The Blogg

You know what’s annoying? When you know that you are a hair’s breadth away from sneezing, but it won’t actually happen until you let your guard down. Of course, that’s not the only thing. I am also annoyed by continuous draining of the sinuses through the nose, the ever-present mild headache that results from sinus pressure, that raw burning in the back of your throat, the asthma-like symptoms that I get with an infection, the dry, cracked lips that result from dehydration, general exhaustion, shivering in an 80 degree climate, and the fact that by this time tomorrow or the next day I won’t be able to lay down flat without causing an uncontrollable coughing fit.

Yep, my ever-frail self has a cold again. At times, it boggles my mind that with all of our advanced technology and culture, we are still basically helpless against these viruses. Sure, we have medicines designed to make the symptoms less severe, but recent studies have shown that, at least in young children, they are no more effective than placebo. Because of my inherent distrust for pharmaceuticals, I probably have a reverse-placebo effect: if I think I have been medicated, I start to feel worse. Others will claim that super-high doses of Vitamin C or other herbal remedies are effective, but the science seems to indicate otherwise. Basically, our immune system gets no help from modern man.

Thankfully, the only thing that I *needed* to do this week was today, so I can spend the next two days sleeping as much as possible and wiling away the miserable hours when I cannot. I am quite aware that I am a pathetic sick person, and that many people go about their lives relatively unchanged while suffering from the common cold or much worse. I know that I am lucky to have this disease over others; I haven’t had serious gastroenteritis in 7 years, and I’d take a week of the cold over a day of that. If I were to ever contract a *real* illness, the kind that other people manage regularly, I would probably see how petty my complaints are. But that doesn’t make it easier.

About this time two years ago I wrote a post about my frustration with Broderick Bunkley and other professional athletes and agents who act like children when it comes to their compensation. Sure enough, Bunkley was basically useless that year, primarily because he chose to not report for training camp on time. He had a good season in 2007-2008, but I am sure he would re-do his rookie season if he could.

This year the same sort of things are happening. The Eagles managed to sign all of their draft picks on time and had no contract holdouts, but we still have the same sorts of attitudes. After some speculation that he would not do so, star running back Brian Westbrook showed up. While talking to the press, however, he revealed that “holding out was a possibility”. I applaud his decision not to do so, but he should not have even considered the option. He is currently in the third year of a 5 year, $24.9 million contract. As one of the very best running backs in the league (the best, depending on the statistics you use to make the determination), he should be paid more. But the point is that he signed a contract, and he should honor it. (His current contract was, surprise surprise, the result of an extended holdout.) If he continues to play at the level he has over the last few years, he will be able to get a very lucrative contract when this one expires in 2010.

Lito Sheppard has similarly come to camp but made it clear that he is not satisfied with what he is being paid or the possibility of not starting. If I were a very good worker and my employee planned to demote me and hire someone from outside to take my position, I would be understandably upset. I would not, however, go whining to the media or threatening to stop doing my job.

I realize that professional football players have a somewhat unusual situation in that their employer may terminate their contract at any time, but if the player does so he may not sign with any other team. It is not a level playing field, and that sucks, but it does not excuse this behavior. In exchange for this, players get paid exorbitant sums of money to do jobs that at least 50% of men would do for free if given the opportunity.

Around the league is more of the same. I have not followed other teams nearly as closely as the Eagles, but a quick perusing of the sports page reveals that running back Steven Jackson is holding out from the Rams camp, Devin Hester considered doing the same with the Bears and will do so if negotiations break down, Kurt Warner is unsatisfied with his compensation in Arizona, and so forth. But certainly the biggest story is Brett Favre’s second un-retiring. After quite clearly telling management he was retiring from football, he now wants them to accept him back as their starter or cut him loose for another team to pick up. It should be obvious why either of those options would be unacceptable to the Packer administration. But Favre is not interested in what is best for the team; he is looking out for number one (or in this case #4). He has said he is tempted to “call their [Packers management] bluff” and show up for training camp, recognizing that it would cause a media circus. In the same interview he admitted to feeling “a little bad” for Packers quarterback Aaron Rodgers. The situation his selfishness is putting Rodgers in should make him feel much worse than that. Favre has a reputation as a superb athlete and a good guy, and I am not sure why he is so desperate to throw that away.

There is one Eagles player who did not come to camp, although for “personal”, non-contract reasons. I can think of lots of good personal reasons to skip camp: maybe he is at his mother’s deathbed, or helping a friend through a psychological crisis, or something. The fact that the team is calling this an unexcused absence, however, makes it seem unlikely that he has such a noble reason for not being there. If he would just make a press release, or authorize the Eagles to do so, stating exactly why he is not there we could all rest a lot easier. He might be the best right guard in the league, and not having him for the season would be a significant blow to our Super Bowl aspirations.

On a completely unrelated note, Chris Cocca pointed out this to me. High-larious.

I had the opportunity to attend the Association for the Advancement of Artificial Intelligence’s AAAI Conference in Chicago from July 12 – July 17. For those friends and family who were interested, I kept a fairly detailed log of my time there, although the quantity and quality deteriorate as time goes on. You can follow my travels on the CTA’s Central System Map, which covers most of the area I was in.

I could not fall asleep until after 4:30 am on Saturday night, which was not a good omen. I woke up before my 10:00 am alarm, somehow, and was on the road by 10:30. Traffic on 476 S and 95 N was not bad, so I make it to the economy parking lot at Philadelphia International Airport just before my self-imposed deadline of 12:00. The US Airways ticket counter at Terminal B was absurdly long, but had very few people working. Those that were all seemed to be catering to specific groups of which I was not a member: first class passengers, corporate fliers, people with small children, etc. I had been hoping to deal with a live person, since this was my first time flying and I was unaware of the proper procedures, but eventually gave in and used a self-service kiosk to print my boarding pass. When doing so I was offered the choice to upgrade to first class for $100, which I find to be absurd. I do not know what percentage of the airplane’s volume is reserved for first class passengers, but I certainly hope it is small. Anyone who is willing to spend an extra $100 for a slightly more comfortable seat for 3 hours is far too wealthy.

