|
[08 Jun 2015|01:48am] |
|
this entry is backdated
|
|
| Omni Shoreham hotel in Washington DC |
[24 Jul 2008|09:38pm] |
The hotel is pretty luxurious, but as a conference site, it has shown some problems:
* Internet is flaky
* microphones are flaky
|
|
| on a bus, with WiFi |
[23 Jul 2008|05:16pm] |
They have a password, as if someone would steal bandwidth from a moving bus. I'm being forced to watch The Bourne Ultimatum again.
Scheduled to arrive at my DC stop around 2pm.
|
|
| Greenland |
[22 Jul 2008|06:56pm] |
My Helsinki->Boston flight flew about 1000km over of the coast of Greenland, a significant Northwards deviation from the geodesic (we reached 70N). This was at about ~2pm Greenland time. The date was July 21, i.e. middle of summer.
 
 
These light blue blobs intrigued me. What are they?
 
|
|
| GMail problem |
[22 Jul 2008|03:01pm] |
Times are wrong, but timezone is right...
Someone posted this in May, nobody gave a good answer, and I'm having the exact same problem right now. I'm trying to get a Google person to look at this.
|
|
| "data science" |
[16 Jul 2008|02:50pm] |
The Swedish word for Computer Science is "datavetenskap" (data(?)-science).
Generally speaking, Computer Science is known as the "science of information". The "science of data" is obviously Statistics. "data" and "information" are different views of the same thing.
Therefore, when talking to laypeople, instead of trying to explain what "Machine Learning" is, I should describe myself as a "Data Scientist".
|
|
| Finnish cuteness |
[14 Jul 2008|01:44pm] |
Well, it turns out that Spoken Finnish is somewhat different from the written language.
English: Thank you!
Written Finnish: kiitos!
Spoken Finnish: kiitti!
The spoken numbers are mercifully shorter than in writing.
English: two twenty-five
Written Finnish: kaksi kaksikymmenta viisi
Spoken Finnish: kaks kakskyt viisi
On about 5 occasions now, I had to deal with people (mostly old people) in Finnish. I'm proud of myself for managing to convey:
* "I want a cell phone under 60Euro that is compatible with DNA". (keywords: puhelin, kuusikyt euroa)
* "I am looking for Patrik from office A342".
* "Does this bus go to Viikki?" (-mene Viikkiin? -jaa, mene!)
* "Does this bus go to Central Station?" (-mene Rautatientorille? -jaa, mene!)
* "Where is Central Station?" (-missä on Rautatientori?)
* "That bill is for me!" (-minulle!)
Like Dutch and German, knowing a few words can go a surprisingly long way. Some examples:
( Read more... )
|
|
| Church |
[13 Jul 2008|10:08pm] |
Church is a new stochastic programming language based on Scheme (Church is to Scheme what IBAL is to OCaml). The idea is that programs represent stochastic processes (they specify how the data is generated).
Statisticians always want to compute the likelihood the data under certain hypotheses (often plotted as a function of the model's parameters).
The easy way to do this is by rejection sampling, i.e. estimate the likelihood by simulating with the model and parameters, and count how often you get the same data. But this is "exponentially slow".
So these languages support smarter inference methods like MCMC. I would also like to see some support for analytic methods, possibly by integrating with algebraic packages.
Such languages tend to be based on functional languages because functional programs are easier to reason about (unlike Scheme, Church forbids mutation).
Stochastic programming languages seem like a good way to organize algorithms that work in specific situations, seeing how much they overlap, how well they generalize; making it possible to better evaluate the novelty/improvement of new approaches.
|
|
| Helsinki |
[13 Jul 2008|03:48pm] |
The conferences are over. My talk went well. Feedback was sparse (that's the disadvantage of having a talk rather than a poster), but from later conversations, I can tell that some people understood important parts of it. In the last 10 days, I met about half of the important people in machine learning, though I wish I'd had more time to talk to some of them.
I have a phone again, and have many Helsinki-ers to meet in the next week. (for once, I know the Finnish word better than the English one: "helsinkiläinen")
|
|
| frequencies in time series (auto-regression) |
[03 Jul 2008|05:58pm] |
low-pass filter: average over the last n time steps
high-pass filter: the residual of the low-pass filter
|
|
| my phone number |
[02 Jul 2008|10:15pm] |
My number in Helsinki is 044 961 8295.
The pre-paid SIM card cost 17Euro and came with 10Euro in credit; All calls I can make cost me 16c/min (except for International and service numbers).
|
|
| digital video bleg: getting a picture from a movie |
[02 Jul 2008|01:07am] |
Why doesn't a screen-capture of VLC Player / Media Player work?
What would work?
|
|
| 1 minute past midnight in Helsinki |
[02 Jul 2008|01:02am] |
It's 1:04am now, and it's somewhat darker than that, but I could probably walk around without artificial lighting.
|
|
| flight update |
[30 Jun 2008|12:56pm] |
* Pitt-Cincinnati pilot got a stomach flu, unfit to fly; severely delayed. I figured I better not go West if I'm going to miss my connection.
