You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mahout.apache.org by Grant Ingersoll <gs...@apache.org> on 2008/03/31 23:14:23 UTC

Fwd: New FF slides version

Message from Isabel, who is having server problems:

Begin forwarded message:
> From: "Isabel Drost" <is...@gmail.com>
> Date: March 31, 2008 5:07:00 PM EDT
> To: "Grant Ingersoll" <gs...@apache.org>
> Subject: New slides version
>
> Hello Grant,
>
> the provider hosting my web page currently has problems -
> I have put a new version of the slides online. This time at a server
> of another provider:
>
> http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf
>
> The "who we are" slide is still missing several photos, but the rest
> has improved quite a bit, I think. Would you be so kind as to forward
> the url to the list?
>
> Isabel (I urgently need to get the mailserver running at the strato
> server to be able to move my domain as quickly as possible... ;) )


Re: New FF slides version

Posted by Isabel Drost <ap...@isabel-drost.de>.
On Tuesday 01 April 2008, Stephan Seufert wrote:
> http://en.wikipedia.org/wiki/Buoyancy

Thanks :) I will update the slides tomorrow or on Friday.

Isabel

-- 
Unix is the worst operating system; except for all others.		-- Berry Kercheval
  |\      _,,,---,,_       Web:   <http://www.isabel-drost.de>
  /,`.-'`'    -.  ;-;;,_
 |,4-  ) )-,_..;\ (  `'-'
'---''(_/--'  `-'\_) (fL)  IM:  <xm...@spaceboyz.net>

Re: New FF slides version

Posted by Isabel Drost <ap...@isabel-drost.de>.
On Thursday 03 April 2008, Karl Wettin wrote:
> It doesn't. But these people were able to take one data set and make it
> in to something else that was useful for things nobody probably
> considered when the original data sources was created. Eureka! and that.

Ah. Now I see your point. I was thinking about one point that are still 
missing: One would be the one exact cool application that would not be 
possible without our project - in terms of scale and available algorithms. 
Something like a look into the future: What will the world look like in two 
years with Mahout having grown to full potential?


> > I a  kind of lost here - POI?
>
> Points of interests, pins on electronic maps.

Ah. Thanks.

Isabel


-- 
Life is difficult because it is non-linear.
  |\      _,,,---,,_       Web:   <http://www.isabel-drost.de>
  /,`.-'`'    -.  ;-;;,_
 |,4-  ) )-,_..;\ (  `'-'
'---''(_/--'  `-'\_) (fL)  IM:  <xm...@spaceboyz.net>

Re: New FF slides version

Posted by Karl Wettin <ka...@gmail.com>.
Isabel Drost skrev:
> On Wednesday 02 April 2008, Karl Wettin wrote:
>> I'm thinking of more or other data sources to put under "huge amouts of
>> data" that anybody might have access to but don't think about it.
>>
>> Like the WLAN geopositioning used by Apple products: there is this
>> company that drives around with cars and record WLAN signal strength
>> with lat/long. Think of all the cool things you can do with that data
>> and people reporting in where they are and how you figure out that a hub
>> has a new geoposition from only that and who knows what sort of big
>> brother related thingy.
> 
> I am sure that is a huge amount of data, I am not sure which tasks machine 
> learning could solve in this context. Maybe for finding people who have 
> visited the same places as I have?

It doesn't. But these people were able to take one data set and make it 
in to something else that was useful for things nobody probably 
considered when the original data sources was created. Eureka! and that.

I'm not saying this is whay you should be saying. Just that I'm trying 
to think of data sources people don't think of, but they are not so easy 
to come up with. This was an example of such a source.

> By the way, there are even phones that have WLan as well as GPS (just think of 
> the OpenMoko free software based smart phone). I think there were even 
> navigation devices (?) that contain both, GPS and WLan chips.
> 
> 
>> At least there might be room left to squeeze in POIs?
> 
> I a  kind of lost here - POI?