The new security screenings were lots of fun. There is no signage whatsoever to tell you what you should do. Instead, there was a woman at the metal detector occasionally yelling instructions, but you were lucky to have heard them in time. One of her directives was to remove all laptop computers from carry-on bags. Unfortunately, I did not hear this until my luggage was already being screened. Another worker seemed like I had added a great strain to her day as she removed the laptop and then sent it and the bag through a second time while I muttered apologies. Extrapolating from what I saw, I would have to guess that they have more than 100 Transportation Security Administration employees working simultaneously at that airport, which seems excessive. I vaguely remember riding along to either drop someone off or pick someone up from an airport when I was much younger, and my recollection is that we were inside the terminal with them. I suppose these were more innocent times.

After a brief stop for lunch, I made it to the gate with about 20 minutes before they were starting boarding. The flight itself was as uninteresting as the seat was uncomfortable (that is, very). I did enjoy the beginning and end of it, however. I had an excellent vantage point to see what was going on with a window seat just a few feet behind the wing. The view just as you break through the top of the cloud cover is one that really must be experienced to be understood. That two small wings and a lot of forward thrust can propel a giant cylinder up through the air is quite remarkable.

Upon landing in Chicago I had the fun of baggage claim. As far as I can tell, anyone could take any luggage they felt like and, unless the actual owner of the bag noticed and made a point, no one would stop you. In fact, you would not even need to have been on a plane. Anyone could walk in from the street and take whatever they desired. The fact that this does not happen constantly restores a bit of my faith in humanity.

I took a 10 minute bus ride to the blue line train, which I then rode for at least a half hour to get to the stop nearest where I would be staying. Along the way I learned two things: First, you do not actually need to take the ‘A’ train if you want to get to Harlem. The Blue Line will get you there also, although it may not be the Harlem you intended. Second, the Chicago Transit Authority is not just one of the greatest albums of the 1960s; it is also a remarkably thorough and well-oiled machine. Within the city they run 8 different trains, 2 subway lines, and more than 200 different bus routes. That beats what I have seen of SEPTA and LANTA hands-down.

I then had to walk perhaps a mile and a half before I found the University of Illinois at Chicago’s Marie Robertson Hall, where I would be staying. This is by far the nicest residence hall I have ever seen. The hallways have the exact look and feel of a hotel. The rooms themselves are actually apartments containing four private bedrooms as well as a kitchen, living room, and two bathrooms that are shared. Being quite exhausted from carrying my luggage around, I laid down for about a half an hour.

Marie Robinson Hall	My Bedroom
The Kitchen	Common Area

After this mini-nap I went for a walk. My first stop was an Italian cafe, where I was surprised to find “Panzerottis” on the menu. (The Panzarotti with an ‘a’ is what I thought to be a fairly regional food to the Philadelphia / Southern New Jersey region. It consists of what is essentially a calzone or stromboli that has been deep fried.) Franco’s will be pleased to know that while this was good, it definitely could not compete with the home offering. The outside lacked that pitted, salty, greasy goodness, and while the crust was large, there was more air inside than anything else. After another mile walk I found a grocery store, where I bought myself some iced tea and sunchips for the week. Walking back to my housing with groceries reinforced my knowledge that I would not be able to hack urban living.

Massa, Source of 3 Dinners

When I got back, both sweaty and exhausted, I was very much looking forward to a shower. As I was just about to start taking one I noticed something odd — we lacked a shower curtain. Rather than spraying water all over the bathroom, I decided to see if the front desk could help me out. While I was there, another guest was complaining that he had not even received towels or bed linens. They searched various supply stations for a shower curtain for a half hour, but found nothing. Eventually, they decided to steal one from a room into which no one had yet checked in. After my eventual shower I went to bed at the absurdly early hour of 9pm, due to lack of sleep the night before, an hour lost to Central Time, and the fact that I needed to be at my volunteer position by 7:30 the next morning.

On Sunday I experienced much more of the CTA, not all of it positive. I dragged myself out of bed at the unholy hour of 6am and began following the instructions for taking mass transit from the student housing to the conference center: walk 2 blocks north to Roosevelt, take the #12 bus east to Michigan, then take the #3 bus south to McCormick Place. By 6:40 or so I was at the stop on Michigan, but I waited and waited for a #3 bus to no avail. Around 7:15 I pulled out my massive transit map, where I found out that on Sundays the #3 bus does not start running until 7:25. Not wanting to be late, I got on a #4 bus that I thought would take me at least close to my destination. I disembarked near Michigan & Cermak and started walking in the direction I thought I needed to go, asked someone for directions, then turned around and walked the correct direction. The result of all this was that I arrived at 7:45, probably no earlier than if I had just waited for the #3 bus.

Working the registration desk was a rather simple task, and there were 5 of us doing what was really a job for 3. There was much to be done until the workshops started at 9:00, but after that the morning dragged on as people continued to trickle in. It was interesting thumbing through the registration envelopes while searching for a name and finding those of luminaries whose papers have been influential to your own work. In my section (J-Q) I noticed Richard Korf, Steven Minton, Tom Mitchell, Ray Mooney, and a few others. Unfortunately, the only one of them that arrived while I was working was Korf, and I had only a very brief chance to introduce myself to him. There was also exactly one person that I remembered from ICAPS, Siddharth Srivasta.

