Watch videos with subtitles in your language, upload your videos, create your own subtitles! Click here to learn more on "how to Dotsub"

Machines Helping Humans Helping Machines...

0 (0 Likes / 0 Dislikes)
- Machines helping humans helping machines. Welcome David.

[David Orban] Thank you very much for the opportunity. Thank you for being here today.

My name is David Orban. For me these opportunities of being at these conferences are always not only a pleasure but also beginnings of new conversations. That is why I am showing you my WeChat code. Please add me. And I hope that you will have potentially an opportunity then to keep up with the conversation after the conference as well. I have 168 slides to go through in 15 minutes. [laughter] Not true. I cut a lot of slides for respecting our wonderful simultaneous translators. I would like us to give a hand and thank them because they are doing a wonderful job. [applause]

Thank you very much. What I want to talk to you about today is how in the forthcoming Internet of Things the challenges of the localization industry are going to be very interesting, and they will be completely new challenges compared to before. I am also recording my presentation on video as well as I will be uploading it and also uploading my slides. So feel free to just refer to your favorite search engine to find them after a while.

All of this is going to be based on the understanding that exponential change is driving today's world. What is happening in the world is due to the fact that the power of our computers is increasing. Every 18 months, according to Moore's Law, we are going to be able to buy a cheaper and more powerful mobile phone. However, there are only 7 billion people on the planet. And the electronics industry wants to keep selling what they are producing. And this is the very simple economic reason for the birth of the Internet of Things because as the power of exponential change keeps growing and growing and growing, we are going to enter a very new era of unexpected consequences.

My company, Dotsub, is specialized in online video captioning and translation. And we leverage the power of this exponential change. Our client YouTube releases periodically updates to their statistics of what is the rate of uploads of video, and these are staggering numbers. And I'm sure that for local video platforms like Youku and Tudou, these numbers are similar and exploding really. So one of the latest data points is 72 hours of video are uploaded on YouTube every minute.

And Cisco is also releasing incredible statistics. Ninety percent of the global Internet traffic on the consumer side, according to Cisco, is going to be dedicated to video this year.

And the reason is very simple. We are human, and video allows a very powerful evolutional connection, even if we are not sharing the same physical space. However, videos are a black box for search engines. It is very hard for a search engine to understand what is going on within a video. And even more, whether your video is in Chinese or your video is in English or any other language, only a fraction of the global audience is going to be able to understand what is the original language of that video.

So that is why we are partnering with people like Kay Koplovitz who is the founder of the cable industry in the United States, together with Ted Turner, and they are creating interesting ventures specialized in vertical fields like Videum, which is in the vertical field of health and pharma, providing access to information in innovative ways.

The changes around us do not occur simultaneously all over the planet. As we build our global civilization going from the pyramids in the Egyptian times to the Middle Ages and the Industrial Revolution and today our globally connected world, technology is key. And we have to understand that hardware, software, and design go hand in hand. And generations of mutual solutions are what are deciding the exponentials, whether it is vacuum tubes, integrated circuits of successive generations, or even today now we are about to enter the age of quantum computing. We have to leverage the new platforms.

When computers were born, they were behind in that they couldn't understand—they couldn't even remember— anything from one day to another. When you turned them off, they forgot everything, and they had to start from scratch, and you had to feed them your punch cards again. And then little by little computers started to acquire memory. And then they started to understand, the desktop metaphor told computers what were documents and what were folders and how we organized information and that we like color. We like aesthetics, because we are aesthetical animals. And when the touch interfaces arrived, it coupled the power of computers to understand the world from the limitations and constraints of the human hand. And we are now giving computers sight so that they can understand our gestures and our emotions. And we have given computers the ability to hear and to understand our words. And things are not stopping there because already computers are starting to understand our thoughts.

Every object is going to be smart. There is no object that is not going to have these kinds of senses because the cost of adding these to objects is exponentially decreasing, but the value that derives from that is huge. And the old paradigm of separated, isolated, hierarchical, centralized organizations is being replaced by a new paradigm of networks and interconnected, distributed, and decentralized organizations.

The consequence of this is that whether you are in New York, like is our case where Dotsub headquarters are, or whether you are as we are together with you today in Shanghai, the challenges we are facing and the opportunities that we can build together are very, very similar regardless of where you are on the planet.

When I talk to audiences like you—recently I was the keynote speaker at Cisco's Internet of Everything event which is a worldwide event that they hold—what I talk about is how these changes are designing threads and understanding around the world and how the machines that are going to be composing the Internet of Everything in those need to be very different in how they are designed, deployed, and managed. They have to be totally autonomous. It is easy to understand why. You already have to charge your phone. You already have to move the pictures from your phone to the computer. You already have to update the operating system and the different updates and maybe install and update and monitor your antivirus software and everything else.

