Robert Staubs goes to Mitre

Congratulations to Robert Staubs, who has been hired as a Senior Data Scientist by Mitre Corporation.

Staubs goes to MITRE

Robert Staubs has recently started a position at MITRE as a senior data scientist. MITRE is a not-for-profit organization that exclusively does work in the public interest for government clients. Congratulations Robert!

Masters in Computational Linguistics at Rochester University

The Computational Linguistics MS program at Rochester trains students to be conversant both in the analysis of language and in computational techniques applied to natural language. The curriculum consists of courses in Linguistics and Computer Science for a total of 32 credit hours.

Graduates from the Computational Linguistics MS program will be prepared for both further training at the PhD level in Computer Science and Linguistics and Computational Linguistics positions in industry. A growing number companies such as Google, Amazon.com, Nuance, LexisNexis, Oracle and many others are searching for employees with advanced degrees in Computational Linguistics for positions ranging from speech recognition technology to improving translation systems to developing bettermodels of language understanding. 

For fullest consideration, applicants should submit completed applications by March 1, though applications will continue to be considered as space permits. All documents and application materials should be uploaded in the online application (https://apply.grad.rochester.edu/apply/). Fees for applications received by February 15 waived (more information: https://apply.grad.rochester.edu/register/ASEwaiver). 

Direct questions to jeffrey.runner@rochester.edu

For more information about the program, go here.

Sam Bowman talks at Computer Science

Sam Bowman of Stanford University will be giving the talk “Modeling Natural Language Semantics with Learned Representations” on Thursday, February 11, at 4PM in CS room 151. An abstract of his talk follows

The last few years have seen many striking successes from artificial neural network models on hard natural language processing tasks. These models replace complex hand-engineered systems for extracting and representing the meanings of sentences with learned functions that construct and use their own internal vector-based representations. Though these learned representations are effective in many domains, they aren't interpretable in familiar terms and their ability to capture the full range of meanings expressible in natural language is not yet well understood.

In this talk, I argue that neural network models are capable of learning to represent and reason with the meanings of sentences. First, I use entailment experiments over artificial languages to show that existing models can learn to reason logically over clean language-like data. I then introduce a large new corpus of entailments in English and use experiments on that corpus to show that these abilities extend to natural language as well. Finally, I briefly present ongoing work on a new model that uses the semantic principle of compositionality to more efficiently and effectively learn to understand natural language.

Howard Karloff speaks in Computer Science

Variable Selection is Hard

Who: Howard Karloff

When: Feb 9, 1-2pm

Where: Computer Science Building, room 150/151

Website: https://www.cics.umass.edu/event/theory-speaker-variable-selection-hard 

Abstract:

Consider the task of a machine-learning system faced with voluminous data on m individuals.  There may be p=10^6 features describing each individual.  How can the algorithm find a small set of features that “best” describes the individuals?  People usually seek small feature sets both because models with small feature sets are understandable and because simple models usually generalize better.

We study the simple case of linear regression, in which a user has an m x p matrix B and a vector y, and seeks a p-vector x *with as few nonzeroes as possible* such that Bx is approximately equal to y, and we call it SPARSE REGRESSION.  There are numerous algorithms in the statistical literature for SPARSE REGRESSION, such as Forward Selection, Backward Elimination, LASSO, and Ridge Regression.
We give a general hardness proof that (subject to a complexity assumption) no polynomial-time algorithm can give good performance (in the worst case) for SPARSE REGRESSION, even if it is allowed to include more variables than necessary, and even if it need only find an x such that Bx is relatively far from y.
This is joint work with Dean Foster and Justin Thaler  and was done when all coauthors were at Yahoo Labs.

Bio: After receiving a PhD from UC Berkeley, Howard Karloff taught at the University of Chicago and Georgia Tech before leaving Georgia Tech  to join AT&T Labs–Research in 1999.  He left ATT Labs in 2013 to join Yahoo Labs in New York, where he stayed till February, 2015. Now he does data science for Goldman Sachs in New York.

A fellow of the ACM, he has served on the program committees of numerous conferences and chaired the 1998 SODA program committee. He is the author of numerous journal and conference articles and  the Birkhauser book “Linear Programming.”  His interests include  data science, machine learning, algorithms, and optimization.