The previous conference that I attended (ICAPS 2007) provided continental breakfast at the beginning of each day and snacks during the coffee breaks between each paper session. I was hoping for the same thing here, but I found no breakfast and the only things served at the coffee breaks were coffee and tea. By the time my shift ended at 12:30 I was quite hungry. (On Tuesday through Thursday, the coffee breaks did include snacks.)

I walked west on Cermak to find some lunch and stopped at a walk-up establishment called Baba’s. Here I had my first sampling of “authentic” Chicago cuisine — the Italian beef sandwich. This consists of thinly sliced beef and cheese, served warm in a hoagie roll, and yet is nothing like a cheesesteak. Rather, it is much closer to a French dip. The difference is that the au jus is poured directly onto the sandwich as it is being made (making it much easier to walk around while eating) and that the beef is cooked with tasty Italian spices. It was quite delicious, and I predict I will have another one before leaving. I declined a bag and sat down on a bench outside the convention center to eat. At this point I was immediately reminded that I was in the Windy City, as everything not nailed down tried to blow away. Next time, I’ll take the bag.

The Hyatt Regency McCormick Center

Baba’s Famous Steak, Source of 4 Lunches

After lunch I got on the #21 bus, which took me to Cermak / Chinatown. There I boarded the red line train with expectations of exiting it at the Clark / Division stop. Unfortunately, I discovered that, due to construction, the red line was being re-routed over the purple line instead. I instead got off at Chicago & Franklin, then walked to Wells & Franklin, from which I took the #66 bus to Dearborn and Franklin. From there I got on the #22 to Clark & Goethe. I walked a bit more, again got directions and discovered that I was going the wrong way, and eventually made it to the Chicago History Museum.

Skyline from Cermak / Chinatown Station

The museum was interesting, but I would have preferred more emphasis on those things that were uniquely Chicago-an (the importance of being a transit hub, first by sea and then by railroad, to the city’s growth; the meat-packing and steel industries; the Great Chicago Fire; the Haymarket Square incident; the Gangland era and the St. Valentines Day Massacre; the ’68 Democratic National Convention; the World’s Fair; etcetera). In particular, I thought the museum’s coverage of Chicago’s importance in the history of jazz and blues music was quite poor. They had a small section on this, but it was filled mostly with concert posters captioned with “[performer] was from [place], but frequently played in Chicago”. Any musician that toured North America would perform frequently in Chicago, simply by virtue of its size and location. There was no mention at all of the important migration of jazz out of New Orleans to Chicago or the recording of Louie Armstrong’s influential Hot Fives and Hot Sevens albums, and only scarce mention of the fact that the city inspired one of the most important sub-genres of blues music. This should probably merit a museum of its own. All of this was indeed covered in the museum, but there was also large portions of it devoted to things such as the Native American settlements in the area and fur trade throughout the 18th century. This is important, of course, but not really that much different from the history of every other North American settlement during the time period. There was also a large section devoted to visual art made either about Chicago or by artists associated with Chicago. This was interesting, but somewhat unnecessary due to the fact that the city has a Museum of Contemporary Art, the Art Institute of Chicago, the National Museum of Mexican Art, the Vietnam Veterans Art Museum, and at least one other facility devoted solely to visual art. There was also a small temporary exhibit dedicated to the band that took its name from the transit authority and then, under threat of litigation, the city itself, containing concert posters, instruments, and commendations from the city council and state legislature. As a fan of the band, I appreciated this.

Leaving the museum, I took the #36 bus south. Both the transit map and the marquee on the bus stated that it went south to Polk & Clark, but when we reached Van Buren & State, the driver kindly asked if I knew that was the last stop. When I responded that I thought it went to Polk, she said that it never does. I walked for a bit, then took the #29 bus down to Roosevelt & State, then got the #12 to Roosevelt & Halstead, from which I could walk back to the housing. I ended the day enjoying a gelato and complimentary wifi at Massa, the Italian cafe where I had dinner the night before. Once again, I was in bed and asleep not long after 10:00.

On Monday I woke up at 6:00 again, although there was no need to on this day. I stayed in my room, taking care of various things, until around 8:30. Then I walked to Roosevelt & Halstead, took the #12 bus to Roosevelt & Michigan, and walked north on Michigan to a curious restaurant naked Yolk, where I purchased a delicious freshly-made cinnamon roll. I then walked across Michigan and through a section of Grant Park, snapping pictures as I went, before reaching 11th & Columbus. Here I waited about 25 minutes for a #10 bus, which I took to the Museum of Science & Industry.

Aaron Montgomery Ward Gardens Plaque	Walking Figures Perspective 1
Walking Figures Perspective 2	Grant Park Statue
Panoramic View Part 1	Panoramic View Part 2
Panoramic View Part 3	Panoramic View Part 4
A Flock of Segway-ers	Central Section of Museum of Science & Industry

I found the museum to be not properly about science or industry, but about the intersection of the two: engineering. This suited me quite well, because I suspect that my general interest in the area and education would mean that I already have the superficial knowledge that can be provided by a facility such as the Franklin Institute in Philadelphia. I had the option of purchasing admission to a variety of special exhibits and demonstrations with my general admission, and I am glad I chose not to; after the passage of 3.5 hours I realized that I had only seen the first floor of three, and that I needed to leave in about 2 hours. Fortunately, the exhibits on the second and third floors were both smaller and less interesting to me.

The first section I visited was filled with a collection of photographs of various industries and feats of engineering. This turned out to be actually one of the more interesting segments. I paid special attention to the four or five photographs of the Packard automobile factory, since I have seen “Number One” in the lobby of the building where I work (Packard Lab).

