Conference on Logic and Machine Learning in Natural Language (LaML)

 

1561330_clasp_webbanner

We are very happy to announce the first official conference organized by CLASP, Logic and Machine Learning in Natural Language (LaML). The conference hopes to bring together researchers working in both logical  and machine learning/deep learning approaches to Computational Linguistics. We aim  at initiating a discussion with the two sets  of researchers that have been largely unconnected up to now. The conference will take place on June the 12th-14th 2016 in Gothenburg, Sweden. For more information please take a look at the call for papers.

Call for Papers

Conference on Logic and Machine Learning in Natural Language (LaML)

Conference dates: June 12-14, 2017

Venue: Wallenburg Conference Centre, University of Gothenburg

Organised by CLASP, University of Gothenburg

The past two decades have seen impressive progress in a variety of areas of AI, particularly NLP, through the application of machine learning methods to a wide range of tasks. With the intensive use of deep learning methods in recent years this work has produced significant improvements in the coverage and accuracy of NLP systems in such domains as speech recognition, topic identification, semantic interpretation, and image description generation. 

While deep learning is opening up exciting new approaches to longstanding, difficult problems in computational linguistics, it also raises important foundational questions. Specifically, we do not have a clear formal understanding of why multi-level recursive deep neural networks achieve the success in learning and classification that they are delivering. It is also not obvious whether they should displace more traditional, logically driven methods, or be combined with them. Finally, we need to explore the extent, if any, to which both logical models and machine learning methods offer insights into the cognitive foundations of natural language. 

The Conference on Logic and Machine Learning in Natural Language will address these questions and related issues. It will feature invited talks by leading researchers in both fields, and high level contributed papers selected through open competition and rigorous review. Our aim is to initiated a genuine dialogue between these two approaches, where there they have traditionally remained separate and in competition.

The conference proceedings will be published online, with an ISSN, on the CLASP website. Authors will retain the copyright of their papers and be free to publish them elsewhere, with acknowledgement.

Registration is free and participation is open. We warmly invite everyone to attend. 

Invited Speakers:

Marco Baroni, Trento
Alexander Clark, King’s College London
Devdatt Dubhashi, Chalmers
Katrin Erk, University of Texas, Austin
Joakim Nivre, Uppsala
Aarne Ranta, Gothenburg
Mehrnoosh Sadrzadeh, Queen Mary University of London

Submission of papers:

EasyChair submission address: https://easychair.org/conferences/?conf=laml2017

Contact email address: laml2017@easychair.org

LaTeX style files: http://eacl2017.org/images/site/eacl-2017-template.zip

Camera copies will be up to 5 pages + references.

We anticipate accepting 17 papers for oral presentation, and up to 20 papers for poster presentation.

Important dates: 

1. Deadline for submission: April 4, 2017

2. Notification of authors: April 27, 2017

3. Camera ready papers due: May 9, 2017

The LaML Programme Committee:

Marco Baroni, Trento
Islam Beltagy, University of Texas, Austin
Jean-Philippe Bernardy, Gothenburg
Gemma Boleda, Universitat Pompeu Fabra 
Stergios Chatzikyriakidis, Gothenburg
Alexander Clark, King’s College London
Robin Cooper, Gothenburg
Simon Dobnik, Gothenburg
Devdatt Dubhashi, Chalmers
Katrin Erk, University of Texas, Austin
Julian Hough, Bielefeld
Christine Howes, Gothenburg
John D. Kelleher, Dublin Institute of Technology
Shalom Lappin, Gothenburg
Staffan Larsson, Gothenburg
Julian Michael, Washington
Joakim Nivre, Uppsala
Barbara Plank, Groningen
Matthew Purver, Queen Mary University of London
Aarne Ranta, Gothenburg
Mehrnoosh Sadrzadeh, Queen Mary University of London
Anders Søgaard, Copenhagen
Charalambos Themistocleous, Gothenburg

 

Talk @Chalmers: Eliciting information from speech signals: Linguistics, signal processing, and machine learning.

Abstract: Listeners during an interaction can elicit multimodal linguistic and social information about a typical or atypical speaker, such as information about her/his gender, age, place of origin, social class, emotional state, and behaviour. The process of information extraction and classification takes place on the spot and involves the decomposition and classification of complex, heterogeneous, and multimodal parameters from the audiovisual channel.

In this talk, I will be discussing approaches from linguistics, signal processing, and machine learning for the quantification and understanding of linguistic and social information encoded in the speech signals of utterances produced in controlled experimental settings. The potential applications of this research in personalised discourse systems, in speaker identification, and as diagnostic tools of atypical speech productions, such as aphasic speech, are also discussed.

Randomizing Experimental Material

Sometimes we need to randomize our experimental material for every speaker and repetition. To this purpose, I wrote a quick script in R,  which you can modify to fit your own purposes. The script saves the experimental material of each speaker in a different file in the path you should define at the top of the code.

setwd("/Users/")   # Export Path
a <- c(
"distractor1",
"distractor2",
"distractor3",
"distractor4",
"distractor5",
"distractor6",
"Phrase1",
"Phrase2",
"Phrase3",
"Phrase4",
"Phrase5",
"Phrase6",
"Phrase7",
"Phrase8",
"Phrase9",
"Phrase10",
)
for (num in 1:10) # Number of Speakers.
{
  sink(paste("SPEAKER ",num,".txt", sep=""))
  cat(paste("SPEAKER", num),sep="\n")
for(i in 1:6) # Number of Repetitions per Speaker. 
{
  b <- sample(a)
  cat(paste("Repedition ",i),sep="\n")
  cat(b,sep="\n")
}
sink()
}

