Next Heavy Reading group

Next reading group will take place on:

11 December 2009

Presented by:

Oya Aran

Paper to read:

Gaussian Process Classification for Segmenting and Annotating Sequences
Altun, Hofmann & Smola
http://www.cs.brown.edu/~th/papers/AltHofSmo-ICML2004.pdf

This will be the last reading group of the year. We will continue the reading groups next year, on mid January, when everybody is back from holidays.

Next Heavy Reading Group

Next reading group will take place on:
Thurs. Dec 3rd at 11am

Presented by:
Kate

Paper to read:
Dynamic Topic Model:
www.cs.princeton.edu/~blei/papers/BleiLafferty2006a.pdf

Next light reading roup: The Combining Classifier: to Train or Not to Train?

Next reading group will take place on:

27 November 2009

Presented by:

Oya Aran

Paper to read:

The Combining Classifier: to Train or Not to Train? by R. Duin
/homes/oaran/Public/papers/duin02combining.pdf

Weighted Kappa: Nominal Scale Agreement with provision for scaled disagreement or partial credit Jacob Cohen, Psychological Bulletin,1968

Next Light reading group will take place on:

6th November

Presented by:

Hayley

Paper to read:

Weighted Kappa: Nominal Scale Agreement with provision for scaled disagreement or partial credit Jacob Cohen, Psychological Bulletin,1968
/home/userdirs/hhung/E000783444_1.pdf

This paper should be useful to a number of people who may be collecting any type of human annotations.

Suggestion for heavy reading group themes

In continuation of the CRF topic and also concurrently with the dynamic topic models paper we were reading, I suggest that we find a nice tutorial on gaussian processes,
or maybe even spice up the reading group by doing a 'watching group' for next time:
http://nips.cc/Conferences/2006/Program/event.php?ID=4
There's also a text book online by the presenter here:
http://www.gaussianprocess.org/gpml/chapters/

Following on from this, we can continue Kate's paper on dynamic topic models and after this combine what we've learnt from Gaussian processes with CRFS with this paper:
http://www.cs.brown.edu/~th/papers/AltHofSmo-ICML2004.pdf

I haven't read/seen any of this stuff yet so I have no idea whether it would be any good. However, they are by good people and so my gut reaction is that this would be quite a nice theme to continue.

Let me know what you think!

Next reading group will take place on:

[october 23th]

Presented by:

[Dayra]

Paper to read:

[/homes/dscortes/Documents/Papers/readingGroup/IEEEXplore.pdf]

Request for Reading group protocol

Following today when only 2 people turned up for the reading group, could everyone please register their interest (positive or negative) whenever a reading group is coming up? We had agreed that if people were ok to attend, they wouldn't reply to the post.. however, people are just not replying when they can't make it so it can feel a bit like talking to a brick wall.

I am not sure if anyone is actually subscribed to these blogs either so please reply to this post if you have read it!

Hayley

Next heavy reading group on CRFs

Next Heavy reading group will take place on:

[October 16th 2pm]

Presented by:

Hayley

Paper to read:

www.idiap.ch/~hhung/crf.pdf

Following our last heavy reading group conversation, I think this was the paper that was suggested...

Next Light Reading Group

Next reading group will take place on:
Thursday (I'm not here on Friday!) Oct. 1st - 11am

Presented by:
Kate

Paper to read:
NOTE: EXCLUDE section 3 to make it a light read!
www.cs.princeton.edu/~blei/papers/BleiLafferty2006a.pdf

Heavy Reading Group on CRF?

Are enough people here this Friday to read a paper? We have missed a few weeks already.

Choices for the next light reading group

Hi ppl,

Looks like i have to (unfortunately:-) do the next light reading group (4th Sep). But I have 3 choices, all of which i enjoyed reading.

(1) Overview paper on computational social science by pentland et al.,
(2) Statistical basis for performance evaluation.
(3) A classic paper on how and what can you measure?

Take a look at the files present in /homes/sparta/reading_group and comment on what you would prefer to read.

Thanks,
Hari

tagging and calendar reminder

Hi Everyone
Can you please make sure you always tag your posts and also if it's a reading group, please put it into the google calendar so that everyone can see when it's happening!

Heavy Reading Group on CRF

Next reading group will take place on:

[21. August 2009]

Presented by:

[Kate]

Paper to read:

[http://www.scai.fraunhofer.de/fileadmin/images/bio/data_mining/paper/crf_klinger_tomanek.pdf]

Lite Paper - Collective Modeling of Social Behavior

Next reading group will take place on:
14 th August, 2009

Presented by: Dinesh

Paper to read:
paper can be found at
http://www.aaai.org/Library/Symposia/Spring/2009/ss09-04-015.php

Poll for Heavy Reading Group

A poll was added to the blog, but it doesn't have the option to add new items or to see who voted what. So we'll still have to pick the next topic and presenter somehow. Maybe we'll have volunteers once we see what the most popular topic is.