Points of interests, pins on electronic maps.



     karl

Re: New FF slides version

Posted by Isabel Drost <ap...@isabel-drost.de>.
On Wednesday 02 April 2008, Karl Wettin wrote:
> I'm thinking of more or other data sources to put under "huge amouts of
> data" that anybody might have access to but don't think about it.
>
> Like the WLAN geopositioning used by Apple products: there is this
> company that drives around with cars and record WLAN signal strength
> with lat/long. Think of all the cool things you can do with that data
> and people reporting in where they are and how you figure out that a hub
> has a new geoposition from only that and who knows what sort of big
> brother related thingy.

I am sure that is a huge amount of data, I am not sure which tasks machine 
learning could solve in this context. Maybe for finding people who have 
visited the same places as I have?

By the way, there are even phones that have WLan as well as GPS (just think of 
the OpenMoko free software based smart phone). I think there were even 
navigation devices (?) that contain both, GPS and WLan chips.


> At least there might be room left to squeeze in POIs?

I a  kind of lost here - POI?

Isabel

-- 
Flon's Law:	There is not now, and never will be, a language in	which it is the 
least bit difficult to write bad programs.
  |\      _,,,---,,_       Web:   <http://www.isabel-drost.de>
  /,`.-'`'    -.  ;-;;,_
 |,4-  ) )-,_..;\ (  `'-'
'---''(_/--'  `-'\_) (fL)  IM:  <xm...@spaceboyz.net>

Re: New FF slides version

Posted by Karl Wettin <ka...@gmail.com>.
Grant Ingersoll skrev:
 >On Apr 1, 2008, at 2:24 PM, Isabel Drost wrote:

 >> http://www.isabel-drost.de/mahout_fast_feather.pdf
 >> http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf
 >>
 >> (Both links should be the same document.)
 >>
> Very cool!  Thanks for taking this up Isabel.

Indeed!

 > You might

 > I don't know

I'm thinking of more or other data sources to put under "huge amouts of 
data" that anybody might have access to but don't think about it.

Like the WLAN geopositioning used by Apple products: there is this 
company that drives around with cars and record WLAN signal strength 
with lat/long. Think of all the cool things you can do with that data 
and people reporting in where they are and how you figure out that a hub 
has a new geoposition from only that and who knows what sort of big 
brother related thingy.

But there is no one good word for it. I'm sure you all understand what I 
mean. Semantic circles? I have no idea. Banerry.

At least there might be room left to squeeze in POIs?



      karl

Re: New FF slides version

Posted by Grant Ingersoll <gs...@apache.org>.
Very cool!  Thanks for taking this up Isabel.

You might consider making the "Where is ML used" sooner, up towards  
the front.

I don't know that I would "name names" on the GSOC page, as we don't  
know who will be selected.

Under initial contributions, we also have the Canopy Clustering!

Also, not sure when your talk is, but _if_ it is before the Mahout  
BOF, you should mention that.

-Grant

On Apr 1, 2008, at 2:24 PM, Isabel Drost wrote:

> This time the message I am able to send the message myself. The  
> updated
> version of the slides is online. If you have any further comments -  
> let me
> know.
>
> The only piece missing is the formula/model that Archimedes  
> discovered sitting
> in his bath tub just before the moment in history when he (supposedly)
> shouted Eureka! and ran through the city - naked. Anyone who can  
> help me out
> there?
>
> The URLs again who have missed it:
> http://www.isabel-drost.de/mahout_fast_feather.pdf
> http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf
>
> (Both links should be the same document.)
>
> Isabel
>
> -- 
> "You can't get very far in this world without your dossier being there
> first."		-- Arthur Miller
>  |\      _,,,---,,_       Web:   <http://www.isabel-drost.de>
>  /,`.-'`'    -.  ;-;;,_
> |,4-  ) )-,_..;\ (  `'-'
> '---''(_/--'  `-'\_) (fL)  IM:  <xm...@spaceboyz.net>



Re: New FF slides version

Posted by Stephan Seufert <se...@pydev.org>.
Hello Isabel,

http://en.wikipedia.org/wiki/Buoyancy

:-)