I next went through a room with models of a great many sailing ships and steam-powered vessels, as well as small artifacts from seafaring in those ages. This was fairly interesting, but would have been more so if they could manage to have all of the models built to the same scale, allowing the viewer to see the relative sizes of the different ships. Beyond this was a room with several generations of automobiles, both those designed for general passenger use and ones designed specifically for racing. Apparently there was an early competition for self-propelled vehicles that ran a course starting in Chicago, and of scores of applicants only two complete the track. Interesting stuff. There was also a large Foucault pendulum hanging through one of the staircases, much like the one at the Smithsonian.

On the other side of the first floor was a fascinating exhibit on the U-505, a German U-boat that was captured by the United States Navy during World War II. At this point I realized that photography was permitted throughout the museum and started taking advantage of it. There was much information about the remarkable operation that allowed the sub to be damaged but saved from destruction and towed from the coast of Africa to Bermuda. In addition, the actual submarine was held in a large hall in the museum, with many placards explaining the uses of its various external and some internal components. Although it would seem very tiny to be inside of it within the vast ocean, I am sure, the submarine was staggeringly large. In addition to displays on how the engine drive, ballast tanks, diving planes, weapons systems, and other important technical components worked, there was also much information about life in a submarine, the typical foodstuffs that would be packed for a long journey, and other, more personal aspects. Tying everything together were exhibits of many small artifacts recovered from the ship such as binoculars, uniforms, tools, code books, and even two Enigma machines. Along the way were displays explaining the more general history of both submarining and World War II, and an extensive collection of American propaganda posters from the time urging men to join the Navy, women to join the WAVES, and everyone to prevent loose lips from sinking ships. I declined the additional-fee guided tour of the inside of the submarine, but still easily spent an hour and a half absorbing everything in this section.

Port Side of U-505	Starboard Side of U-505
Sail of U-505	Rudder and Screw of U-505
Aft Torpedo Tube of U-505	Engine Removed from U-505

I next went to the Henry Crown Space Center. It had a number of interesting artifacts, including an engine from one of the Apollo rockets and one of the Gemini capsules. While not as complete as the Air & Space Museum in Washington, D. C., it had good coverage. I found the sections on plans that NASA and their equivalents around the world have made for replacing the Shuttle program and continuing to expand the ISS most interesting, because I was largely unfamiliar with them. The first floor also contained a section on farming technology, which I only gave a cursory exploration, and several children’s sections.

Apollo Rocket Engine

Gemini Capsule

The museum contains a cafe where I stopped for lunch. Not wanting to eat too much and desiring to experience another traditional Chicago food, I got a hot dog. (I am not, however, quite sure of its authenticity.) My plans were stymied when I was delicious-looking baked potato soup and was forced to get some of that as well.

I hurried through the Networld section of the second floor, presuming that I should already know anything explained there. It appeared to be a reasonably complete overview of computing technology and especially the Internet. I had to fight to resist the urge to run through screaming, “It’s not a dump truck, it’s a series of tubes!”. The second floor also had a section on genetics, with a focus on the ethical questions surrounding the exploitation of it, and one on railroads. Most of the rest of the second floor required either waiting in long lines, waiting for a demonstration to start, or purchasing admission to a special exhibit.

Significant portions of the third floor were closed for construction, but there was a wing devoted to aircraft and one devoted to chemistry. For me, the most interesting part of this was a wall-sized copy of the periodic table of elements, with each entry highlighting any unusual properties of the element and a few uses of it that a typical person might encounter, when appropriate. As it was approaching 4:00 and the soles of my feet were crying out in agony, I took the #10 and #12 buses and my feet back to housing.

Jetliner, Similar to One in which I Arrived

Jet Engine

I found the museum fascinating, and appreciated their permissive stance on photography. If I had been able to take the time to explore the second and third floors as thoroughly as I did the first, I could easily have filled the time between when the museum opened and when it closed without doing any of the optional extras. I can only speculate, but it seemed that they did an excellent job of balancing the needs of children and adults. I do, however, have a few suggestions for ways the museum could be improved. I noticed three embarrassing grammatical errors on placards within the museum. There were also many places where placards were placed so low that it was necessary to kneel or squat in order to read them. It was not always clear whether I was viewing an actual artifact or a reproduction. In some cases, it almost seemed as if they went out of their way to obscure this. For example, there were videos in the section on the U-505 that were of a quality that they could easily have been shot from one of the airplanes used in the capture. It was only after they started showing the actions of the sailors inside that it became clear that the video was a reconstruction. Finally, both here and the Chicago History Museum made extensive use of audio/visual stations. I recognize the importance of multimedia in drawing people’s attention, but the audio can be very distracting as you are trying to read something else. It would also be nice if a transcript of these videos were displayed beside them for those who would like to simply read the material and move on.

After a shower and much-needed rest, I headed to the conference center for the opening reception, where I attempted to overcome my innate introversion and make some contacts. For the most part, I was not successful in finding the sort of people that it would behoove me to know, although I got to speak briefly with Pat Langley, whom I had already met. I started by taking my plate of cheeses and breads to a table of Master’s students from Michigan, then moved to one filled with neural network researchers from Canada, Germany, and Russia.

AAAI Opening Reception

Following the reception was the video competition, in which videos designed to educate and motivate students to enter the AI field had previously been judged and awards were given out. Unfortunately, only a 30 second trailer of the winning videos was shown, so it was rather difficult to understand why they were selected.