Postdoc position in Computational Linguistics at Tübingen

The Department of  Linguistics at the University  of Tübingen, Germanyinvites applications for  a PhD (or Postdoc) position in  the group ofProf. Dr. Detmar Meurers: http://sfs.uni-tuebingen.de/en/tcl

This is a  university position, with an initial  period of appointmentof  three years,  with the  possibility of  renewal for  another three years. The  position  involves  independent  research  and  teaching (2hrs/week during the semester)  and related departmental duties. The degree programs in computational linguistics are taught in English, so knowledge of German is not a prerequisite. The position is at the rank of Wissenschaftlicher Mitarbeiter  (German scale TV-L E13  50% = 19.75hrs/week), for  which salary starts  at € 22.000 per  year, increasingwith experience and marital status.

Ideal candidates  have an  excellent Master's degree  in Computational Linguistics or a related field and  link to the strengths of the group and the excellent research  context in Tübingen. Candidates  with an interdisciplinary interest in Second  Language Learning, Education, or Psychology are particularly encouraged to  apply and may be associated with  the  Excellence Initiative  LEAD  graduate  school in  Empirical Educational Science: http://purl.org/lead/4

Applications should include:-

Cover letter spelling out research experience & interest in position

- Academic CV

- Copy of MA degree and transcript

- Names and emails of two referees

Send applications by email as a single pdf attachment to:

Lisa Becker <sekretariat-meurers@sfs.uni-tuebingen.de>

For questions, please email Detmar Meurers <detmar.meurers@uni-tuebingen.de>

Applications  received   by  December  22,  2015   will  receive  fullconsideration,  although interviews  may start  at any  time and  will continue until the position has been filled.

The University of Tübingen is  an equal opportunity employer regarding professional  opportunities  for  women and  men.   Equally  qualified candidates with physical challenges will be given  preference. Employment will be conducted by the central university administration. The NLP CALL list is housed at the University of Waterloo (Canada) and administrated by Mat Schulze (mschulze &#228; uwaterloo.ca). See http://artsservices.uwaterloo.ca/mailman/listinfo/nlp call for more information.

Cog-Sci Workshop this Thursday

Join us on Thursday, November 13th from 2:00 until 5:30 in Computer Science 150/151 to learn about and contribute to Cognitive Science efforts on the UMass Amherst campus. The Computer Science building is at the far north end of campus, and there are metered parking spots available directly across the street. A schedule of events follows.

2:00  Social 15 minutes to meet others interested in Cognitive Science and to see a short video introducing the new website

2:15  Lisa Sanders (Co-Director Cognitive Science Initiative) will outline our current plans to become an Institute of Cognitive Science and introduce John McCarthy

2:20  John McCarthy (Vice Provost for Graduate Education and Dean of the Graduate School) will talk about interdisciplinary research and graduate education

2:30  Andy Barto (Computer Science) Computational Clues to the Brain’s Reward System

2:50  Louise Antony (Philosophy) Epistemology and Psychology: Can Justification be “Naturalized”?

3:10  Caren Rotello (Cognitive Psychology) Why Cognitive Psychology is Important for Neuroscience: An Example from Research on Reasoning

3:30  Poster session and social hour

4:30  Rajesh Bhatt (Linguistics) The Importance of Treebanks in Cognitive Science

4:50  Erik Cheries (Developmental Psychology) Foundations of Mind: Infants’ Knowledge of Objects, Agents, & Identity

5:10  Dave Huber (Cognitive Neuroscience) Testing a Perceptual Habituation Model with Electrophysiology

Call for papers: Computational Phonology Workshop

The implications of computation and learnability for phonological theory

Crystallizing a feeling widespread in the field, Mark van Oostendorp (2013) writes: ''It is fair to say that we know much more about sound patterns in human language than people did at the beginning of the 20th Century. At the same time, many phonologists seem to feel that we have not yet reached the standards of some of the ‘hard’ sciences.'' An obstacle to progress seems to be the fact that competing phonological theories are underdetermined by sheer typological and linguistic data. Moving beyond descriptive adequacy, Alan Prince thus proposes that ''rational arguments about two theories’ comparative success […] depend on a broad assessment of their properties.'' Among the formal properties of a phonological theory which are crucial for its comparative assessment are its computability and learnability properties. 

