posted
Well, I thought I'd respond to Noemon's request to hear about my trip to Japan. Sorry about the misspellings and bad grammar that will probably be in this long message.
So first of all it's probably important to explain that I have never been out of North America. Second, I have never presented my research at a conference before. Well on Sept. 26th I got on a plane to Japan to do just that.
The flight over mostly involved me trying to get my homework done for a class. I was pretty stressed and didn't sleep the entire 13 hr flight from Chicago. I arrived in Japan, tired and very sore. I immediately encountered something you see quite frequently in Japan, the line. This particular line was to go through immigration. It turned out to take 1.5 hours since 3 flights all arrived at the same time. Luckily I'd carried on all my luggage so I was immediately able to go down into the airport and begin my getting lost.
After some amount of work I found cash, and a Japan Rail pass. So I get on a train and find to my joy that there's someone who is obviously American. It turns out he's lived in Japan for 6 years with his wife and child. His little boy is very cute and is very excited when he finds out I'm taking the Hayate train to Sendai, the location of the conference.
It turns out that the hotel staff at my western style hotel speaks no english. This makes things interesting. It takes me a while to find out that yes indeed there is wireless at the hotel so I can actually submit my homework assignment from Japan. It also means that later on in the week when I want to make a local call I'll never figure out how to.
I spent a lot of time being jetlagged in Japan. I finally settled on what I called the all nighter schedule. I'd usually fall asleep between 8-10 pm and be up at around 4-5 am. Since Japan is 13 hours ahead of EST I was feelign a lot of pain.
The first full day I was there I went, checked in at the conference and went and saw the site of Sendai castle. I then proceeded to obsess over my homework assignment, darn algorithms, and turn it in.
The next day was basically free since I wasn't signed up for a workshop. So, I decided to go north of the city to Matsushima. This was supposed to be one of the prettiest sites for the whole time. And, it turned out to be the case.
I got on what I *thought* was the correct train. The only think I managed to do right was take a train on the correct line in the right direction. I guess, 2 out of 3 aren't too bad. Sadly my train stopped much earlier than I wanted. I was sitting on the train and suddenly the car emptied except for 3 people. My confused tourist warning sense immediately indicated that something was wrong. Fiddling with my map I made eye contact with the nice looking Japanese gentleman across the way. Pointing to the train map I asked, "doko?" Basically where are we. He immediately responds, to my relief in English and asks where I am going. It turns out he speaks 4 times more English than I speak Japanese. It also turns out this is just enough for us to figure out that 1) my train is stopping at the next stop. 2) I will have to get off b/c this is the wrong train. 3) I need to transfer. This is scary. I can not read the signs and at this point I'm out of the city where even less people speak english. He says I should ask the conductor. I'm not very certain that this will work but I nod in agreement. Then he suggests that perhaps he should ask the conductor. This sounds like a great idea to me. The conductor tells him to tell me to take the 9:22 train that arrives on the next track. This is one of the best things about Japan. If the train is supposed to arrive at 9:22 it's there at exactly that time.
I enter the car to discover that there are 3 very foriegn(to Japan) guys on the train. In fact I hazard a guess that they are European and maybe German. It turns out that 2 of them are from Germany and that one is from Portugal. It also turns out that one of them is dating a woman who does very similar work to mine but is unfortunately at the conference. So I got to run around with them and see the islands and the Zen Temple. We then headed back to make it to the Opening reception at the conference. Before going there I ended up at the mausoleum.
The opening reception was pretty neat. We had some free sushi and some free sake. I also ran into someone who knew me. And I had no idea who he was at first. After he introduced himself I realized I'd organized a visit for him at our school. His name is Rus Tedrake, and he does passive walkers at MIT. Passive walkers are robots that don't even need a computer. They can perform walking just based on how they are built. His stuff is pretty cool. He's also working on some reinforcement learning stuff built on top of these types of walkers to get interesting gait (if I am remembering this correctly). What this means is that he uses a type of learning that's inbetween supervised learning and unsupervised. Instead of always telling the learner things like, that action was bad or that action was good, you can just say that series of actions succeeded here's some positive reward. Or that failed here's some negative reward. It was nice to see a friendly familiar(after I remembered him) face since I was alone at the conference.
The next day was the first full day of the conference. It was also the first day of my talk. Naturally I was up at 4am so I had lots of time to obsess. I got to the conference and watched some talks remotely related to some of the stuff I do. Then after lunch I was up!
My research involves basically watching people move around an environment and training hierarchical models to recognize the activity. One of the things I definately need to do is get more complex activities that have more obvious hierarchy. We'll have to see how that goes. The talk went pretty well but I ended up going pretty short. This meant there was lots of time for questions which was good and bad. I was really relieved that inspite of the throat problems I was having I managed to survive my talk!