Kind regards,
steps



Am 01.04.2008 um 20:24 schrieb Isabel Drost:
> This time the message I am able to send the message myself. The  
> updated
> version of the slides is online. If you have any further comments -  
> let me
> know.
>
> The only piece missing is the formula/model that Archimedes  
> discovered sitting
> in his bath tub just before the moment in history when he (supposedly)
> shouted Eureka! and ran through the city - naked. Anyone who can  
> help me out
> there?
>
> The URLs again who have missed it:
> http://www.isabel-drost.de/mahout_fast_feather.pdf
> http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf
>
> (Both links should be the same document.)
>
> Isabel
>
> -- 
> "You can't get very far in this world without your dossier being there
> first."		-- Arthur Miller
>  |\      _,,,---,,_       Web:   <http://www.isabel-drost.de>
>  /,`.-'`'    -.  ;-;;,_
> |,4-  ) )-,_..;\ (  `'-'
> '---''(_/--'  `-'\_) (fL)  IM:  <xm...@spaceboyz.net>


Re: New FF slides version

Posted by Ted Dunning <td...@veoh.com>.
>From the wikipedia as well:

Another famous mathematician, Carl Friedrich Gauss, echoed Archimedes when
in 1796 he wrote in his notebook, "ΕΥΡΗΚΑ! num= Δ + Δ + Δ", referring to his
discovery that any positive integer could be expressed as the sum of at most
three triangular numbers.[3]


On 4/1/08 11:24 AM, "Isabel Drost" <ap...@isabel-drost.de> wrote:

> This time the message I am able to send the message myself. The updated
> version of the slides is online. If you have any further comments - let me
> know.
> 
> The only piece missing is the formula/model that Archimedes discovered sitting
> in his bath tub just before the moment in history when he (supposedly)
> shouted Eureka! and ran through the city - naked. Anyone who can help me out
> there?
> 
> The URLs again who have missed it:
> http://www.isabel-drost.de/mahout_fast_feather.pdf
> http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf
> 
> (Both links should be the same document.)
> 
> Isabel


Re: Fwd: New FF slides version

Posted by Isabel Drost <ap...@isabel-drost.de>.
This time the message I am able to send the message myself. The updated 
version of the slides is online. If you have any further comments - let me 
know.

The only piece missing is the formula/model that Archimedes discovered sitting 
in his bath tub just before the moment in history when he (supposedly) 
shouted Eureka! and ran through the city - naked. Anyone who can help me out 
there?

The URLs again who have missed it:
http://www.isabel-drost.de/mahout_fast_feather.pdf
http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf

(Both links should be the same document.)

Isabel

-- 
"You can't get very far in this world without your dossier being there 
first."		-- Arthur Miller
  |\      _,,,---,,_       Web:   <http://www.isabel-drost.de>
  /,`.-'`'    -.  ;-;;,_
 |,4-  ) )-,_..;\ (  `'-'
'---''(_/--'  `-'\_) (fL)  IM:  <xm...@spaceboyz.net>

Re: Fwd: New FF slides version

Posted by Isabel Drost <ap...@isabel-drost.de>.
On Monday 31 March 2008, Karl Wettin wrote:
> > From: "Isabel Drost" <is...@gmail.com>
> >> http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf
> >>
>  >> We need you – the Java developer with
>  >> – A good part of enthusiasm.
>  >> – Solid mathematical knowledge to understand ML papers.
>  >>– Either proficient in or willing to learn about Hadoop.
>
> I also want people that show up on the users list and ask questions
> about what ML stuff they can do with their data.

+1 I will add that to the next version. (To appear on the URL above again).

Isabel


-- 
"The only way for a reporter to look at a politician is down."		-- H. L. 
Mencken
  |\      _,,,---,,_       Web:   <http://www.isabel-drost.de>
  /,`.-'`'    -.  ;-;;,_
 |,4-  ) )-,_..;\ (  `'-'
'---''(_/--'  `-'\_) (fL)  IM:  <xm...@spaceboyz.net>

Re: Fwd: New FF slides version

Posted by Karl Wettin <ka...@gmail.com>.
> From: "Isabel Drost" <is...@gmail.com>
>> http://h1349259.stratoserver.net/~isabel/mahout_fast_feather.pdf
 >>
 >> We need you – the Java developer with
 >> – A good part of enthusiasm.
 >> – Solid mathematical knowledge to understand ML papers.
 >>– Either proficient in or willing to learn about Hadoop.

I also want people that show up on the users list and ask questions 
about what ML stuff they can do with their data.


     karl