I then went to dinner with a prestigious group of people: David Aha [Naval Research Lab] and his wife and son, Marc Ponsen [Massstricht University] (a briefly former colleague of mine), Susan and Ian Craw [Robert Gordon University & University of Aberdeen], David Wilson [University of North Carolina, Charlotte], Michael Youngblood [University of North Carolina, Charlotte], Kalyan Gupta [Knexus Research], Michael Floyd [Carleton University], Sabine Hauert [Swiss Federal Institute of Technology, Lausanne], Hector Munoz-Avila (my advisor), and one other person whose name I unfortunately cannot remember. We were going to take taxis to Gino’s East at Ontario & Wells, but ended up squeezing the 14 of us into a limousine designed for 12. At Gino’s East, where the walls were all plastered with approved graffiti, I had my second authentic Chicago food experience: pizza. Unlike the familiar dish, this had cheese directly on the crust, an 1/8 inch thick sausage patty above that, and then about an 1/8 inch of sauce on top. An interesting and tasty idea, but I think I prefer our East Coast variety. While waiting for dinner and then eating I had several good conversations.

14 AI Researchers in a 12 Passenger Limousine

Geno’s East at Closing Time

Everyone else was returning to the convention center, where they had hotel rooms, while I looked for CTA transportation back to the student housing. The only bus that used the local stop only ran until late afternoon and it was nearly 11:00, so I was searching my map for an alternate route when someone interrupted to offer me directions. I ended up allowing him to walk me several blocks to the closest blue line train stations, listening to his story. Being the softie that I am, I allowed him to convince me to give him $20, ostensibly to be used for a fee to acquire the state identification card he would need for a job interview the next day. Was his story true? Probably not, but he did save me nearly that much in cab fare, was pleasant conversation for 15 minutes, and was clearly substance-free. As soon as I got home it was straight to sleep.

Tuesday started at 7:00, when I was able to help out some other conference attendees who had just arrived and not yet figured out the transit system to McCormick Place. I was able to enjoy about 45 minutes of Internet access before the first session of the day, enough to read a backlog of emails but put only a small dent in my RSS feed. The opening ceremony was as uninteresting as it sounds, and Eric Horvitz’s presidential address was not much better, focusing on the historical context that most of the attendees surely knew quite well.

I then attended a session on Planning & Scheduling. The first paper was on temporal planning, using an idea of splitting durative actions into sets of simple actions with additional preconditions and effects to ensure they obey the proper durations in spite of any interleaving that may happen, and using some insights about these temporal constraints to constrain the search for a plan. The second was about probabilistic planning through compiling probabilistic domains into equivalent deterministic domains and using statistical sampling to reason over plans. Both of these were very good. The third presentation was about scheduling any-time algorithms, which was less interesting to me. The session chair arrived late and was ineffective, so the first two presentations each took 25 minutes, allowing only 10 minutes for the third.

My second session was on search, and contained one particularly good talk by Richard Korf. He described the process of space-optimizing a search to find the longest shortest-path in an example problem domain, reducing the space complexity to only 2 bits per reachable state. It must have been an excellent presentation, because I was able to anticipate the central idea several slides before it was provided.

At lunchtime I ate another Babo’s Italian beef sandwich, and convinced Dr. Munoz-Avila to do so as well in spite of his cholesterol concerns. We dined with a group of people from the Carlos III University in Madrid, several of whom are working in the same hybrid of planning and learning that I am. They had several entries in the current ICAPS competition on planning and learning, and their explanations of how much time this took away from basic research confirmed our decision not to attempt to enter ourselves.

After lunch I attended an invited talk by Mark Newman on studying graphs. I was already aware of much of the background that he provided through my interest in web hyperlinks and online communities, but he presented it in a very interesting and useful way. His own work focused on the connection between social networks and the spread of disease, and the ways that properties of the network could be used to protect against pandemics. This was also quite interesting, but I had much difficulty staying awake and focused through the second half of his talk.

For the third block I went to a session on search in planning, but stayed for only the first talk. After the one that looked particularly interesting to me (a theoretical study of the effectiveness of various heuristics across a variety of planning domains, by Malte Helmert), I went out and spent 40 minutes recovering from lecture fatigue and emptying the rest of my RSS inbox. My final session of the day was a Nectar (important papers that were originally presented in other, more specific conferences, but have applicability to the general AI community) session on planning. I should have realized this in advance, but was disappointed to find that the three papers were all ones I was familiar with from ICAPS the previous year.

At the end of the day Dr. Munoz-Avila and I had a meeting with Dr. Qiang Yang and one of his students from Hong Kong University, with whom we are collaborating on a project. The student was unaware of the recent rescheduling of the meeting, so the three of us had drinks in a restaurant that was part of the hotel and discussed the work only sparingly.

After observing several talks, mostly Malte’s, I realized that mine could be much improved with a relatively small amount of work, so I spent the rest of the evening working on changing its aesthetics in a room at the conference center where I could get Internet access and then at Massa, where I also ate another Panzerotti with Nick, a student from the University of Illinois at Urbana-Champaign, who was staying in the same apartment as I. I got to bed around midnight again.

Wednesday morning was much the same. The morning invited talk was about extrapolating back the human genome with the genomes of other species as they branched off from the same path we took back to a common ancestor of all mammals, fish, and birds. Very little of the work has been done yet because only the human and mouse genomes have been completely sequenced, but based on incomplete data they have already found certain genes that are very similar among all non-human species in this line that have been studied and quite different in humans, and have been found to be somehow involved in the construction of a certain part of the brain during fetal development. The talk was quite interesting, but did not discuss any specific AI techniques used.

I then attended the IAAI invited lecture, which was about cognitive prostheses. Despite the title, this was not about cybernetics, but simply computer programs that offload some reasoning that person would normally need to do themselves to computer technology. The main example was a new computer-based instrument for military aviation that would synthesize the data provided by many sensors into a supposedly intuitive interface. The presentation was very well written and explained, but I did not find the topic especially interesting. I then attended a session on ensemble learning, mostly because a paper on clustering looked interesting. Unfortunately, none of the authors of that paper showed up.