I then got to enjoy the rest of the conference without any problems.
Let's see if I can recall some of the high points and then I'll finish up with my loonnngggg trip home.
There were a lot of people doing Simultaneous Localization and Mapping or (SLAM). This means that as the robot walks around the world it not only tries to build a map it also tries to figure out where it is in this map it is building. Actually a lot of people were doing mapping at the conference. One of the plenary speakers was Sebastian Thrun who got a lot of this going using statistical methods and particle filtering. His talk was pretty interesting.
The banquest was quite tastey and we ended up eating a 5 course meal. QRIO, a miniature humanoid robot made by Sony, entertained us for a while after dinner. They had the robot programmed to talk and do a traditional Japanese dance. It was really cute but what would be more interesting would be if there was some idea of learned behaviour. Well, that would be more interesting to me at least.
The little pet seal robots that were being used for therapy were really really cute! Sadly there were no Aibo's so I couldn't see the puppies.
One of the talks from MIT was on Leonardo, their robot built by the guy who made Teddy in AI as well as many other movie robots. This group, part of the Media Lab, is doing work on social robotics. They had a neat movie and while I knew some of the underlying systems were shakey it was definately an interesting idea. They were showing how you might be able to teach a robot through social interaction. And how the robot should be able to use social cues to help the person with this.
Anyway this is getting really long so I'll skip to my story of the trip home. I had to catch the first train to the airport. As it turns out I got off at the wrong terminal and then had to take the bus to the correct international terminal. I then managed to convince the woman that I really could carry on both pieces of luggage since they both fit the size requirement. I usually stow my backpack under my seat and my small luggage above. She seemed a little hesitant but let me get away with it. It turned out our flight time had changed to 12:20 although, my ticket said 12. I ony had a 2 hr layover in Dallas/Ft. Worth so now I was down to 1 hr and 40 min. Then once we got in the air it turned out that we were running late . This ended up giving me a 54 min layover to transfer between an international to domestic flight. I got through immigration quickly because there were very few americans on my plane and the line was short. Luckily I didn't have to get any luggage since I carried all of mine on and managed to get through customs quickly too. As soon as I got through customs I saw a huge line to go through a seperate set of metal detectors. I was in shock. I'd already walked through so many different security things in the last 21 hrs I just couldn't understand why I had to deal with this too. Well I could but I didn't have to like it. I scowled at the line trying to determine if there was some secret way around that I was missing. Suddenly this woman was asking me what I was doing and after she saw my connecting flight I was moved immediately to the head of the line. This was fantastic since I now had approximately 20 min to get to my flight. I ended up walking/running as fast as I could through the airport trying to ignore the blisters on my feet from all the walking I'd done in Japan. I managed to just make my flight and was quite glad that there was no hassle with anything. I was feeling a little stressed and tired and a little out of control of my emotions. But the real fun was about to come. Once I got back to Hartford, CT I took my bags and walked out of the airport. I got on an escalator right before the door. Finally, I could see the light at the end of the tunnel. I saw the escalator was about to end thought okay just one or 2 more steps took a step forward and was caught in the escalator. Yup that's right my shoe was caught. And it wasn't like there were laces, my black zip up boot was being eaten, with my foot *in* it by the escalator. Luckily some part of my brain was still functioning. I managed to unzip my shoe and pull my already pinched foot out of it. Someone else told me they had got help and I just stood around. However none of the airport staff ever came by to tell me that someone had been called. After waiting for a while I looked at the 2 escalators and decided since the were both supposed to be going down that I was not dragging my stuff back up the jammed escalator. So I walked up to see what the deal was. This woman acted very surprised and was trying to get me to sit down. I was not exacty in the best state of mind at this point and ended up telling her that as long as someone had been called I'd go sit by my stuff. I think my exact words were "I've been traveling for 24 hours, I just want to go home. I'm going to go sit by my stuff and wait"
Then another woman started to chew me out and tell me I shouldn't have left my bags. I just looked at her and said, "I don't care. I really don't care." I think at this point the airport staff realized I was on the verge of a break down and left me alone. They got my basically destroyed shoe out. I put it back on my foot since it was the only pair I had and went outside to wait for my ride. Naturally when outside I completed my nervous breakdown cried for 15 min and apparently worried the other patrons because they were constantly asking if I was okay. This only made it worse since what I really wanted was a small dark place to sleep in. But after everyone left me alone I started to feel a bit better and was recued by my ride.
