in ride-sharing services enabling businesses that could not otherwise exist)
  • Autonomous intelligence: Automating decision making processes without human intervention. (e.g. self-driving vehicles)

What will be the economic impact of AI?

Currently London is Europe’s largest AI start-up hub. Below is detail of the best funded European AI companies and the industry they serve:

Screen Shot 2017-10-03 at 15.53.10


According to PwC research carried out for this report, global GDP could be up to 14% higher in 2030 as a result of AI. This is the equivalent of an additional $15.7 trillion. Thus making it the biggest commercial opportunity in today’s fast changing economy.

The economic impact of AI will be driven by:

  • Productivity gains from businesses automating processes (including use of robots and autonomous vehicles)
  • Productivity gains from businesses augmenting their existing labour force with AI technologies (assisted and augmented intelligence)
  • Increased consumer demand resulting from the availability of personalised and/or higher-quality AI-enhanced products and services

 The greatest gains from AI are likely to be in China (boost of up to 26% GDP in 2030) and North America (potential 14% boost).

The biggest sector gains will be in retail, financial services and healthcare as AI increases productivity, product quality and consumption.

 Screen Shot 2017-10-03 at 15.54.11

The AI Education Landscape

Education is often seen as an area with big AI promise, in particular regarding the personalisation potential of education.

Machine learning technology means acquiring a new set of skills; particularly the ability to successfully interact with the machine. This will be key for the machine to learn from the human and vice versa. The skills required by jobs in AI are different to the traditional computer scientist for whom demand will already outstrip supply by 50% in 2018.

Upskilling employees and students is of utmost importance to keep up with the demand for these in specialised skills. Educational institutions are already starting to offer some AI training:

There is still a long way to go in order to ensure the skills gap is not as huge as 50% supply versus demand. Teachers also require the resources (which are not necessarily available in the traditional education model) in order to effectively train the next generation of machine learning engineers and researchers.

The below chart maps some of the education tools currently available in the field and the age ranges it is available for.

Screen Shot 2017-10-03 at 15.55.15.png

What is to be expected from AI in the near future?

According to the document produced following this year’s AI for good global summit here are some predictions for AI use in 2018:

Screen Shot 2017-10-03 at 15.56.55

AI for Good

The summit explored the huge potential AI has to solve some of the world’s most complex problems. A broad range of examples of AI for good, especially in agriculture and health, exist already. These positive uses for AI however don’t come without their obstacles and challenges. Subject matters which cause anxiety around AI include; the ethical issues, security risks and the disruptive impact of AI on employment (more than 60% of jobs will be fully automated in the near future). We will explore this in much more detail in the next blog post in the artificial intelligence and machine learning series.

*Federica is a freelance consultant specialising in market research, partnership strategy and intrapreneurship. She is passionate about enabling organisations to innovate and collaborate to grow their social impact. In her 10-years experience she has worked with a broad range of organisations, from start-ups to multinationals, in the private sector and social sector.