February 2023 - Corpora

[PhD-Position] Research Position in Speech- and Text Anonymization in Medical Domain in Berlin
by Polzehl, Tim, PD Dr.-Ing. 16 Feb '23

16 Feb '23

Dear colleagues, we have a new PhD vacancy in the field of speech- and text anonymization in the medical domain in Berlin, Germany. The position is in the “Medinym” project of the department of Quality and Usability Labs of Berlin Institute of Technology. We’re looking for a Researcher or Junior Researcher level, offer a 2 years contract with optional prolongation and PhD perspective. Application deadline: Feb 28 More details and how to apply: TU Berlin: https://www.jobs.tu-berlin.de/en/job-postings/161912 Please circulate upon potentially interested. Many thanks! In case of questions pls contact me, I'm happy to help. Best regards from Berlin, Tim -- Dr.-Ing. Tim Polzehl Associate Senior Researcher Technische Universität Berlin Quality and Usability Lab Ernst-Reuter-Platz 7 D-10587 Berlin, Germany Email: tim.polzehl(a)qu.tu-berlin.de Web: www.qu.tu-berlin.de

1 0

The 28th International Conference on Natural Language & Information Systems - Second call for papers
by Farid Meziane 16 Feb '23

16 Feb '23

NLDB 2023 The 28th International Conference on Natural Language & Information Systems 21-23 June 2023, University of Derby, United Kingdom. https://www.derby.ac.uk/events/latest-events/nldb-2023/ About NLDB The 28th International Conference on Natural Language & Information Systems will be held at the University of Derby, United Kingdom and will be a face to face event. Since 1995, the NLDB conference brings together researchers, industry practitioners, and potential users interested in various application of Natural Language in the Database and Information Systems field. The term "Information Systems" has to be considered in the broader sense of Information and Communication Systems, including Big Data, Linked Data and Social Networks. The field of Natural Language Processing (NLP) has itself recently experienced several exciting developments. In research, these developments have been reflected in the emergence of neural language models (Deep Learning, Word Embeddings, Transformers) and the importance of aspects such as transparency, bias and fairness, a (renewed) interest in various linguistic phenomena, such as in discourse and argumentation mining, and in new problems such as the detection of disinformation and hate speech in social media, as well of mental health disorders that increased during the recent pandemic. Regarding applications, NLP systems have evolved to the point that they now offer real-life, tangible benefits to enterprises. Many of these NLP systems are now considered a de-facto offering in business intelligence suites, such as algorithms for recommender systems and opinion mining/sentiment analysis. It is against this backdrop of recent innovations in NLP and its applications in information systems that the 28th edition of the NLDB conference takes place. We welcome research and industrial contributions, describing novel, previously unpublished works on NLP and its applications across a plethora of topics as described in the Call for Papers. Call for Papers NLDB 2023 invites authors to submit papers for oral or poster presentations on unpublished research that addresses theoretical aspects, algorithms, applications, architectures for applied and integrated NLP, resources for applied NLP, and other aspects of NLP, as well as survey and discussion papers. This year's edition of NLDB also introduces an Industry Track, to foster fruitful interaction between the industry and the research community. Topics of interest include but are not limited to: * Social Media and Web Analytics: Opinion mining/sentiment analysis, irony/sarcasm detection; detection of fake reviews and deceptive language; detection of harmful information: fake news and hate speech; sexism and misogyny; detection of mental health disorders; identification of stereotypes and social biases; robust NLP methods for sparse, ill-formed texts; recommendation systems. * Deep Learning and eXplainable Artificial Intelligence (XAI): Deep learning architectures, word embeddings, transparency, interpretability, fairness, debiasing, ethics. * Argumentation Mining and Applications: Automatic detection of argumentation components and relationships; creation of resource (e.g. annotated corpora, treebanks and parsers); Integration of NLP techniques with formal, abstract argumentation structures; Argumentation Mining from legal texts and scientific articles. * Question Answering (QA): Natural language interfaces to databases, QA using web data, multi-lingual QA, non-factoid QA(how/why/opinion questions, lists), geographical QA, QA corpora and training sets, QA over linked data (QALD). * Corpus Analysis: multi-lingual, multi-cultural and multi-modal corpora; machine translation, text analysis, text classification and clustering; language identification; plagiarism detection; information extraction: named entity, extraction of events, terms and semantic relationships. * Semantic Web, Open Linked Data, and Ontologies: Ontology learning and alignment, ontology population, ontology evaluation, querying ontologies and linked data, semantic tagging and classification, ontology-driven NLP, ontology-driven systems integration. * Natural Language in Conceptual Modelling: Analysis of natural language descriptions, NLP in requirement engineering, terminological ontologies, consistency checking, metadata creation and harvesting. * Natural Language and Ubiquitous Computing: Pervasive computing, embedded, robotic and mobile applications; conversational agents; NLP techniques for Internet of Things (IoT); NLP techniques for ambient intelligence * Big Data and Business Intelligence: Identity detection, semantic data cleaning, summarisation, reporting, and data to text. Important Dates: Full paper submission: 14 March, 2023 Paper notification: 10 April, 2023 Camera-ready deadline: 24 April, 2023 Conference: 21-23 June 2023 Submission Guidelines Authors should follow the LNCS format (https://www.springer.com/gp/computer-science/lncs/conference-proceedings-gu… ) and submit their manuscripts in pdf via Easychair (https://easychair.org/conferences/?conf=nldb2023 ) Submissions can be full papers (12 pages maximum including references), short papers (8 pages including references) or papers for a poster presentation or system demonstration (6 pages including references). The programme committee may decide to accept some full papers as short papers or poster papers. The reviewing process of NLDB 2023 is double-blind, i.e., submissions must not contain author names or other identifying information, such as funding sources, acknowledgments and must use the third person to refer to work the authors have previously undertaken. System demonstration papers may not be anonymous. [University of Derby] Professor Farid Meziane, PhD, FHEA Professor of Data Science Data Science Research theme lead Head Data Science Research Centre<https://www.derby.ac.uk/research/about-our-research/centres-groups/data-sci…> Chair, College Research Committee College of Science and Engineering University of Derby, Markeaton Street, Derby DE22 3AW NLDB2023 Conference at Derby <https://www.derby.ac.uk/events/latest-events/ndlb-2023/> Te: 01332 594031 f.meziane(a)derby.ac.uk<mailto:f.meziane@derby.ac.uk> The University of Derby has a published policy regarding email and reserves the right to monitor email traffic. If you believe this was sent to you in error, please reply to the sender and let them know. Key University contacts: http://www.derby.ac.uk/its/contacts/