This workshop thus aims at investigating the implications of computation and learnability for phonological theory. The issues addressed include (but are not limited to): the computability/intractability of phonological grammars and the debate among derivational, representational, and constraint-based frameworks; learnability guarantees and the debate between competing modes of constraint interaction; the characterization of phonological patterns within the sub-regular hierarchy and the expressive power of phonological formalisms; the learnability filter and its implications for the evaluation of the typologies predicted by competing phonological theories; methods for constraint induction and the phonetic grounding of the phonological constraints; statistical methods, probabilistic grammars and the divide between categorical and gradient models of phonological competence; the learnability of phonological processes conditioned by prosodic domains and its implications for the syntax/phonology interface. The workshop adopts an inclusive perspective, open to any computational approach and any phonological framework. 

Invited speakers:

Paul Boersma (University of Amsterdam)

Bruce Tesar (Rutgers University)

Important dates:

abstract submission deadline: December 1st, 2014

notification of acceptance: February 15th, 2015

workshop date: April 18th, 2015

Submission:

We invite abstracts for a 20 minute oral presentation followed by a 10 minute discussion. Abstracts should be anonymous and should not exceed 2 pages in length (A4 or letter-size, in 12 pt. font, with 1-inch/2.5-cm margins), including examples and references. The language of the workshop is English. Abstracts should be submitted through the GLOW 38 Easychair page (https://www.easychair.org/conferences/?conf=glow38), specifying that the submission should be considered for the workshop.

Website: https://sites.google.com/site/2015glow/home

Organizers:

Joaquim Brandao de Carvalho,

Mohamed Lahrouchi,

Giorgio Magri,

Michela Russo

Brendan O'Connor on text analysis today

Brendan O'Connor, CMU, will give a talk today, February 10, at
4 PM in Computer Science Building 150 entitled “Statistical Text Analysis for Social Science.” An abstract follows.

What can text analysis tell us about society? Corpora of news, books, and social media encode human beliefs and culture. But it is impossible for a researcher to read all of today's rapidly growing text archives. My research develops statistical text analysis methods that measure social phenomena from textual content, especially in news and social media data. For example: How do changes to public opinion appear in microblogs? What topics get censored in the Chinese Internet? What character archetypes recur in movie plots? How do geography and ethnicity affect the diffusion of new language? In order to answer these questions effectively, we must apply and develop scientific methods in statistics, computation, and linguistics.

In this talk I will illustrate these methods in a project that
analyzes events in international politics. Political scientists are
interested in studying international relations through *event data*:
time series records of who did what to whom, as described in news
articles. To address this event extraction problem, we develop an
unsupervised Bayesian model of semantic event classes, which learns the verbs and textual descriptions that correspond to types of
diplomatic and military interactions between countries. The model
uses dynamic logistic normal priors to drive the learning of semantic
classes; but unlike a topic model, it leverages deeper linguistic
analysis of syntactic argument structure. Using a corpus of several
million news articles over 15 years, we quantitatively evaluate how
well its event types match ones defined by experts in previous work,
and how well its inferences about countries correspond to real-world
conflict. The method also supports exploratory analysis; for example,
of the recent history of Israeli-Palestinian relations.

Second Call: Conference on Formal Grammar

The 19th Conference on Formal Grammar

Tuebingen, Germany, August 16-17 2014

http://fg.phil.hhu.de/

Collocated with the European Summer School in Logic, Language and Information

Background

FG-2014 is the 19th conference on Formal Grammar, to be held in conjunction with the European Summer School in Logic, Language and Information, which takes place in 2014 in Tuebingen, Germany.

Previous Formal Grammar meetings were held in Barcelona (1995), Prague (1996), Aix-en-Provence (1997), Saarbruecken (1998), Utrecht (1999), Helsinki (2001), Trento (2002), Vienna (2003), Nancy (2004), Edinburgh (2005), Malaga (2006), Dublin (2007), Hamburg (2008), Bordeaux (2009), Copenhagen (2010), Ljubljana (2011), Opole (2012) and Duesseldorf (2013).