* called Delta Reservations while inside the airplane. After talking to unhelpful person (who refused to look up alternative routes), the Delta gate attendant made a phone call and put me on a direct flight to Boston (USAirways). The line to talk to him was huge, is probably going to last >1 hour*: I'm lucky I was third (and that a Tennessee-bound gentleman let me go ahead of him)
* since this is none of FinnAir's business, I could theoretically get stuck in Boston indefinitely, or forced to buy yet another very expensive flight. For now, sounds like I will be ok.
The luggage is probably still with Delta, but they said it should catch up with me in Boston. (It's checked all the way to Helsinki, so I think I don't need to worry about it)
Updates here.
* - 30 minutes later, ~20 people were still standing in line to see him. He was still completely alone! I couldn't locate any Delta agents in other gates.
|
|
| practice talk |
[28 Jun 2008|03:22am] |
My practice talk went so-so. "Only" 4 people came. I got lots of good feedback.
My talk needed more rhetorical structure (letting audience know what to expect), better motivation, and the scope of our contribution wasn't clear enough.
I find it hard to coordinate the content of the slides with the spoken part without either:
* sounding like I'm just reading the slides, or
* deviating from the slides and presenting information at the wrong time
Are my slides too good?
|
|
| silly electronic machines |
[27 Jun 2008|02:40pm] |
Why does my laptop insist on making a "tada" sound on start-up from "Hybernate", turning off my "mute" button if that's what it takes? Just now, I had to mute it twice, to no avail! I miss having an analog volume dial.
UPDATE: if I "Stand By"(a.k.a. "Sleep" or "Suspend") instead of "Hybernate", it is possible to mute it right after I turn it on, just in time before the "tada". Also, "Stand By" starts up faster!
UPDATE: Also, I'd like to be able to trace which program / Firefox window/tab is making certain sounds. e.g. by seeing a list of programs making calls to the sound APIs + sound drivers.
|
|
| “THEORETICAL MATHEMATICS” |
[27 Jun 2008|11:33am] |
Arthur Jaffe and Frank Quinn (1993) - “THEORETICAL MATHEMATICS”: TOWARD A CULTURAL SYNTHESIS OF MATHEMATICS AND THEORETICAL PHYSICS
Is there a place for speculative mathematics? This paper does a better job than I could at expressing the analogy between formal proof and physical experiment.
From what I've read so far, their main message is: "Physics has a division of labor between theorists and experimentalists. Why shouldn't math work this way too?"
To whet your appetite, this is how the paper begins: << Modern mathematics is nearly characterized by the use of rigorous proofs. This
practice, the result of literally thousands of years of refinement, has brought to
mathematics a clarity and reliability unmatched by any other science. But it also
makes mathematics slow and difficult; it is arguably the most disciplined of human
intellectual activities.
Groups and individuals within the mathematics community have from time to
time tried being less compulsive about details of arguments. The results have been
mixed, and they have occasionally been disastrous. Yet today in certain areas there
is again a trend toward basing mathematics on intuitive reasoning without proof. >>
Being fond of speculation, intuitions and informal logic arguments, I personally feel like the "theoretical physicist" here. My interest in logic and formalization stems from trying to prevent/catch mistakes (and also partly from my desire to resolve disagreements).
|
|
| $920 of Lufthansa credit; "validating carrier" |
[26 Jun 2008|07:40pm] |
Until June 2 2009, I can use $920 of credit on any flight for which Lufthansa is the "validating carrier for the entire ticket" (a.k.a. "issuing carrier").
It seems to be impossible to tell for which itineraries this is the case (even *after* you've bought the ticket), except that they tell me that if I select "Lufthansa" as my preferred airline, it sorts it that way for you . I take this to mean that this way, I'm more likely to get itineraries in which the "validating carrier" is Lufthansa.
Trying to learn more...
( Read more... )
Generally speaking, life has been teaching me that booking things through agencies tends to lead to inflexible arrangements, and partly provides them with excuses for bad service: they can always point fingers at each other, or cite complex rules that the customer is powerless to contest.
|
|
| UBC Housing: 4BR? |
[26 Jun 2008|05:58pm] |
I can check my place in the queue online. For a studio, 1BR, 2BR (my current choices), it's not looking very good.
But I could get a room right now, in a 4BR Thunderbird apartment, with 3 other Joe Randoms, i.e. "upper-year undergraduate and graduate students".
Advantages:
* location: next to CS building
* facilities: gym, etc.
* if I hate it, I only need to stay 4 months (just give 30-days-notice after the 3rd month).
* It's cheap: $580/month
Disadvantages:
* people
* people
* people
OTOH, groups of people can choose to live together. Each good person I convince to room with me is one less disadvantage.
Q: if I accept the 4BR, do I lose my place in line for the studio?
A: as soon as I accept an offer, I may be placed on an internal transfer waitlist ("usually much shorter than waitlist from the outside").
Generic waittimes cited, for this list:
* 1BR: 4-6 months
* studio: 4-8 months
Since Thunderbird studios are normally ~9-12 months and I've already waited 2, this is not "much shorter", IMHO.
|
|