Determining what we want from artificial intelligence

Determining what we want from Artificial Intelligence

Successful artificial intelligence modeling

Before tackling artificial intelligence; let's first talk about Thomas Edison, who is said to have tried 10,000 times before succeeding in creating a light bulb filament; but he rejected any talk of failure. "I didn't fail"; he corrected his detractors. "I just found 10,000 ways that don't work."

It seems we like hearing about mistakes more than celebrating success. Although artificial intelligence technologies have transformed our lives; at home and at work; many recent media reports focus on the failures of smart devices; from disappointing gadgets exhibited at CES to faulty hotel robots. Some of the stories are very funny; but all they tell us; is that the technology is still developing and some products are better designed than others.

Using artificial intelligence

The Wall Street Journal writes about a guest in a robot-equipped hotel in Japan who was woken up every few hours by the room assistant asking him to repeat his order. The hotel manager finally realized that the guest's loud snoring had triggered the robot's voice recognition system.

For every innovation, however, there are success stories. Even when a machine achieves something as mundane as winning board games; it can be a harbinger of transformative benefits. For example, a chess-playing program called AlphaZero; developed by DeepMind; an artificial intelligence research company owned by Alphabet (Google's parent company); has made significant progress.

The chess game

AlphaZero has developed a new style of chess that is much closer to human improvisation than traditional computer chess. That's because AlphaZero learns from its past successes and mistakes; rather than calculating millions of possible permutations as it plays. According to Wikipedia, AlphaZero searches for 80,000 chess positions per second, compared with 70 million for the Stockfish chess engine.

AlphaZero uses (deep) neural network technology; sometimes referred to as deep learning; which has resulted over the last decade in significant improvements in machine learning. As computing power has increased, deep neural networks have produced machines capable of performing tasks in a way that would not have been possible with traditional programming techniques.

Indeed, it has transformed technologies such as computer vision and natural language processing (NLP); which are now deployed on a large scale in many different products and services. Manufacturing; healthcare and finance are just some of the sectors using deep learning to discover new patterns, make predictions and guide decision-making.

Positive impact in several areas

"In the field of intelligent manufacturing, AI can help streamline efficiency," says Wael Diab; who leads international standardization work in this area. "It can help provide information on where improvements can occur and; more importantly; it can provide information on the direction a particular organization may want in terms of production planning."

Sales of industrial robots have doubled over the past five years, according to the International Federation of Robotics. The IFR forecasts that by 2021, the annual number of robots supplied to factories worldwide will reach around 630,000 units. Industrial robots meet a real need.

In contrast, consumer electronics is still focused on the novelty value of gadgets. To a large extent; this is because we haven't quite figured out how we intend to use AI-enabled devices; in our daily lives or what we expect from them.

Korea Joongang Daily reported in October that Koreans are not only using their smart speakers to change TV channels; but also to discuss their feelings. In homes, 15% of staggering things to smart assistants seemed to be attempts at conversation; including "I'm bored" and "I'm sad". The newspaper notes a similar trend in hotel rooms, where more than 18% of the commands were attempts at conversation. Joongang Daily acquired the data from KT Corporation, the country's largest telephone company.

Trend towards standardization in artificial intelligence

In 2017, IEC and ISO became the first international standards development organizations (SDOs) to set up an expert group to carry out standardization activities for artificial intelligence. Subcommittee (SC) 42 is part of the joint ISO / IEC technical committee JTC 1.

SC 42 works with other JTC 1 subcommittees, such as those dealing with the Internet of Things, IT security and IT governance, as well as with the IEC Systems Committee (SyC) for Smart Cities. SC 42 has set up a core standards working group to provide a common framework and vocabulary. Several study groups have been set up to examine computing approaches and AI system characteristics, reliability, use cases and applications and megadata.

IEC standards play a key role in the transition to the fourth industrial revolution. IEC TC 65, for example, carries out important work relating to the measurement, control and automation of industrial processes.

"We look at the different components that go into AI, from the IT side to the ethical side. Having standards allows a common language and way for different stakeholders to interact," Diab explains.

"This leads to the ability to innovate in addition to the standards widely adopted in the market."

Author: NAJI Faouzi

Follow our innovations on social networks

We frequently publish on social networks (Linkedin, Twitter and Medium) our innovations and the new functionalities of our industrial management solutions.

Also, we would be happy to share with you the latest trends in industrial management 4.0 through high quality content that you could share with others.