Next "lite" reading group will take place on:

[24 july 2009]

Presented by:

[Radu]

Paper to read:

[http://www.google.com/url?sa=t&source=web&ct=res&cd=1&url=http%3A%2F%2Finfolab.stanford.edu%2F~mor%2Fresearch%2FNov_Naaman_Ye_ICWSM_2009_final.pdf&ei=xdheSuaBCeKgjAfNv93iDQ&usg=AFQjCNFinkjLfPwhdaaA5Kl-DQumv2RuHg&sig2=NQgGn6NpP9LEycA-7menHQ]

How to read a paper

Hey guys,

I found the following to be a very interesting short reading:

"How to read a paper" by S. Keshav, from University of Waterloo
http://blizzard.cs.uwaterloo.ca/keshav/home/Papers/data/07/paper-reading.pdf

Our new google calender

Not sure if this is going to work but I have created a google calender for the reading groups..

I haven't edited anything yet so please feel free to add the upcoming reading groups to the calender. I have added all the people from the blog owners list to the google calender.


hayley

Our new google calender

Not sure if this is going to work but I have created a google calender for the reading groups..

hayley

How to get your paper rejected

http://www.siggraph.org/publications/instructions/rejected

Makes an interesting read I think.

Machine Learning Reading Group (or Heavy Reading Group) Topics

Please add the topics you're interested in discussing for the 'Heavy' Reading Group then we can make a rough schedule.

The topics I'm interested in discussing are:

1. Conditional Random Fields
2. Sampling Techniques
3. Approximate Inference Techniques

Heavy Reading Group

Chapter 12 - Bishop till Kernel PCA
July 17th 2pm...

Next light reading group: Gokul

Hi everyone,
I'm posting this on behalf of Gokul

Next reading group will take place on:

[Either 19th or 26th June.... still to be decided. please post preferences here]

Presented by:

[Gokul]

Paper to read:

[www.idiap.ch/~hhung/IS07-enos-p1085-1.pdf]

Light Reading group order of presenters..

Hi,
I just thought I'd repost the ordered list of presenters for the light reading group. Please try to choose short papers of around 4 pages. The idea of this reading group is to present work that interests you and gives others a 'window into your world'...

paco,
gokul,
radu,
hari,
dinesh,
kate
dayra,
hayley,
nikhil,
oya,
jaggan

next heavy reading group

Next reading group will take place on:

[26 June, 2pm, 309]

We'll continue with Chris Bishop's book chapter 12, reading sections 12.2 and 12.3

Next normal reading group

Next reading group will take place on:

[June 19th]

Presented by:

[Paco]

Paper to read:

[Sketch-Based Retrieval of Drawings using Topological Proximity]
http://www.inesc-id.pt/ficheiros/publicacoes/4887.pdf

proposal for next heavy reading group

We were thinking of reading section 12.1 on principal component analysis from Chris Bishop's book next Tuesday(9th June) at 2pm. I can send out the pdf of the chapter.

Please reply to this post if you are interested in coming.

Hayley

Outcome of last reading group

I'll let Joan put up the new times since the ICA reading has been postponed.

So, this is what I think we came up with after much debate..
the 'lite' reading group will be run on a rotation system in this order, every two weeks starting from 5th June.

paco,
gokul,
radu,
hari,
dinesh,
kate
dayra,
hayley,
nikhil

The more in-depth reading group will start next week with a continuation of the discussion on ICA. Hopefully we can go on to read other stuff to do with clustering but this will be decided next week. The general consensus is that there are those who are interested in reading more in-depth tutorial-style papers but since this was 5 out of 9, the other (slightly louder) members decided in-depth papers would be a waste of time for them.

During the 'lite' reading circuit, there is no reason why you can't present a more in-depth paper. However, the general rule is to keep the paper short so everyone can read it in about 1 hour or 1.5 hours.

Hope that sorts everything out,
Hayley

Clustering series: ICA

Hayley proposed to start a series of readings on clustering algorithms.
We'll start reading about ICA in the next reading group.
The paper looks long but it should be easy to read. (If you have any comment
post it below this entry)

Next reading group will take place on:

Friday May 22nd, at 2pm in Room 405

Presented by:

Joan

Paper to read:
Independent Component Analysis: Algorithms and Applications
http://www.cs.helsinki.fi/u/ahyvarin/papers/NN00new.pdf

Internet Advertising

Next reading group will take place on:

Firday, May 8, 2009 at 2pm in Room 405

Presented by:

Nikhil Garg

Paper to read:
Internet Advertising and the Generalized Second-Price Auction: Selling Billions of Dollars Worth of Keywords
http://faculty-gsb.stanford.edu/ostrovsky/papers/gsp.pdf

Open for business

Hi all,

I think this will be the easiest way to keep track of the reading group stuff as suggested. Any time one of us will present a paper, he or she should post a message here, with something along the lines of:

Next reading group will take place on:

[insert date here]

Presented by:

[insert name here]

Paper to read:

[insert link here]

I suppose everyone has a Google account, and if not, create one and become the slaves of our all-knowing master!

As discussed today at lunch, first reading group is on Nikhil! (sorry man, you should have come with us to lunch!)

P.S. Colour theme dedicated to Paco