After a lunch of a hot dog from Baba’s, I went to the afternoon invited talk by Stuart Russell (of Russell & Norvig fame). I was expecting an excellent talk, and was not disappointed. Most of his time was dedicated to arguing for the use of first-order logic over less expressive reasoning systems. Fortunately for me, he also talked briefly about some work on a new formalism for HTN planning, which helped to motivate my own work.

I presented my paper in the next session. The presentation went well, and I got several good questions from people who seemed to find the work worthwhile, which was encouraging. We also got a bit of a dressing-down from our friend Pat Langley due to our lack of a slide listing related work (of which his is the most important). I have never considered this information to be within the scope of a 15 minute talk, and judging from the presentations I have seen while here neither do most people, but his point is valid. In the future I will include this information at the end, where I can discuss it if there is time and if not at least show that I intended to so as not to offend anyone.

I then attended a session on recommendation and collaboration on the web. I was hoping there would be something here about recommender systems, but found this not to be the case. There was, however, a good paper by Jennifer Golbeck, who made up at least a third of the bibliography of my depth study survey of models of trust for the web.

After the technical program for the day, the ballroom was opened for a reception, poster session, and system demonstrations. I first stopped at a demonstration by Dan Morris of Microsoft Research of a program called MySong, which takes as input monophonic sound data and uses hidden Markov Models to create a chord progression to match the provided melody. It has several variable settings that can control this by changing “happiness” — essentially whether it uses a major or minor key — and “jazziness”, which balances the concerns of choosing a chord that will be consonant with the melody notes and choosing a chord that makes reasonable harmonic motion with those around it. I sang jazz standard “Secret Love” into his microphone, and with the proper tuning of parameters it produced something quite musical, although not identical to the traditional arrangement. I talked to him quite a bit about his work and my related honors thesis from Ursinus College.

I then looked for the buffet and was pleasantly surprised to find simply excellent food. I ate macaroni & cheese, shredded pork, three cheese ravioli, chicken ravioli, garlic bread, capicola, prosciutto, romano cheese, chocolate cake, and a chocolate-covered strawberry. There were also many other options that I did not try. There was really no comparison to the meager provisions of the opening reception. I observed several other demonstrations and posters, but the only other notable things were two posters on teaching AI: one on a re-implementation of the FF planner in well-documented Java that students could modify, and the other an introductory course in which a dozen students collaborated over a single long day to design and partially build a text-based MMORPG.

Thursday morning I checked out of the student housing and dragged my luggage over to the conference center, where my advisor was kind enough to allow me to store it in his hotel room for the day. There were two competing invited talks this morning; I chose one on sentiment analysis, which essentially meant mining the web for people’s opinions on products and classifying them as positive or negative. It was interesting but not ground shattering. The next session I attended was on ontologies, and I got almost nothing out of it. The next was reasoning about action and change, which was actually pretty good, although mostly irrelevant to my work.

I had lunch at Baba’s again (there are basically no other options in walking distance) and had another Italian beef sandwich. While the others had been delicious, this one had apparently been made earlier and sitting, which resulting in a dry, mostly cool meal. I then went to a session on human-computer interaction on the basis of a recommendation from my advisor that any paper by one of the presenters was worth seeing. I found none of this work particularly useful, although it was presented well. Finally I went to a mixed session on reinforcement learning and an invited short talk. I was able to follow the first RL paper but not the second, and the invited talk was a condensed version of a seminar that I saw Tom Mitchell give when he visited Lehigh last year. There was nothing in the last session that looked relevant, so I stopped in the GGP competition for a moment, then took the #21 bus to the red line train to the blue line train to a shuttle to O’Hare International Airport. I tried giving away my transit pass, which was still useful for a day and a half, but no one would take it. I arrived around 6:30 and found out my flight was delayed to 9:10.

General Game Playing Competition	The Tarmac Outside my Gate
No Loaded Firearms? What A Rip-Off	The Lazy Inventor of the Horizontal Escalator is My New Hero

Flying in the dark was rather less interesting than in daylight, and my ears were in a state of discomfort for the entire 40 minute descent to Philadelphia. This surprised me, because I would have attributed this problem to changing atmospheric pressure, and I assume that the cabin of an airplane is kept near 101.3 kPa throughout its journey. I did not make it back to my apartment in Bethlehem until after 2:00 am, which made for a long day.

In summary, my experiences at AAAI 2008 were great. I saw many interesting talks and met many knowledgable people, although the opportunity to engage them in significant conversation was more limited. Most of the papers that were relevant to my work were all crammed into the first day, which is good because my ability to concentrate on presentations decreased throughout the conference. Overall, I think I found ICAPS last year more valuable. The breadth of AAAI is both a blessing and a curse. It is good to see what is going on in diverse fields, but most of it is presented at a level such that I lack the background knowledge necessary for more than a superficial understanding of what is being discussed.

On the weekend of June 28-29 Rachel and I went to a cabin near Williamsport, PA with some of my closest friends from high school (Bob & Sam Groff and Adam Shivery), as well as my best friends from college (Ryan & Lisa Michaluk) and two friends of the Groffs named Bob (II) and Katie. Ryan, Lisa, Rachel, and I drove up on Friday night as soon as all of us were finished work. The drive went smoothly until we were perhaps 5 miles away from the cabin. At this point a fog like I have rarely seen billowed in, obscuring my view to about 15 feet. Even within that limit it was difficult to make out much of anything. Our directions were not quite right, so we were quite fortunate to have Bob and Sam drive up behind us. Between their knowledge of the area and more powerful headlights we got there safely.

After unpacking there was a rousing game of poker, interspersed with Sam dealing a game called “golf”, none of which was for any money. I did quite well, frequently throwing my winning pots to a less fortunate player. When we grew tired of this we moved on to a game of something else that I cannot currently recall, and then to bed.

