Weeknote for 8/23/2020

For having done so little, I sure have a lot to say.

Productivity

😐

A temporal black hole passed through my week. It’s the only explanation for where the time went and how little I got done. I suppose there’s one other explanation, that my prednisone ended and left me to my own devices to deal with the sleep schedule it had scrambled for me. But I had a reset at the end of the week in the form of an involuntary early bedtime, so maybe this week I can recover.

I’m trying a new anti-procrastination technique. A lot of tasks involve physical objects. In the past I’ve tried putting these objects out of place in plain sight so I’ll be reminded to do the task and the materials I need will be easy to grab. But this isn’t enough, because if I’m determined to procrastinate, I get used to having the objects out of place, and they fade into the scenery.

So I’m trying something new—putting the objects in the way of other everyday tasks. For example, a few weeks ago I needed to vacuum. So I plugged the vacuum cleaner into my bathroom outlet, and I made a rule that I couldn’t unplug it until I’d done the vacuuming. This made it inconvenient to recharge things like my electric toothbrush. I still put it off a few days, but the vacuuming got done.

The next rule I need is to spend my tired and lazy moments making a bit of progress on my projects rather than scrolling through social media.

Math

😐

I started transferring my notes to RemNote. My time for this project fell into the black hole, so I’m still working on the notes. Hopefully I’ll finish that early this week and I can return to learning.

AI

🙂

I finished The Quest for Artificial Intelligence by Nils Nilsson. Some thoughts:

• I didn’t realize the author was the inventor of the well-known A* pathfinding algorithm, which he developed for a well-known robot named Shakey.
• Competition was a major driver of research at certain points. For example, the reaction of the US to Japan’s Fifth-Generation project in the early 1980s. Another was DARPA’s Grand Challenges for automated driving in the early 2000s. Well don’t look now, but China wants to be the AI king. I should read Kai-Fu Lee’s AI Superpowers soon.
• Military agencies have been another major source of funding in the US since way back in the 1950s. Early on it was from the Office of Naval Research and ARPA.
• Even though AI winters sound like doom for research, AI research continues during them, just with less funding, and I’d say we’ve recovered from the major one in the late ’80s. Judging by Nilsson’s account, progress plodded before that period and zoomed afterward.
• The book reinforced my observation that AI is less about perfect reasoning and more about making the most of the system’s limitations. Here and there a theme of idealism vs pragmatism surfaced. A prominent example was the debate in computer vision between reconstructing entire 3D scenes and processing only the elements relevant for action (see the faculty page of Yannis Aloimonos).
• Expert systems could help with my modeling project. Creating them involves a step called knowledge acquisition. The old approach consisted of interrogating subject matter experts to build the knowledge base. The questioning methods they used could be instructive.
• Thinking about all this AI reminds me that I have a long way to go and I need to keep moving. Maybe I should pretend I’m already in grad school and study like I have deadlines.

People

🙂

Monday night my family played Go Fish over Zoom. Since Go Fish involves passing cards between players and we were each playing with a physical deck, we took some time to improvise new rules.

1. Since we didn’t have a seating arrangement, our turn order was from youngest to oldest.
2. Instead of passing cards, each player had a discard pile for cards they had “passed” to another player. The player receiving the cards would look through their draw deck for the corresponding cards. If it didn’t have enough of the right cards, the player would take them from the discard pile.
3. Since everyone had the same amount of cards and could all theoretically collect every set, the winner was the first to five sets. The game still took an hour.

One interesting effect of using individual decks is that it was helpful to know which sets each person had collected, because other players were still collecting those sets, and only the players who hadn’t collected them could pass those cards. Our cameras didn’t show the sets we’d collected, so we had to be reminded of other players’ collections or remember on our own.

Another effect is that the game worked less well if some of the players shared a deck, because someone might pass cards to one of the deck-sharing players that were used up by the sets that their deck-sharing partner had collected.

And a final thing I learned is that playing games in a recurring video call is a nice way to pass the time.

Tuesday I went in to work for a team picnic. We have a team lunch a couple of times a year. It was a fun conversation, as usual, and for social distancing we were able to spread ourselves out over three picnic tables and an uncomfortable canvas folding chair. I wasn’t in the chair, but it was a conversation topic at one point.

Nature

😰

Thursday morning I found a wasp in my bathroom. It was hanging out on my toothbrush. Scary. It took a long time to figure out what to do about it. Thankfully it seemed to not like flying, and eventually as it crawled across the floor, I plunked a food container over it, slid a piece of paper underneath, and took it out to the balcony. I nudged the container partway over the balcony’s edge so the bug could fly out. Since it was such a sluggish creature, I expected it to take a while, but a few minutes later it was gone.