Expert Opinions |


Promoted Content
Send in e-mailSend in e-mail
Promoted Content

Amichai Antebi: A.I. spurs progress in so many fields

Artificial Intelligence (A.I.) has made huge strides in recent years and is increasingly being implemented in many different fields. There are many examples. Lets start with some that are well known and continue with those that are more surprising. Of course, there are all the digital assistants such as Google Now, Microsofts Cortana, Alexa from Amazon and others They use A.I. to learn from every search and request as well as from the users behavior (for example, what did he click on and how long did he stay on the same result) in order to learn about the quality of the result and to continuously improve the results for that user specifically and also for all users in general.

Obviously, these learning capabilities are also used for video. YouTube and Facebook, for example, know how to recommend the most relevant clips.

In the area of sales, Amazon really dominates today, mainly due to its use of machine learning. Amazon knows how to build a unique shopping page for each user, featuring the most relevant products and deals.

In the automotive sector, there are already autonomous cars without drivers made by Tesla, Google, Uber and others – as well as driverless trucks.

In addition, machine learning helps connect between people using transportation by finding common routes, thereby improving travel efficiency. Using Uber and other companies, A.I. already helps to significantly reduce traffic jams in big cities where Uber operates.

In industry, there are now robots that, thanks to machine learning, can see and understand how to manufacture various pieces, which has a significant impact on lowering production costs and improving quality.

The environment is another area positively affected by A.I. Google recently announced that they succeeded in reducing the energy consumption of their server farm by 40%. This is a considerable savings, resulting from the use of machine learning to determine the optimal temperature for operating the servers.

Today, AI is able to learn additional parameters in various factories, thereby improving efficiency and increasing savings in materials and many other resources. Furthermore, thanks to AI, it is now possible to predict air pollution in advance and to receive information on the best way to act.

One of the most fascinating fields for A.I. is medicine. The rate of progress in this field is tremendous thanks to A.I. – at first without new technologies; only through data collection related to genetics and to diseases, treatments and their results, personalized treatment of every patient can immediately become more efficient and enable huge improvements in both quality of life and life expectancy.

Moreover, A.I. is the best way to interpret the results of all types of medical tests. For example, it already analyzes MRIs and X-rays better than doctors because it learns from countless examples that are equivalent to more experience than 1,000 doctors put together

And if this werent enough, A.I. also helps discover new medicines. For example, it can learn which gene should be repaired in which person, based on the vast amount of data that has been collected, and it also learns the best way to try to formulate the medicine. Thanks to these technologies and others in the field of medicine, scientists predict a significant increase in human life expectancy as well as quality of life.

Translated and reprinted from

Amichai Antebi is a software engineer and artificial intelligence engineer and has a degree in computer science. He is an expert on technological innovation and writes for the tech website

Dr. Hagit Perry: Personal shoppers reinvent retail using A.I.

In recent years, a new retail channel has taken off quite dramatically with growth rates exceeding 1,000% – subscription e-commerce. This evolving channel is already available in a variety of retail categories, including beauty, toys, kids clothing, womens fashion, groceries, personal grooming and pet products, and is provided by companies such as Birchbox, StitchFix and Winc. Using A.I., companies in this growing segment fit products to their subscribers and prepare boxes containing specially curated items. Prices range between $10 and $300 per box and the boxes are sent right to the subscribers doorsteps. In fashion, for example, the new subscription services can be thought as automated personal stylists.

This new retail channel is predicted to continue growing rapidly and to become a major shopping method. Customers of this channel say that it makes life easier for them. For example, young parents that work full time, find it hard to shop for their kids who frequently need new clothing, books and toys. Thus, they subscribe to personal shopping services that send them the items that have the best fit for their kids taking into account their needs and budget.

When they subscribe to the service, customers complete a questionnaire about their tastes and needs. Predictive systems use the given information together with collaborative filtering, Bayesian learning, and other algorithms to define a personal box for each subscriber. Then, the box is prepared and sent to the customers.

