Restructure class?

[Kanoelani Pilobello]

I passed by Bridge Street today around 5:10, but didn't see a sign on the door.  I waited a bit, but it was cold.  Maybe I was at the wrong location?  Sorry, if I missed you.

Here's my suggestion for restructuring the class if anyone is still interested.  We can each make 15-20 minute presentations for a selection of the remaining lectures.  We pick one related problem or exercise and an interesting example to help summarize the lecture. The problems are emailed a week before the presentations and people can try to glean from the course material whatever they wish.  We can do 3-4 presentations a week for 2 more weeks.

I that might be more fun and more rewarding.  Personally, I learn a lot when I'm trying to teach something and it's a valuable skill.

I think it would be a shame to quit now.  I know we had some rough weeks, but i think this is a great opportunity to experiment with the format and to build a core group for future attempts.

Rally! Rally! Rally!

Lani

On Sat, Jan 30, 2010 at 7:33 PM, Kelly Farrell <kel...@gmail.com> wrote:

That sounds like it could be a lot of fun. We can also talk about
whether we want to keep slogging through the lectures, skip around to
ones we like, or decide we've had enough of Prof. Boringpants and TA
Spaz.

Kelly Farrell
http://www.everythingtiny.com/

On Sat, Jan 30, 2010 at 7:16 PM, Kanoelani Pilobello <ktp...@gmail.com> wrote:
> I am seriously behind too, but I'm trying to catch up by tomorrow.
> Would anyone be interested in doing an algorithms show and tell at the
> beginning of class just to make things interesting?  We could just showcase
> an algorithm/something related to algorithms that we know of or is related
> to our work?
> I bring this up because I saw a lecture on a cool microorganism.  I keep
> meaning to send it out in a email, but I might as well present it.
>
> Lani
> On Sat, Jan 30, 2010 at 3:49 PM, Kelly Farrell <kel...@gmail.com> wrote:
>>
>> So, admittedly I have been slacking a bit... the last lecture I've
>> watched is 12, but we can watch whatever the next interesting looking
>> lecture is this week. What's everyone else up to?
>>
>> Kelly Farrell
>> http://www.everythingtiny.com/
>>
>>
>>
>> On Fri, Jan 29, 2010 at 3:20 PM, Kanoelani Pilobello <ktp...@gmail.com>
>> wrote:
>> > Hi,
>> > I had some flight delays coming back from vacation last Sunday.  I just
>> > wanted to check in and see if you were still meeting this coming Sunday
>> > and
>> > what lecture we are on.  17 or 18?
>> >
>> > Thanks,
>> > Lani
>> >
>> >
>> > On Sun, Jan 24, 2010 at 9:58 AM, Kellbot <kel...@gmail.com> wrote:
>> >>
>> >> Hey folks,
>> >> Just checking in to see if people will be making it to algorithms
>> >> today. We can pick up wherever folks have left off.
>> >>
>> >> We're still in the Bridge St space, which hasn't been quite so frigid
>> >> lately. Hooray!
>> >>
>> >> -Kelly
>> >
>> >
>
>

[Adam Mayer]

Oh, crap! Sorry, Lani. We were there, and Kelly did put the sign
out, but she may have been a little late. Next time we should mail out
phone numbers to people on the list before the class.

We ended up reading the lecture notes for skip lists, and then going
over it together, and that worked pretty well-- it was a lot more fun
than watching a lecture, and we got through the material in half the
time. We read the lecture notes for the amortized analysis lecture
after that, but agreed that the textbook or lecture would probably do
a better job of explaining it.

I really, really like the presentation idea, although we might want to
stick with 2-3 lectures as week, since we're finally starting to get
to fun stuff (graph algorithms, hooray!). That would take us through
early March.

Scheduling: there are a lot of scheduling collisions in Februrary
between algorithms and NYCR. On the weekend of the 13th-14th, we're
doing our 48 hour hackathon (because we are dumb nerds, and forgot
when valentines day was until long after we scheduled it), so the
space will probably be a chaotic shambles by the end of that (and
Kelly and I will probably be walking wounded). On the weekend of the
27th-28th, NYCR will be moving to its new space, and the evening of
the 28th will be Ranjit's lantern festival. We'll probably have to
reschedule or relocate both of those sessions. (Y'all are invited to
the hackathon and lantern festival, which should both be awesome.)

