29 July 2016

A quick comment on structured input vs structured output learning

›
When I think of structured input models, I typically think of things like kernels over discrete input spaces. For instance, the famous all-s...
1 comment:
26 July 2016

Decoding (neural?) representations

›
I remember back in grad school days some subset of the field was thinking about the following question. I train an unsupervised HMM on some ...
6 comments:
12 July 2016

Some picks from NAACL 2016

›
Usual caveats: didn't see all talks, didn't read all papers, there's lot of good stuff at NAACL that isn't listed here! Tha...
05 July 2016

Rating the quality of reviews, after the fact

›
Groan groan groan reviewers are horrible people. Not you and me. Those other reviewers over there! tldr: In general we actually don't ...
3 comments:
30 June 2016

Subgradient descent, or something....

›
Last semester from grad ML , I totally revamped stuff and started using the awesome book Understanding Machine Learning by Shai Ben-David a...
6 comments:
24 June 2016

Language bias and black sheep

›
Tolga Bolukbasi and colleagues recently posted an article about bias in what is learned with word2vec , on the standard Google News crawl (h...
9 comments:
14 May 2016

A bad optimizer is not a good thing

›
A very popular style of research in NLP and ML is the math abstraction. You cast your learning problem as some sort of objective function th...
5 comments:
26 March 2016

A dagger by any other name: scheduled sampling

›
Scheduled Sampling was at NIPS last year; the reviews are also online. (Note: I did not review this paper.) This is actually the third tim...
3 comments:
27 December 2015

NIPS 2015 Retrospective

›
There have already been lots of really nice overviews of NIPS 2015. I don't have much to add, largely because I missed almost every talk...
1 comment:
30 October 2015

How sausage got made once

›
When I was working on what turned into an old short paper ( Markov Random Topic Fields ) I decided it might be pedagogically interesting to ...
20 October 2015

Results of NAACL 2015 post-conference survey are up

›
http://aclweb.org/adminwiki/index.php?title=2015Q3_Reports:_NAACL#UPDATE_20_OCT_2015.2C_NAACL_POST-CONFERENCE_SURVEY
05 October 2015

A small observation for prepubs on arxiv

›
The question of how "traditional conference publication" should react to arxiv prepublications is raised quite frequently. I'm...
11 September 2015

How long'll it take to say that?

›
tl;dr: Given a string of text in some language you might want to know how long it would take to speak it. Here are some perl / python ...
08 September 2015

Overfitting

›
Many students, in response to their assigned reading for today's undergrad ML class , asked me for a formal definition of overfitting. T...
28 comments:
09 June 2015

Some NAACL 2013 statistics on author response, review quality, etc.

›
NAACL 2015 has just passed, NAACL 2013 is long in the past. One bonus of being a program chair is that you get to have fun with data. In t...
3 comments: