This is the first day we woke up while it was snowing. According to the staff, it usually is cold this time of year but snow is pretty much unheard of. Luckily for us we are traveling quite a while today which means we probably be walking around in the sun rather than the snow.
We walked past ATR and NICT multiple times (they are in the same town as for example Omron, the first company we visited in this study tour) but today we actually get to visit them. And to be honest, these are the ones we have high hopes for.
When we got there we first walked towards ATR as we thought they were located in the same building. After a minute of so, one of the receptionists walked us to the right building one kilometer down the street.
NICT uses a lot of research from ATR to speed up their own research. One of the cool things we saw was a prototype of the ubiquitous home. In here, the whole home is controlled by a computer (with pressure sensors, RFID readers and camera’s) which turns on the lights, makes sure you don’t forget stuff and records shows you like. Something like HAL2000 but then in 2008.
Another project uses pressure sensors in the floors of the entrance hall to track occupants and determine their intent from just the pressure patterns: is someone waiting, hasty, idling around etc. Even though the idea is cool there is not much to see except for the floor of pressure sensors.
After NICT we walked back to ATR (which we now knew how to find) and we were greeted by Nick Campbell, one of ATR’s big shots. He turned out to be a really nice guy and he told us to follow him into the restaurant for lunch.
Another weird thing about Japan is are the rules regarding presents and such: we had to pay our own lunch because otherwise it could be considered a bribe… We heard that in some cases it would be possible to get a free lunch but that has to be approved by numerous large organizations and it would take months.
Lunch consisted of a kind of fish – salmon I believe, soup and rice. I ditched the rice, drank the soup and was then pleasantly surprised to find out the the fish was prepared great. After lunch we moved towards the back where we found a coffee parlor. With a bit of help from a Japanese employee we managed to find the right button on the ticket machine to get a ticket for coffee and some cookies. After we got some fresh made coffee from the lady at the counter we heard we could walk around for a bit because we were early (that was a first) so we decided to walk around on the ATR site.
When we passed the parking lots we turned the corner and found ourselves next to a huge grass field with some stone barbeques next to a small terrace. This was meant for employees that wanted to have a party at ATR. According to the signs next to it, the party was allowed to continue until 10 PM – some party…
Next we walked by some huge pond with a small forest around it and some benches along the way – no doubt the place to go if you need some peace and quiet. After we passed that, we emerged from the trees to find a soccer court and a few tennis courts. ATR sure knows how to keep their employees happy.
We finally walked back to the entrance to find the rest of the group. ATR does research on a wide range of topics including robotics (they have full size humanoids and robot faces with realistic skin) but we came to see ubiquitous technology so we had to skip that section. Mostly.
We got to see the voice recognition and syntheses that ATR build. The computer is a UMPC and the program works within a certain context to make sure the translation is accurate, in this case the topic was travel. The software even copies the intonation of the speaker when it translates.
Next we got a presentation from Nick Campbell himself and his research on following a conversation with out listening to the words. He found out that we use a certain intonation when we speak to certain people (father, mother, sister, aunt, lover etc) and he created an algorithm that determines the intonation and deducts the softness of the voice. It was pretty cool to see how accurate the software worked.
ATR also has a group that is working on learning algorithms. They programmed a robotic arm with some stereo camera to learn how to do a certain move. For example we taught the software what the command "stack" means or "move": you simply show the computer what you mean and it can then do the movement, even if the number of blocks differ, their color differ and even if their position is different.
After the very successful visit to ATR we traveled all the way back to Kyoto for a great dinner in the hotel we were staying at, the Palace Side Hotel. We finally got to eat Western style prepared food. Well mostly Western anyway: the Japanese think its normal to serve the garments (the vegetables and the potato’s) cold. Even with the cold garments, the food was great.