Front of cabin

Back of cabin

Bunk room

Main room, with Lisa walking to sink

I was displeased to be woken around 9:00, but some yummy egg, cheese, and bacon sandwiches made it a bit better. After breakfast we went down to the very cold creek and spent some time wading and otherwise enjoying it. When we returned we made lunch of hamburgers, hot dogs, and macaroni and cheese. Toward the end of lunch Adam surprised us by showing up early, but without his wife. After lunch some people took naps while Bob, Adam, Ryan, and I taught Bob II to play Magic: The Gathering. At some point through the evening there was dinner of shish-ke-bobs and baked potatoes and games of Pit, Catch Phrase, and Trivial Pursuit. When the womenfolk went to bed the men played poker for actual money. Now that it mattered I became neither good nor lucky, and lost my $5 investment twice.

Ryan bravely tests the water

The pristine creek

You can pick your wife, and you can pick your nose, but only Ryan can pick his wife’s nose

Sam and Katie go wading

Ryan and I take the “man way” through the rapids

Time for pollination

View from the porch

On Sunday we were again roused early. We stayed only long enough to eat a breakfast of pancakes and pack our things before people (including myself) began to go their separate ways. A good time was had by all; thank you to the Groffs for their planning! See the rest of Lisa’s photography here.

Thanks to a new jarring reminder that our bank account is scraping the bottom, I’ve decided to get serious about keeping track of exactly where we earn and spend our money. (Based on my track record for deciding to do things like this, Rachel gives it about a week.) I had written my own basic accounting system long ago as a way to teach myself database programming, but I stopped using it about 6 months after we got married because of the hassle of keeping track of two people’s expenditures. This time around I am using GnuCash, a mostly excellent (see note below) double-entry personal finance system in the style of QuickBooks.

Importing all of my old data and getting account statements for the intervening months was a bit tedious, but now I can do all sorts of interesting data analysis. (I have an unusual love for such activities, and probably should have become a statistician.)

The first thing I noticed and decided to look at more closely was my expenditures on gasoline over the last 5 and a half years. In early 2003 a typical fill-up cost me $10, and this last month it got precipitously close to $40 each time. With a few fluctuations, it seems to have increased roughly linearly during that time period. This is not an exact measure, because the tank is not equally empty each time I replenish it, but in the aggregate the trend is very clear. Below this paragraph is a graph of these values. In July 2005 I replaced my 1990 Geo Prizm (may it rest in peace) with a 2001 Ford Escort. As you can see this made little difference, since both have an approximately 10 gallon tank. The national average price was extracted from a Department Of Energy resource at http://www.eia.doe.gov/oil_gas/petroleum/data_publications/wrgp/mogas_history.html, and it is remarkable how closely my own data tracks it.

Next I decided to look at overall expenses over the last slightly more than two years of our marriage. (After all, that is ostensibly why I started this entire project.) I have ignored the largest expense, medical school tuition, because it would dwarf everything else and because it is being paid entirely with debt. There are no real surprises here; about a third goes to rent and utilities; prescriptions, gasoline, and groceries each take about a tenth. The nearly a quarter spent on “other” things (meaning either that it was categorized into something that did not make up a significant percentage of total expenses, or that I could not figure out from statements what it was for) is somewhat disconcerting. Also disturbingly, rent accounted for only 29.5% of money spent, but is equivalent to 43.6% of income during the same period. Thankfully, our years of deficit spending should be rapidly drawing to a close, as Rachel has job applications out currently.

Speaking of rent, I noticed that in the 4 years that I have lived in Bethlehem I’ve spent a total of 30,400 on having a place to live. It certainly would have been nice to have all of that in the equity of a house, but I am not actually rethinking my decision. I hate to have just thrown this money away, but I am too risk-averse to have considered buying, even if I had been able to make a down payment. Given the precipitous fall in home values over the last year, I am really in no hurry to stop renting any time soon. Also, according to Rentometer we are paying about 10% more than average for similar rental units in the area. Check out your own results.

Now, for my complaint about GnuCash. It does an excellent job of storing data and allowing data entry, but a very poor job of opening that data up for analysis. The built-in reporting system seems very limited, and there is no automatic way to export data for analysis by an external tool. Ideally I should be able to either run (read-only) SQL queries directly or easily export to a file of database INSERTs. Not only is this not possible, there is not even a built-in function to export to a simple comma-separated values format. I was forced to manually copy transaction information out into a text file and use GnuPlot to generate the graph of gasoline purchases. For the pie chart I was able to get some information out of a built-in report, but still had to make some adjustments and generate the chart in OpenOffice.org Having the data is great, but being able to easily aggregate and view it would be much better.