Aims and Scope

FG provides a forum for the presentation of new and original research on formal grammar, mathematical linguistics and the application of formal and mathematical methods to the study of natural language. Themes of interest include, but are not limited to,

* formal and computational phonology, morphology, syntax, semantics and pragmatics;

* model-theoretic and proof-theoretic methods in linguistics;

* logical aspects of linguistic structure;

* constraint-based and resource-sensitive approaches to grammar;

* learnability of formal grammar;

* integration of stochastic and symbolic models of grammar;

* foundational, methodological and architectural issues in grammar and linguistics;

* mathematical foundations of statistical approaches to linguistic analysis.
Previous conferences in this series have welcomed papers from a wide variety of frameworks.

Submission Details

We invite **electronic** submissions of original, 16-page papers (including references and possible technical appendices).  Authors are encouraged to use the Springer-Verlag LNCS style.

The submission deadline is **February 23 2014**. Papers must be submitted electronically at https://www.easychair.org/conferences/?conf=fg2014

Papers should report original work which was not presented in other conferences. However, simultaneous submission is allowed, provided that the authors indicate other conferences to which the work was submitted in a footnote. Note that accepted papers can only be presented in one of the venues.

Submissions will be reviewed anonymously by at least three reviewers. Accepted papers will be published as a volume in the Springer LNCS series, under the FoLLI subline, either separately or jointly with the papers from FG-2015, depending on the number of accepted papers.

Important Dates

* February 23, 2014: Deadline for paper submission

* April 20, 2014: Notification of acceptance

* June 1, 2014: Camera ready copies due

* August 16-17, 2014: Conference dates

Call for papers: CCLCC 2014

Computational, Cognitive, and Linguistic Approaches to the Analysis of Complex Words and Collocations (CCLCC 2014)

http://www.sfs.uni-tuebingen.de/~vhenrich/cclcc_2014/

Workshop organized as part of the European Summer School on Logic, Language and Information ESSLLI 2014 (http://www.esslli2014.info/), August 11-15 2014 (ESSLLI first week), Tübingen, Germany


Workshop Organizers:
--------------------

Erhard Hinrichs and Verena Henrich
esslli-workshop@sfs.uni-tuebingen.de


Workshop Purpose:
-----------------

The analysis of complex words, compounds, and collocations has received considerable attention in linguistics, cognitive science and computational linguistics. Research on these phenomena concerns theoretical, experimental, and applied aspects relevant to all three disciplines. This diverse and interdisciplinary perspective lends itself particularly well to an ESSLLI workshop on this topic.

The aim of the workshop is to stimulate a cross-disciplinary discussion that will be of mutual benefit to the three fields sketched above and that will provide a forum for junior and senior researchers alike.

Relevant topics for the workshop include, but are not limited to, the following issues:

- Cognitive and computational models for the interpretation of compounds and collocations
- Experimental and computational methods for the analysis of compounds and collocations
- Statistical models for the identification and classification of collocations
- Word formation, lexical semantics, and semantic compositionality
- Word formation at the interface of morphology and syntax


Submission Details:
-------------------

Authors are invited to submit an EXTENDED ABSTRACT for a 30-minute presentation (including discussion). Submissions should not exceed 3 pages, including figures, data, and references. The submissions will be reviewed anonymously by the workshop's program committee. The abstracts accepted for presentation will appear in the workshop web site and be published as part of the ESSLLI 2014 proceedings. In addition, we are considering the possibility of compiling a journal special issue from selected papers presented at the workshop.

Please upload your submission at EasyChair:
https://www.easychair.org/conferences/?conf=cclcc2014


Workshop Format:
----------------

The workshop is part of ESSLLI and is open to all ESSLLI participants. It will consist of five 90-minute sessions held over five consecutive days in the first week of ESSLLI. There will be 2-3 slots for paper presentation and discussion per session. On the first day the workshop organizer will give an introduction to the topic.

All workshop participants, including the authors, will be required to register for ESSLLI.