In kids clothing stores, for example, if customers live in New York, it is January, and the kids ages are five and eight, the A.I. service will know that the children need winter clothing. Moreover, using collaborative filtering it will find the brands and styles that are popular with other kids of the same ages from New York. The A.I. service will add the personal information that the customers entered in the questionnaire and using Bayesian learning it will determine the best items that are available at the store for the customers kids.

The accuracy of the current A.I. algorithms is decent and is improved rapidly through varied innovations. The best way for companies to improve the accuracy on their end is by gaining more data through acquiring a large customer base and by developing panel data about each subscriber. Moreover, when customers return items, the automated predictive services immediately adjust their information about the customers personal tastes and improve themselves for the next boxes.

The new retail channel, which is based on A.I., is a win-win venture for many customers and companies. However, it also has some other economic implications. This channel needs a fraction of the common retention marketing budget in retail. This is because the subscription services continue on a frequent basis unless the customers cancel them. Thus, retention rates in this channel are much higher than those in the regular online stores. While big stores can greatly benefit from this, small stores will have a harder time to break through. This can impact the competition and prices in many markets. Moreover, marketing and advertising budget will be cut, which will require marketers to re-invent themselves and innovate. Nevertherless, it seems unavoidable that most of the future marketing innovations will use A.I.

Dr. Hagit Perry is the Head of the Big Data for Business specialty of the MBA program, IDC Herzliya

Alon Talmor: Will A.I. really take over the world?

Artificial intelligence has recently emerged as a prominent part of our conversation. From conquering the last human-dominated game of GO to self-driving cars, A.I. is now increasingly influencing our lives. What enables this recent breakthrough is neural networks or Deep Learning – a technology that tries to mimic some of the brains behavior, such as recognizing objects in images or controlling a walking robot, by learning from examples.

Some consider the Turing Test to be the ultimate goal of A.I. This test is passed when a human tester is unable to distinguish whether he is chatting with a human or an A.I. Chabot. And though we are still a long way from passing this test, it has long been speculated that machines will inevitably take over the world.

So are we really doomed? In my opinion, the fact that we are ALL worried about it suggests the answer is NO. The argument is simple. Humanity, as a group, is extremely unsuccessful in collectively predicting the first time a certain catastrophe occurs, historically speaking. In 1894, everyone predicted the Great Horse Manure Crisis – that in 50 years time, every street will be buried under nine feet of horse manure. Before the year 2000, we all thought that Y2K, a change in one computer date digit, will cause all computers to halt. Positive predictions such as The Titanic is unsinkable were not extremely successful either. Even more sensitive predictions, such as total nuclear annihilation, has luckily not yet occurred, although it was thought to be highly likely 70 years ago.

On the other hand, real events such as the drop of the first atomic bomb, the outbreak of World War Two, the 2008 financial crisis and others, were not foreseen by the majority of people. We even don't really know when frequently reoccurring events such as earthquakes, hurricanes and financial crises will happen, until they are in front of our eyes....

My point is that the fact that we are worried perhaps makes us all come together and prevent the worst. And the field of AI is no exception. 'OpenAI,' a company backed by Elon Musk, is aimed at preventing any one company from controlling a superior A.I. by open-sourcing and distributing state-of-the-art solutions. Conversation is also beginning to ignite government regulation on A.I. research.

Even from a practical stand point, when it comes to worrying about A.I., some of our conceptions are not entirely relevant to machines. For instance, the thought that shutting down a computer will cause it to revolt is not likely to happen since, unlike people, a computers mind is usually saved, thus we are actually always putting it to sleep

So, finally, it seems that history proves that worrying about A.I. is essential to our future existence – so keep up the good worries!

Alon Talmor is a software entrepreneur and was the CEO and co-founder of BlueTail, an A.I. start-up acquired by He has a Masters in Brain Sciences and is pursuing a PhD in Artificial Intelligence at Tel-Aviv University, and is a lecturer on A.I. and Brain Sciences.