Vitor Carvalho - Home Page (Vitor Rocha de Carvalho) .............................................................................................................................................................................................................(frequently confused with Victor Carvalho)

Vitor R. Carvalho

contact:

I defended my PhD thesis in Carnegie Mellon in the Summer 2008, working under William Cohen. Currently I'm a scientist at Microsoft Bing. Before that, I was a scientist at Microsoft Live Labs.

I'm interested in applied research interfacing Machine Learning and Information Retrieval. In particular, I have worked on learning ranking functions, text mining, information extraction, classification models (stacked, online, collective, etc.), user recommendation models for email (speech acts, email leaks, recipient recommendation, etc.), and a handful of other interesting topics.

What's new:

Current Program Committees: EMNLP-09, IEEE CEC-09 (E3C), CEAS-09, IJCAI-09, ICML-09

Recent papers:

SIGIR 2009 , Reducing Long Queries Using Query Quality Predictors
CEAS 2009 , Information Leaks and Suggestions: a Case Study using Mozilla Thunderbird (accepted)
CIKM 2008 , Suppressing Outliers in Pairwise Preference Ranking
WS-08-04, Proceedings of the AAAI 2008 EMAIL Workshop
SIGIR-2008 LR4IR, A Meta-Learning Approach for Robust Rank Learning
AAAI-2008 EMAIL Workshop , CutOnce - Recipient Recommendation and Leak Detection in Action
ECIR-2008 , Ranking Users for Intelligent Message Addressing
WSDM-2008 , Fast Learning of Document Ranking Functions with the Committee Perceptron

All publications

Check out our new IR group blog....Probably Irrelevant
Now that I graduated, my "academic lineage" traces back all the way to Maxwell, Poisson, Lagrange, Bernoulli & Euler (thanks to William Cohen)

All Publications:

All Publications in chronological order.
Google Scholar entry: Vitor R. Carvalho
DBLP entry: Vitor R. Carvalho

Software:

Ciranda - Java package for email-speech-act prediction
Jangada - Java package for extraction of signatures (sig files) and reply-to (quotes) lines in email messages
Cut Once - A Mozilla Thunderbird plug-in for email information leak prevention and recipient recommendation
I contribute to Minorthird, a package for text learning, classification, extraction and annotations

Datasets:

617 messages from the 20 Newsgroups, annotated with reply bodies and signatures, used in the CEAS-2004 paper.

Other Stuff:

My " academic lineage" tracing back all the way to Leibniz, James Clerk Maxwell, Poisson, Lagrange, Bernoulli and Euler (compiled by my advisor William Cohen)
Recent program committees: COLING/ACL-06, AAAI-07, CEAS-05-06-07-08 , WWW-08
I organized, with Mark Dredze and Tessa Lau, EMAIL-2008: the AAAI-2008 Workshop on Enhanced Messaging
I was a TA in the Machine Learning (10-601) course during Fall 2007
I used to organize the CMU Machine Learning Lunch organization
I was the TA of the Information Extraction course (MLD 10-707 and LTI 11-748) during Spring 07
I used to organize the CMU Information Retrieval Discussion Series
If you're in Colorado and can't stand MPB (mountain pine beetles) my family had an idea.
Alma Matres: CMU-SCS-LTI, UNICAMP-FEEC, UFPE, Colegio Diocesano - Teresina
Check out our radio program on WRCT (88.3 FM) - Fridays, from 6 to 7 p.m. My last programs were: Torquato Neto (November), Tom Ze (April), Los Hermanos (March), Maracatu (November), Gilberto Gil (July), Paulinho da Viola (June), Cordel do Fogo Encantado, Luis Gonzaga...