Important Dates:
----------------

- December 23: First Call for Papers
- January 11: Second Call for Papers
- February 11: Last Call for Papers
- March 8: *Deadline for Submissions*
- April 21: Notification of Acceptance
- August 11-15: Workshop

Call for Abstracts: TBilisi Symposium on Language, Logic and Computation

***********************************************************************
3rd Call for Papers

THE TENTH INTERNATIONAL TBILISI SYMPOSIUM ON
LANGUAGE, LOGIC AND COMPUTATION

23-27 September 2013
Gudauri, Georgia

http://www.illc.uva.nl/Tbilisi/Tbilisi2013
***********************************************************************
CALL FOR PAPERS

The Tenth International Tbilisi Symposium on Language, Logic and Computation will be held on 23-27 September 2013 in Gudauri, Georgia. The Programme Committee invites submissions for contributions on all aspects of language, logic and computation. Work of an interdisciplinary nature is particularly welcome. Areas of interest
include, but are not limited to:

 * Algorithmic game theory
 * Computational social choice
 * Constructive, modal and algebraic logic
 * Formal models of multiagent systems
 * Historical linguistics, history of logic
 * Information retrieval, query answer systems
 * Language evolution and learnability
 * Linguistic typology and semantic universals
 * Logic, games, and formal pragmatics
 * Logics for artificial intelligence
 * Natural language syntax, semantics, and pragmatics

Authors can submit an abstract of four pages (including references) at
the EasyChair conference system here:

http://www.easychair.org/conferences/?conf=tbillc2013

PROGRAMME

The programme will include the following invited lectures, tutorials
and workshops.

Tutorials:

Logic: Rosalie Iemhoff (Utrecht University)
Language: Daniel Altshuler (University of Duesseldorf)
Computation: Samson Abramsky (University of Oxford)

Invited Lectures:

Balder ten Cate (University of California at Santa Cruz)
Agata Ciabattoni (Vienna University of Technology)
Thomas Colcombet (University of Paris)
Galit Sassoon (University of Jerusalem)
Alexandra Silva (Radboud University Nijmegen)
Sergei Tatevosov (Moscow State University )

Workshops:

Workshop on Algebraic Proof Theory

Invited speakers:
Matthias Baaz (Vienna University of Technology)
Alessio Guglielmi (University of Bath)
Kazushige Terui (University of Kyoto)

Organisers:
Agata Ciabattoni and Rosalie Iemhoff

More information:
http://www.illc.uva.nl/Tbilisi/Tbilisi2013/Workshops/

Workshop on Aspect

Invited speakers:
Roumyana Pancheva (University of Southern California)
Hans Kamp (Stuttgart University/University of Texas at Austin)
Corien Bary (Radboud University Nijmegen)

Organisers:
Daniel Altshuler, Daniel Hole and Sergei Tatevosov

More information:
http://www.illc.uva.nl/Tbilisi/Tbilisi2013/Workshops/

Programme Committee:

Lev Beklemishev (Steklov Mathematical Institute, Moscow)
Nick Bezhanishvili (Chair, Utrecht University)
Dion Coumans (Radboud University Nijmegen)
Frans Groen (University of Amsterdam)
Wiebe van der Hoek (University of Liverpool)
Daniel Hole (University of Stuttgart)
Clemens Kupke (University of Oxford)
Sebastian Loebner (University of Duesseldorff)
Larry Moss (Indiana University)
Prakash Panangaden (McGill University)
Floris Roelofsen (University of Amsterdam)
Susan Rothstein (Bar-Ilan University)
Luca Spada (University of Salerno)

PUBLICATION INFORMATION

Post-proceedings of the symposium will be published in the LNCS series of Springer.

IMPORTANT DATES

Submission deadline: May 1, 2013
Notification: July 1, 2013
Final abstracts due: August 1, 2013
Registration deadline: September 1, 2013
Symposium: September 23-27, 2013

Programme and submission details can be found at:

http://www.illc.uva.nl/Tbilisi/Tbilisi2013/

Call for Papers: Tbilisi Symposium on Language, Logic and Compuatation

THE TENTH INTERNATIONAL TBILISI SYMPOSIUM ON
LANGUAGE, LOGIC AND COMPUTATION

23-27 September 2013
Gudauri, Georgia

http://www.illc.uva.nl/Tbilisi/Tbilisi2013
***********************************************************************
CALL FOR PAPERS

The Tenth International Tbilisi Symposium on Language, Logic and
Computation will be held on 23-27 September 2013 in Gudauri,
Georgia. The Programme Committee invites submissions for contributions
on all aspects of language, logic and computation. Work of an
interdisciplinary nature is particularly welcome. Areas of interest
include, but are not limited to:

* Algorithmic game theory
* Computational social choice
* Constructive, modal and algebraic logic
* Formal models of multiagent systems
* Historical linguistics, history of logic
* Information retrieval, query answer systems
* Language evolution and learnability
* Linguistic typology and semantic universals
* Logic, games, and formal pragmatics
* Logics for artificial intelligence
* Natural language syntax, semantics, and pragmatics

Authors can submit an abstract of four pages (including references) at
the EasyChair conference system here:

http://www.easychair.org/conferences/?conf=tbillc2013

PROGRAMME

The programme will include the following invited lectures and
tutorials.