The good news about the move is that NYCR will have about three times
as much space, in separate rooms, which means we'll be able to run a
recurring class like this much more easily. The other good news about
the move is that it's a fourth-floor walkup, which means we all get to
develop awesome calves!

-a

Also: Resistor is going to be pretty chaotic

[Kelly Farrell]

Sorry about the sign, I was indeed late in putting it out.

Between the freezing weather / lack of heat, moving hell, and a bit of
dwindling enthusiasm I suggested we take February off to thaw out and
reboot in March with a format more like what we did this week.
Presentations sound awesome, and I think if we switched to reading the
notes / watching the videos at home and then coming together to talk
about it, instead of the other way around, it would be a lot more fun
overall.

If folks still want to meet in February I'm all for it, although
between preparing for the hackathon and moving I'm not sure if I'll be
able to make all of them. On the flip side, if we're not watching the
videos when we get together we're not as limited to where we can meet
(not needing a projector/wall). So it's up to you guys, but as Adam
said I'm going to be walking dead for most of February.

Kelly Farrell
http://www.everythingtiny.com/

[Kanoelani Pilobello]

It does sound like February might be chaotic.  I'm cool with taking February off and starting again in March.

I think I might try to do some of the exercises or maybe a mini-project at home.  Would anyone like to join me and discuss online?  

Lani

[Adam Mayer]

Just to confirm-- the consensus is restart in March, so there's no class today.

I'm up for exercises or projects. -a

[Serena Wales]

I'd also be down for exercise or projects- I think it would be a good way to not lose momentum altogether over the next few weeks.

Do people want to work through the exercises from the classes or book, or are there other interesting things we could work on?

Serena

[Kanoelani Pilobello]

I haven't looked at the opencourseware exercises in a while, but here are a couple sites with some project ideas.

http://www.inf.unibz.it/dis/teaching/ATA/#Project Proposals

http://www.cs.uu.nl/docs/vakken/an/

Project 7 on the first site has to do with the stable marriage problem, which I think would be fun and holiday appropriate.

Another fun an potentially useful thing might be to do a project on implementing the Traveling Salesman Problem.  I volunteer for RightRides (http://www.rightrides.org/), an organization that gives free rides home to women, the lgbtq, and gender queer on Friday and Saturday nights.  So people call in at different times and give their current location and home address.  The goal is to take home as many people as possible and currently the volunteer numbers are dwindling.  Sometimes you can take more than one ride home at a time, but the order of pick up and drop off is up to the navigator.  Sometimes that navigator is me and I totally don't know the city well enough to make the best decision.  This is completely unsolicited on their part, but I think it would be fun to apply to a real world situation and maybe we could bring it up with the director after we're done.

Lani

[Pat Farrell]

On Tue, Feb 9, 2010 at 11:16 PM, Kanoelani Pilobello <ktp...@gmail.com> wrote:

Another fun an potentially useful thing might be to do a project on implementing the Traveling Salesman Problem.  

You do know that TSP is officially NP-Hard, right?

[Max]

Exactly why it would be interesting. I've seen smart implementations
which run the algorithm once and store the solutions for lookup.
Geography changes slowly, thankfully.

[Adam Mayer]

There are also some approximations that are guaranteed to come within
a guaranteed bound of the best solution in polynomial time. Companies
that build networks spend millions of dollars researching this stuff.
It's pretty neat. -a

[Pat Farrell]

On Wed, Feb 10, 2010 at 4:48 PM, Adam Mayer <pho...@gmail.com> wrote:

There are also some approximations that are guaranteed to come within
a guaranteed bound of the best solution in polynomial time.  Companies
that build networks spend millions of dollars researching this stuff.
It's pretty neat. 

Yes, it is neat. actually, all real implementations are just approximations. That is the essence of the definition o NP-Hard. Something is NP if any algorithm to verify the correctness of a solution is non-polynormial -- not that the algorithm to come up with the possible solution is non-polynomial. The difference is subtle, but fundamental.

Airlines are one example of commercial applications that cost millions of dollars, and have for many years. TSP is easy as long as N is small, but for a big commercial airline, such as United or American, N is too large for exact solutions.