[CL+NLP Lunch] [SHORT NOTICE] CL+NLP lunch, '11am' Oct.26 Monday

[Kazuya Kawakami]

Hi All

The time should be *11am*.
We will have ML lunch talk by Leonid's talk at noon.
Sorry.

>>
Please join us for the next CL+NLP lunch at *11am on Monday Oct 26th at
7101*,
where Jiang Guo will be speaking about Cross-lingual Transfer Parsing.
Lunch will be provided!

To arrange meetings with Jiang, please see the following document.
(
https://docs.google.com/document/d/1i4s181AWQGY1SJup76BZGgsjpquAH2ZrVe7tHWISPYc/edit
)


-----------------------------------------
ML+NLP lunch

*Monday, Oct 26th at 11:00amGHC 7101*

Speaker: Jiang Guo, Johns Hopkins University

[TITLE]
Representation Learning for Cross-lingual Transfer Parsing

[ABSTRACT]
Cross-lingual model transfer has been a promising approach for inducing
dependency parsers for low-resource languages where annotated treebanks are
not available. The major obstacles for the model transfer approach are
two-fold:

1. Lexical features are not directly transferable across languages.

2. Target language-specific syntactic structures are difficult to be
recovered.

In this talk, I will provide a representation learning framework to address
these challenges. By evaluating on the Google universal dependency
treebanks (v2.0), our best models yield an absolute improvement of 6.53% in
averaged labeled attachment score, as compared with delexicalized
multi-source transfer models. We also significantly outperform the
state-of-the-art transfer system proposed most recently.


[SHORT BIO]
Jiang Guo is a joint Ph.D student at Johns Hopkins University and at Harbin
Institute of Technology. His research interests are in the areas of natural
language processing, machine learning, with special interests on
distributed representation learning and its applications on NLP tasks
(mostly structure prediction problems). His long-term goal is developing
efficient and effective algorithms and softwares for NLP and machine
learning applications.

-----------------------------------------


Up comming talk will be on 17th Nov. 12:00-13:00 by Ndapa Nakashole.

Best regards,
Kazuya

On Fri, Oct 23, 2015 at 4:39 PM, Kazuya Kawakami <
www.kazuya.kawakami at gmail.com> wrote:

> Hi Dallas,
>
> Actually it's a sudden talk. We scheduled it last night ;)
> I have never used cs.cmu.edu, but it should be kkawakam.
>
> Best regards,
> Kazuya
>
> On Fri, Oct 23, 2015 at 4:36 PM, Dallas Card <dallas.card at gmail.com>
> wrote:
>
>> Hey Kazuya,
>>
>> Thanks for organizing this! I just tried adding you as an administrator
>> for this list. Is your cs.cmu.edu address also kkawakam?
>>
>> Thanks,
>> Dallas
>>
>> On Fri, Oct 23, 2015 at 4:30 PM, Kazuya Kawakami <kkawakam at andrew.cmu.edu
>> > wrote:
>>
>>> Hi All
>>>
>>> Please join us for the next CL+NLP lunch at noon on Monday Oct 26th,
>>> where Jiang Guo will be speaking about Cross-lingual Transfer Parsing.
>>> Lunch will be provided!
>>>
>>> To arrange meetings with Jiang, please see the following document.
>>> (
>>> https://docs.google.com/document/d/1i4s181AWQGY1SJup76BZGgsjpquAH2ZrVe7tHWISPYc/edit
>>> )
>>>
>>>
>>> -----------------------------------------
>>> ML+NLP lunch
>>> Monday, Oct 26th at 11:00am
>>> GHC 7101
>>>
>>> Speaker: Jiang Guo, Johns Hopkins University
>>>
>>> [TITLE]
>>> Representation Learning for Cross-lingual Transfer Parsing
>>>
>>> [ABSTRACT]
>>> Cross-lingual model transfer has been a promising approach for inducing
>>> dependency parsers for low-resource languages where annotated treebanks are
>>> not available. The major obstacles for the model transfer approach are
>>> two-fold:
>>>
>>> 1. Lexical features are not directly transferable across languages.
>>>
>>> 2. Target language-specific syntactic structures are difficult to be
>>> recovered.
>>>
>>> In this talk, I will provide a representation learning framework to
>>> address these challenges. By evaluating on the Google universal dependency
>>> treebanks (v2.0), our best models yield an absolute improvement of 6.53% in
>>> averaged labeled attachment score, as compared with delexicalized
>>> multi-source transfer models. We also significantly outperform the
>>> state-of-the-art transfer system proposed most recently.
>>>
>>>
>>> [SHORT BIO]
>>> Jiang Guo is a joint Ph.D student at Johns Hopkins University and at
>>> Harbin Institute of Technology. His research interests are in the areas of
>>> natural language processing, machine learning, with special interests on
>>> distributed representation learning and its applications on NLP tasks
>>> (mostly structure prediction problems). His long-term goal is developing
>>> efficient and effective algorithms and softwares for NLP and machine
>>> learning applications.
>>>
>>> -----------------------------------------
>>>
>>>
>>> Up comming talk will be on 17th Nov. 12:00-13:00 by Ndapa Nakashole.
>>>
>>> Best regards,
>>> Kazuya
>>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/nlp-lunch/attachments/20151023/c4628396/attachment.html>