(Pineau, Gordon, and Thrun 2006)

PBVI

One-Liner

“If we can avoid the curse of history, the curse of dimensionality wouldn’t be a problem”.

Basically - most POMDP problems don’t reach much of the belief simplex. So, can we concetrate planning on more probable beliefs.

Novelty

1. trajectory based approach to select beliefs
2. belief set is fixed through layers: each backup results in the same number of layers

Notable Methods

PBVI

Key Figs

New Concepts

Notes

pipe chains the STDOUT of one command and put it to the STDIN of another command. Typically, we want to do pipe per direction.

command pipelines

1. span two child processes
2. create a pipe to allow the two processes to communicate
3. connect the first child’s STDOUT to the pipe + the second child’s STDIN to the pipe

pipe()

pipe() gives us back two file descriptors, such that whatever is written to one can be read from another.

I, for sure, spend equal if not time writing the data flywheel code compared to architectural changes.

This has been bugging me for awhile, but I haven’t had time to sit down and write about it until now. In my mind there are three things that makes coming up with a good take about this (and also implementing it) hard:

1. This may be a problem that’s solved in much of the frontier labs (I would argue more because of workflow optimizations that sidestep the problem, see below)
2. The “solution” to the problem also seem extremely person-dependent, so it seems hard to make an overall suggestion.
3. Many people write it off as “necessary engineering”

I don’t purport to have a solution to the pipes problem here, but I want to spend a little time now to reflect about what I have been thinking about while spending days at the office writing pipes.

User Story

Sejin is the executive administrative assistant at Nueva, working on scheduling Liza, Lee, Terry, and the other admins against the members of the wider community. Sejin spends most of her day scheduling people, of which, the largest time drawn is spent convincing people to move their schedules “in favor” of that of another person (i.e. manually). The reason why this is done is because her approach to scheduling is one-shot: emailing everybody for general availability, noting in her mind who the high-priority attendees are (say, Liza), and if no times match asking/convincing those in lower priority to move their schedules. Although she enjoys the process of putting events together, she is particularly frustrated that, due to the busy schedules and often back-and-forth emails needed to get and conform everyone’s schedule, response rates to complicated scheduling problems are low.

The PSC is a supercomputing center.