Even though we already have maybe a half a dozen smart objects around us, it is already a chore. And when we will have not a half a dozen but maybe 100 objects per person in any environment, they will have to be very, very silent. We will not want to hear from them ever. Anything they do, anything they decide will have to be done by their own. We will want to live our own life, and they will have to live their life. There are already examples of this type of change. Very popular robotic vacuum cleaners that are going around in rooms cleaning them, do not need us to recharge them because very simply they remember where the charging station is, and when they need it they go back and recharge themselves.

The conversations that are going on in WeChat are mostly amongst people, but I visited Changing on Friday, and I saw the showroom of Internet of Things properties and products by the hundreds of next generation's WeChat notes that are going to constitute conversations amongst machines. And I met the former CEO of Tencent yesterday who confirmed that they are building a platform for machines to start using what used to be a network for people to talk to each other. This is similar to what is happening, for example, with the large particle collider at CERN in Geneva where people don't see the data and don't touch the data. Machines decide what should be done with it. And an important example for my next section is the robotic car which sees and perceives everything necessarily, but the decisions that the robotic car makes are very important for the lives of people.

So when you enter a room where there are hundreds of computers that are more powerful than each or maybe all of the computers in this room put together, what happens since we know that they have to be very, very silent? They cannot say anything to us. They have to be very aware of not only what we explicitly tell them but they have to be very aware of also our emotional state. They have to be able to read us, whether we are happy or we are sad, whether we are in a hurry, and they have to be able to make decisions based on that.

Last year there was an experiment and it became known that an important social network was running. And this experiment created an outrage because it became more visible to everybody that computers not only are reading our emotions. Computers are also writing our emotions. We are programming our behavior, and they are learning to do it better and better. So— - He waves his head like this. You do? Did you do this thing? You did? You did? What? [laughter]

[David Orban] The challenge of the localization industry is to really work together with engineers, with designers, with hardware and software developers because the assumption that there are universals in emotional expression and emotional reactions are false. Just like we have separate languages and these languages express nuances importantly in a different manner, the way we communicate emotions and the way we react to stimuli, changing our emotional stance, are not universal.

So imagine just a thought experiment where your smartwatch as you enter a room tells your calendar that your heartbeat is higher than usual. And your calendar responds to the smartwatch, "Of course, he has an appointment with his boss in 15 minutes." And then they agree to alert your music player to start playing some song that is soothing and they realize together the fear of that is actually fine. Your heartbeat is decreasing. You are acquiring a calm state. It gives you a call when your boss arrives. Since the voice recognition system is continuously running they understand that the call is going very well, and they pat each other on the back, "Good job."

However, as innocuous or positive as this example might look, things are not obviously that simple. When machines start to think and outcompete humans as IBM's supercomputer called Watson did beating the best players ever in the history a couple of years ago. Or when Google's deep mind plays video games better than any human could, not by being told what to do but by learning totally on its own. Or when a robotic car has to decide who should live or who should die in an accident where somebody necessarily will because it's only a question of whether to swerve to the left or a swerve to the right but somebody will suffer.

These are decisions of good and bad that have to be made. And today there is not enough of a global conversation what it means to create a science of morality that must precede an engineering of morality. These machines cannot make decisions without those conversations happening. And obviously we cannot let them make decisions without us also agreeing with what are the criteria for that. And you, as protagonists of the localization industry, understand that the world is not what it was.

A single moral system, a single way of living, is not what we want. That would mean that there is a global planetary dictatorship. What we want is tolerance and inclusivity and understanding of the differences.

And this is the challenge that the localization industry has to face together with the engineers, together with the scientists, to create machines that help us not only with artificial intelligence, AI, but also intelligence augmentation, IA. Because 100,000 years ago, man decided to start to produce fire, to start merging with technology. And that decision fundamentally changed us. That is the decision that we are following through with. Today we are trying to find solutions to our challenges, innovating, inventing new ways of analyzing the world. We cannot go back. This is how we adapt to the new conditions, by adopting new technologies. And today the distance from having a great idea to going to market with a product that embodies that idea is going to zero with really unbounded opportunities for all of us.

So my challenge to you is to go back, if you may, if you want, to your offices and ask yourselves what your role in the forthcoming Internet of Things is going to be. Thank you.

Video Details

Duration: 19 minutes and 58 seconds
Year: 2015
Country: China
Language: English
Views: 59
Posted by: dotsub on May 6, 2015

How do we communicate with trillions of machines? We don’t. It is up to them to find ways to aggregate the information they collect and present it in ways that humans can use. It is the duty of the localization industry to lead the development of guidelines for these interfaces that implement best practices for clarity, cross-cultural understanding, accessibility and usefulness worldwide. In this session we will present examples of how this can be achieved while preserving and enhancing the value of the localization industry and its participants.

Track: Translation Automation
Wednesday, April 15, 2015, 12:00pm – 12:45pm
Held in: Grand Ballroom II
Localization World 2015 Shanghai, China

Caption and Translate

    Sign In/Register for Dotsub to translate this video.