 

Greek vowels and the nature of acoustic gradience

dynamicsIn this talk, I will discuss vowel variation in Standard Modern Greek (SMG) and Cypriot Greek (CG). These two varieties differ most significantly in their consonants, but it is very difficult to perceive differences in the vowels. Young—between 18–23—Athenian and Nicosian speakers participated in the study. The results demonstrate gradi- ent effects of vowel variation and change on vowel acoustic properties: vowel duration, articulatory space, and formant dynamics. I argue that their variation is best understood with respect to Greek vowel raising and vowel reduction, and suggest that vowel variation in Greek applies on specific subsystems (unstressed vowels vary significantly between the two varieties whereas stressed vowels display only minor differences). The findings appear to be the first to gather and compare acoustic material from urban Standard Modern Greek (SMG) and Cypriot Greek (CG) vowels in a unified manner. The implications of these findings for studies on language variation, language change, and language pathology are also discussed.

Date: 2/18/2016

Time: 1:15 PM – 3:00 PM

Categories: Humanities

Organizer: Centre for Language Technology, Gothenburg

Location: Seminar Room, Dicksongatan 4

Talk@University of Gothenburg: Melodic patterns in speech communication: Interactions and heterogeneity

The melodic patterns of a Stockholm Swedish speaker contribute to the successful or unsuccessful interaction of the speaker with other speakers. They also function as a sociophonetic fingertip, which shows how the speaker relates with the specific speech community. In this talk, I will present two experimental studies that shed light on the interaction between prosodic and segmental structure. The first study provides empirical evidence that attempt to clarify the domain of final lengthening and the second study, by exploring the timing of the L*+H prenuclear pitch accent, examines the predictions of three hypotheses about tonal alignment: the invariance hypothesis, the segmental anchoring hypothesis, and the segmental anchorage hypothesis. The findings of the study cast doubt on the invariance hypothesis and the segmental anchoring hypothesis, as well as indicating the need for a modified version of the segmental anchorage hypothesis. The implications of these findings for a computational model of prosody are also discussed.

When: 1 March, 2016, 13.15-14.15

Research seminars

This is a venue for linguists in Gothenburg to share their research and methodological expertise with each other. For example, this would be an excellent place to do a practice talk for a conference, a technical presentation (e.g. on how to use some piece of software), or a presentation on work in progress.

To schedule a slot or for other information please contact Liz Coppock.

 

Upcoming research seminars

The research seminars alternate fortnightly with seminars. Research meetings are on odd weeks (and seminars on even weeks). Both events take place in the seminar room on the ground floor of Dicksonsgatan 4.

Talk@University of Gothenburg: Doing Type Theory in R

RlogoIn this talk, I will present R language (or simply R ), a dynamic, lazy, functional, programming language that was designed in 1993 by Ross Ihaka and Robert Gentleman. R adopts the underlying evaluation model of Scheme with the syntax of S, (a programming language, which was developed by John Chambers at Bell Laboratories).

R is an open-source programming language and the flexible statistical analysis toolkit implemented in R , made it the lingua franca for doing statistics. The R package repository (CRAN) features 7861 available packages, which extent the language. Also, there are guides on CRAN that group sets of R packages and functions by type of analysis, fields, or methodologies (e.g. Bayesian Inference, Probability Distributions, Machine Learning, Natural Language Processing).

The statistical capabilities of R along with its functional capabilities can transform R into a rich environment for doing Type Theory. Thus, I will conclude this talk by discussing possible extensions of R for A Probabilistic Rich Type Theory for Semantic Interpretation (Cooper, Dobnik, Lappin, and Larsson, 2015).

Date: 2/18/2016

Time: 3:00 PM – 5:00 PM

Categories: Humanities

Organizer: Centre for Language Technology, Gothenburg

Location: Seminar Room, Dicksongatan 4

Contact person: Dana Dannélls

Talk@ Princeton University: Language Variation and Change: Evidence from Modern Greek Vowels

If you recognize a New York accent, you can probably name its distinctive characteristics; yet, most of its melodic properties are subtle and lie under the level of consciousness of most speakers. The same is true for all accents. In this talk, I discuss how accentual variation is conveyed in speech productions and present experimental evidence from studies, which were conducted in Athens (Greece) and Nicosia (Cyprus). In fact, these studies appear to be the first to gather and compare acoustic material from urban Standard Modern Greek (SMG) and Cypriot Greek (CG) vowels in a unified manner. These two varieties differ most significantly in their consonants, but it is very difficult to perceive differences in their vowels. The subjects are young—between 18-23—Athenian and Nicosian students. In the talk, I provide evidence about the gradient effects of vowel variation and change in Greek and demonstrate that the vowels of these two varieties differ in their duration, articulatory space, and dynamic properties of their frequencies. I also argue that their variation is best understood with respect to Greek “vowel raising” and “vowel reduction” and show that phonetic variation follows systematic patterns; for instance, it applies on specific subsystems (unstressed vowels vary significantly between the two varieties whereas stressed vowels display only minor differences). The implications of these findings for studies on language variation, language change, and language pathology are also discussed.

Location: Scheide Caldwell House, Room 103

Date/Time: 11/20/15 at 1:30 pm – 11/20/15 at 3:00 pm

Department: Seeger Center for Hellenic Studies