Page 17

EETE JULAUG 2014

that of speech recognition, which took a long time to come through? SG: Yes it is the same. Speech recognition took forever, because we had to recognise that the real problem was that only 90 per cent accuracy was nowhere near good enough. With NUI there are even more areas of misinterpretation than there are with speech. What we do is to combine several NUI features at once (multi-modality) so that intelligent interpretations can be increased to an accuracy close on 100 per cent. HW: What do you see as the one change across the market that has left the biggest impression on you? SG: There have been two revolutions, one of which I have made a minor contribution to. The internet world (access to any data) and later on mobile 3G/4G where there is access to data everywhere. This has changed completely the way we think and the way we are communicating. Twenty years ago if you told someone that you could speak to them while driving your car here in Tel Aviv, while they were driving theirs in Australia, they would not have believed you. They would think you were crazy. And yet, when you look back at the beginning of the internet no one thought it would be so revolutionary. HW: Do you have any regrets about selling your two earlier companies, and how do you decide when the time is right to sell a company? SG: Regrets? Yes and no, because it’s hard to let go of a company. But you know when it is right to sell the moment you think it is not easy to grow by yourself. You reach a point where unless you can see a way of bringing in a definite amount of resources you cannot see how growth will happen. There are not many venture capital investors that are able to take the risk of injecting tens of millions of dollars, and so you have to sell it to someone who can invest that money or you run the risk of the company stagnating, or worse. It’s not just black and white, but you do have to consider if there are 50 families involved in your company whether there are advantages for them in entering an acquisition. HW: How do you start to devise gesture interpretation products? SG: You start with the idea that you own a smart device that is not being used for most of the time. Then the process is to turn that device into a kind of a smart friend that can recognise and advise you when needed and the same time detect when you are angry or excited, for example. The first thing required to implement this is to get different experts together to discuss the issue and draft the architecture of the system, which later on will become the layout of the chip architecture. You need to be aware of the limitations of contemporary technology and ask what is out there on the market that we can bring to our product in a reasonable time. So we are looking for something that can be done in the 1-2 year timeframe. HW: Did you know that you were going to develop a chip? SG: We were aware that we were trying to set a high barrier that no one else could compete with in the next year. To do this, we needed to combine hardware, software, firmware, systems and so on. The only way to do it is with an integrated chip. HW: Fortunately, you never went into EDA, but from where you’re standing, what advice do you have for that market? SG: The big thing that needs to emerge from EDA is the ability to arrive at a solution more quickly and in a more methodological way. When you start to design something, you start with a big problem that you need to breakdown into lots of smaller problems. Solving these problems and then reintegrating the bigger solution is the key. It’s like in early days when people were trying to program a chess game. In most cases the human player will beat the machine because the designer can’t think of every possible requirement and outcome. HW: Do you see the world of automotive design opening up for NUI? SG: What we are doing can be used in automotive design, although we are probably ten steps away from that today. The issue is reliability. Misinterpretation when you are sitting in front of your television may not have the same impact as when there is a misinterpretation in your car. This is why the NUI market started with gaming. In gaming if you miss something, well, you’ve just missed it, that’s all. But in automotive it is all a bit more challenging. HW: What are the main challenges ‘In future we won’t be short of technology: it’s just that many people won’t be able to afford it…’ for new technology over the next two decades? SG: The main problems we will face in the next 2 decades are medical and aging issues. Cell phones will become increasingly important as a health monitoring and diagnosis system. Economic issues will also play a part. We won’t be short of technology: it’s just that many people won’t be able to afford it. Also medicine always lags behind technology, because in most cases it exists to provide a cure, rather than having come into place as a preventative. www.electronics-eetimes.com Electronic Engineering Times Europe July/August 2014 17


EETE JULAUG 2014
To see the actual publication please follow the link above