Posts: 872 | Registered: Mar 2002
| IP: Logged |
posted
Oh, hmm. Yeah I guess I did end it on a bad note. Other than traveling home the trip was GREAT. The conference was exciting and Japan was very nice. Actually even though you can't read the signs Japan is a great place. Everyone is so friendly and helpful. It's one of the few places where obviously being a tourist is a good thing. No one marks you as someone to beg to or hassle and everyone knows you're going to be lost and confused.
Posts: 872 | Registered: Mar 2002
| IP: Logged |
posted
More! More! More on robots! More on robots! It's amazing how little news coverage science conferences attract, and how little info about the topics in those conferences is released to the public in an easily accessible form Thanks for the info on Rus Tedrake. I'd read a few blurbs about his work, and promised myself that I would look further into it. By the next time I thought about it, I'd forgotten his name as well as where I read those blurbs.
And the touristy tidbits on Japan were also nice. More?
So, was it just coincidence that you were almost eaten by an evil robot (a non-humanoid machine automated to take over the burden of walking up&down stairs) after attending a conference on making robots into more obedient servants?
posted
Can I just say that I hate the Dallas/FW airport? I've only ever done domestic there, I can't even imagine international. Every single time I have had to run to catch my flight. Even if we are supposed to have a long layover, and the flight leaves on time, we'll end up taxiing around the terminal for an hour before we can get off the plane. And you were doing this whole trip by yourself? I would have been crying at the escalator.
Glad the trip part was fun and that you now have that first conference under your belt.
*wanders off to look at pictures*
Posts: 3956 | Registered: Jun 2001
| IP: Logged |
posted
Celia, oh don't worry I finally had my breakdown after the escalator and had a good cry. Which of course just meant that everyone was paying *more* attention to me.
More robots. Well I should go through and find all the really interesting stuff I liked. One of the papers I heard about from a different year was an IROS 2002 Paper, "Fish Catching by Visual Servoing and Observed Intelligence of the fish". I can only find links that show this paper exists not the actual paper itself. However, the story I heard was that they had a manipulator trying to catch the fish. And of couse, it was too slow to catch the fish and the fish kept getting away. So, they ran their experiments for a longer time thinking they would do better than they had in the past because the fish would get tired. However it turned out that they did worse because fish learned the pattern of the manipulator and was able to avoid it more easily. Perhaps this is a sign that we were a bunch of geeks but there was a table of 6 of us almost crying we were laughing so hard as one of the students told this story. I mean one can you just imagine the poor fish. And two they could have just randomized their algorithm and performed better. Yeah that's right I'm a geek.
Posts: 872 | Registered: Mar 2002
| IP: Logged |
posted
Sure I read this just hours after finishing "Battle of Corrin".... escalators are the new thinking machines, eh?
Japan sounded wonderful and I hope you share more stories of your time there. The airport fiasco... well, doesn't just about everyone who travels by plane have at least one airport horror story? Hopefully this will be your one and only =)
Posts: 4515 | Registered: Jul 2004
| IP: Logged |
posted
sarahdipidity, this is such a cool account. I'm too sleepy to write much, but I'm sitting here flabbergasted by hw interesting your life is. )
Posts: 14017 | Registered: May 2000
| IP: Logged |
posted
DFW is pretty bad. But I have to say, it's nothing compared to LAX... We always pay the extra money to fly to Burbank when we go home rather than LAX.
Posts: 21182 | Registered: Sep 2004
| IP: Logged |
quote: I just looked at her and said, "I don't care. I really don't care." I think at this point the airport staff realized I was on the verge of a break down and left me alone.
I am thinking of the 14 hours I spent just waiting in airports on my way home from Japan in October. There was a point in San Francisco where I would have gladly roasted any human being who told me once more to go through a line.
But you are right about Japan. Wonderful.
I enjoyed your post, though the robotics are not my field.
Posts: 440 | Registered: Oct 2001
| IP: Logged |
Sadly, I missed the deadline. I just can't seem to get this project to work. Actually that's not true. My previous method is working too well and I'm having a difficult time coming up with a domain that will show improvement. *shrug* I'm not sure what's wrong with it. But anyone that has a lot of complicated activity data should let me know. *laugh*
Posts: 872 | Registered: Mar 2002
| IP: Logged |
Printer-friendly view of this topic
![[Eek!]](eek.gif)
![[Smile]](smile.gif)
![[Kiss]](graemlins/kiss.gif)
I would have been crying at the escalator. ![[Frown]](frown.gif)
Actually that's not true. My previous method is working too well and I'm having a difficult time coming up with a domain that will show improvement. *shrug* I'm not sure what's wrong with it. But anyone that has a lot of complicated activity data should let me know. *laugh*