Flutamide
Viagra use in women
Phentermine 37.5 mg
Lowest price for phentermine
Chlorprothixene
Avapro
Xanax online prescription
Xanax ambien
Colchicine
Viagra doses
Order ambien online
Ambien withdrawal
Xanax and drug testing
Viagra lowest prices
Glycopyrrolate
Natural viagra substitutes
Cheep phentermine
Hytrin
Vicodin picture
Provera
Temazepam
Appetite suppressants equivelant to phentermine
Berman sister female viagra study
Order tramadol online
Nitrofurantoin
Phentermine gynecomastia
Review of herbal phentermine
Clorazepate
Methamphetamine
Levivia vs viagra
Phentermine insulin
How viagra works
Mometasone
Naproxen
Cefprozil
Soma carisoprodol
No prior perscription tramadol
Generic sales viagra
Chlorothiazide
Phentermine lortab online
Butoconazole
Oseltamivir
Compare viagra to cialis
Prochlorperazine
Compare viagra to cialis
Cefatrizine
Cleocin
Cheap tramadol cod free fedex
Thyroid
Phentermine alternative
Buy online prescription viagra without
Viagra pills
Gatifloxacin
Xanax side effects
What is xanax used for
Cheap viagra pills
Generic viagra cheap
Canada xanax
Natural supplement equivalent to xanax
Orlistat
Custom hrt phentermine
Xanax high
Generic viagra and generic drug
Phentermine adipex
Phentermine online without prescription
Triamcinolone
Thiopropazate
Viagra and levivia
Comparison levivia viagra
Cialis free sample
Viagra lowest price
Add link phentermine purchase suggest
Iodamide
How much xanax is a lethal dose
Amaryl
Free prescription sample viagra
Trihexyphenidyl
Lowest phentermine
Fioricet order
Terazosin
Lipitor
Oxycodone vs hydrocodone
Mevacor
Phentermine 37.5 diet pills
Female use viagra
Viagra alternative herbal supplement
Tamoxifen
Diet phentermine pill sale
Levitra vs cialis vs viagra
Fluorouracil
Macrobid
Is xanax addictive
On line pharmacy phentermine
Information loss phentermine weight
Imiquimod
Nexium
Cheapest xanax
Buy phentermine pay cod
Amoxicillin
Motrin
Phentermine success
Tramadol addiction
Buy online viagra
Buy cod diet phentermine pill
Vaccine
Lamivudine
Viagra sale uk
Thyrotropin
Skelaxin
Xanax 2 mg
Viagra levivia alternatives
Effect viagra
Diet pal pay phentermine pill
Viagra for woman information
Phentermine
Purchase tramadol without a prescription
Xanax interaction with paxil
Cidofovir
Cheap xanax
Valium vicodin
Phentermine xenical diet pill
Monopril
Buy online pharmacy viagra
Imuran
Burn fat lose weight diet phentermine pill
Trimethoprim
Tramadol pharmacy
Difference between cialis and viagra
Is it safe to take prozac and phentermine
Herbal phentermine review
Diphenhydramine
Avapro
Phentermine xenical
Order vicodin online
Phentermine weight loss pills
Phentermine delivered cod
Phentermine online doctor prescribed
Levivia vs cialis vs viagra
Order viagra canada
Cheap ambien
Imipramine
Online phentermine order
Dicumarol
Kaopectate
Nicotrol
Muse
Cheaper viagra levivia cyalis
Viagra online shop
Desmopressin
Hydrocodone m357
Generic viagra cialis
Phentermine ups delivery
Dipyridamole
Epoprostenol
Phentermine blue diet pills
Methscopolamine
Lipids
Compare levivia and viagra
Picture viagra
Viagra no prescription
Description of meridia
Lowest price on phentermine
Theophylline
Sophia viagra
Frontier pharmacy phentermine
Phentermine amide
Phenylpropanolamine
Keyword prescription qoclick tramadol without
Good morning viagra commercial
2 mg xanax
Buying phentermine online
Free generic viagra
Ephedrine
Duragesic
Adipex phentermine vs
Herbal online viagra
Meperidine
Aspirin
Effects phentermine side strong
Pink oval pill 17 xanax identification
Gabapentin
Xanax and weight gain
Ansaid
Information about viagra
Estrogen
Viagra pictures
Phentermine compare prices
Viagra pulmonary hypertension
Perphenazine
Phentermine c.o.d. Tomorrow
$50.00 phentermine
Phentermine drug test
Recreational viagra use
Amerge
Female viagra cream
Cod overnight tramadol
Cheapest online viagra
Zileuton
Nadolol
Cheapest phentermine pill
Phentermine overnight delivery no prescription
Cetirizine
Nortriptyline
Tramadol 50 mg
Acyclovir
Pantothenic
Erythromycin
Meridia online
Aldactone
Detox hgh phentermine quit smoking xenical
Viagrafix
Diet drug fenfluramine phentermine
Online pharmacy phentermine xenical meridia
Trimethadione
Flagyl
Buy viagra in uk
Minocycline
Order xanax
Naprosyn
Viagra results
Klonopin
Xanax alcohol
Exelon
Chlorpropamide
Dangers of taking xanax and neurontin
Androgel
Granisetron
Zafirlukast
Viagra soft tabs
Buy phentermine diet pill
Phentermine mexico
Buy viagra now online
Arava
Phentermine pharmacies online
Bayer
Fosamax
Xanax anxiety relief pills order here
Phentermine ephedrine
Alternative new viagra
Nonoxynol
Cortisol
Meridia diet
Grapefruit xanax
Prescription viagra
Is klonopin stronger than xanax
Zetia
Hydrocodone on line
Flonase
Side effects phentermine gynecomastia
Avodart
Vicodin m360
Cheap discount phentermine
Oxyphencyclimine
Cimetidine
Demerol
Acetaminophen
Viagra mexico
Tramadol dosage for dogs
Xanax detox
Nalorphine
Demecarium
Cheap phentermine with online consultation
Suicide xanax
Phentermine 37.5mg tablets
Canadian viagra
Can xanax cause frontal lobe dementia
DiflunisalCyber pharmacy viagra
Grapefruit viagra
Purchase viagra online
Decamethonium
Addiction tramadol
Cheapest prescription viagra
Cod free phentermine shipping
Metrizoate
Mannitol
Phentermine 37.5 mg no prescription
Xenical hgh phentermine quit smoking
Phentermine resident sale virginia
Idoxuridine
Capreomycin
No prescription needed phentermine
Viagra cialis levivia dose comparison
Hydralazine
Cialis levitra viagra vs vs
Hydrocodone side effects vicodin
Fioricet with codeine
Levothyroxine
Ionamin
Celecoxib
Diet hcl phentermine pill
Phentermine online diet pill
Hydrocodone apap
Methsuximide
Cialis compare levivia viagra
Viagra herbal alternative
Sertraline
Clofibrate
Cycloserine
Xanax overdose