Tutorials:

Logic: Rosalie Iemhoff (Utrecht)
Language: Daniel Altshuler (Duesseldorf)
Computation: Samson Abramsky (Oxford)

Invited Lectures:

Balder ten Cate (Santa Cruz)
Agata Ciabattoni (Vienna)
Thomas Colcombet (Paris)
Galit Sassoon (Jerusalem)
Alexandra Silva (Nijmegen)
Sergei Tatevosov (Moscow)

WORKSHOPS

There will also be a workshop on Algebraic Proof Theory organized by
A. Ciabattoni and R. Iemhoff and a workshop on Aspect organized by
D. Altshuler, D. Hole and S. Tatevosov. More information can be found
on the TbiLLC website.

PUBLICATION INFORMATION

Post-proceedings of the symposium will be published in the LNCS series
of Springer.

IMPORTANT DATES

Submission deadline: May 1, 2013
Notification: July 1, 2013
Final abstracts due: August 1, 2013
Registration deadline: September 1, 2013
Symposium: September 23-27, 2013

Programme and submission details can be found at:

http://www.illc.uva.nl/Tbilisi/Tbilisi2013/

John Hale gives department colloq on Friday

John Hale, Associate Professor of Linguistics at Cornell, will present the department colloquium on Friday, October 26, at 3:30PM in Machmer E-37. The title and abstract of his talk follow.

Experience as a Control Strategy for Incremental Parsing

This talk presents a family of computer models that are intended to capture syntactic aspects of incremental human sentence comprehension. These algorithmic models build on the Garden Path Theory of Frazier, Clifton & colleagues as well as the seminal work of Rick Lewis. However, they go beyond these classic approaches by introducing experience as a key element of the control strategy. The talk assesses the implications of this move by looking at attachment ambiguities and garden path sentences. Along the way, we re-encounter some classic questions about the inter-relationships between learning, grammar and cognitive architecture that can be answered, at least provisionally, by expressing the model within Soar 9, where reinforcement learning is applicable at all levels of abstraction. This demonstration illustrates one way that grammar might fit into the rest of cognition.

Miriam Butt presented "Deep Natural Language Processing"

Miriam Butt gave a talk last Friday, October 5. Miriam is
Professor of Linguistics and Chair of the Department of Linguistics
(Fachbereich Sprachwissenschaft) at the University of Konstanz. She is best known for her theoretical linguistic work on complex predicates and on grammatical case, and for her computational linguistic work in large-scale grammar development within the Pargram project. Recently she has been also involved in interdisciplinary work on the visualization of linguistic data.

The title, and abstract, of her talk was:

Deep Natural Language Processing

This talks provides a look at the capabilities of deep linguistic
natural lanuguage processing in the context of ParGram, an
international effort at building grammars for languages as diverse as
English, French, Welsh, Wolof, Georgian, Urdu, Japanese, Indonesian,Arabic and Murrinh-Patha.  The aim is to build these grammars using not only the same theoretical assumptions (based on Lexical-Functional Grammar) but also the same computational methods.  While the process of computational grammar devleopment is, of course, inherently interesting, we will also discuss the broader uses of computational grammars, for example in systems like IBM's Watson and Powerset's information retrieval system.

The underlying development platform XLE/XFR (used by Powerset and in the ParGram effort) was developed at PARC over a number of years.  The talk will focus on the various powerful analysis possibilities this system allows for, including a version of Optimality Theory and an integration of statistical information to constrain and inform both parsing and generation. The talk will also touch on the assumptions about the underlying grammar architecture that are made, e.g., with respect to the morphology-syntax and the syntax-semantics interface.