1 0

[Job opening] Computational linguist - European Parliament
by Elisa Di Nuovo 16 Feb '23

16 Feb '23

Dear list members, I would like to draw your attention on the following job opening at the European Parliament, Directorate-General for Translation, Directorate for Citizens' language, Speech-to-Text Unit. Currently the team is managing a project aimed at acquiring an Automatic Speech Recognition (ASR) and Machine Translation (MT) cascade tool covering all the 24 official languages of the European Union. The main goal of this tool is to make plenary debates accessible to D/deaf and hard of hearing persons. The successful candidate will be responsible for the evaluation of the tool, evaluated using both automatic metrics and human assessment. A job description is attached to this email. The main responsibilities are: - linguistic evaluation of the performance of the tool (automatic metrics + human evaluation); - monitoring the performance of the tool in real events; - writing reports; - data collection, curation and annotation for the purposes of the project; - terminology coordination; - project-related research and technology watch; - other project related activities. The contract length is 1 year (renewable maximum for 6 years) and is identified as Contractual Agent FG IV. The job will be performed in Luxembourg. Possible *commencement of duties* on the 1st of September 2023. If interested, please send your CV to Dr Marcin Feder ( marcin.feder(a)europarl.europa.eu) *by the 16th March 2023*. Best regards, Elisa Di Nuovo

1 1

CfP: Data for the Wellbeing of Most Vulnerable Workshop @ICWSM'23
by yelena.mejova＠gmail.com 16 Feb '23

16 Feb '23

Call for Papers & Participation Fourth Workshop on DATA FOR THE WELLBEING OF MOST VULNERABLE https://sites.google.com/view/dataforvulnerable23 June 5, 2023 at Limassol, Cyprus * Submissions due: March 27, 2022 * At the International AAAI Conference on Web and Social Media (ICWSM) https://www.icwsm.org/2023/index.html The scale, reach, and real-time nature of the Internet is opening new frontiers for understanding the vulnerabilities in our societies, including inequalities and fragility in the face of a changing world. From tracking seasonal illnesses like the flu across countries and populations, to understanding the context of mental conditions such as anorexia and bulimia, web data has the potential to capture the struggles and wellbeing of diverse groups of people. Vulnerable populations including children, elderly, racial or ethnic minorities, socioeconomically disadvantaged, underinsured or those with certain medical conditions, are often absent in commonly used data sources. The recent developments around COVID-19 epidemic makes these issues even more urgent, with an unequal share of both disease and economic burden among various populations. Thus, the aim of this workshop is to encourage the community to use new sources of data as well as methodologies to study the wellbeing of vulnerable populations. The selection of appropriate data sources, identification of vulnerable groups, and ethical considerations in the subsequent analysis are of great importance in the extension of the benefits of big data revolution to these populations. As such, the topic is highly multidisciplinary, bringing together researchers and practitioners in computer science, epidemiology, demography, linguistics, and many others. - We anticipate topics such as the below will be relevant: - Establishing cohorts, data de-biasing - Validation via individual-level or aggregate-level data - Linking data to disease and other well-being - Population data sources for validation - Correlation analysis and other statistical methods - Longitudinal analysis on social media - Spatial, linguistic, and temporal analyses - Privacy, ethics, and informed consent - Data quality issues Important Dates Papers Submissions: March 27, 2023 Paper Acceptance Notification: April 10, 2023 Final Camera-Ready Paper Due: May 6, 2023 ICWSM-2023 Workshops Day: June 5, 2023 Submission Instructions We welcome both 2-page abstracts, as well as Long (8 pages) and Short (4 pages) papers - excluding references (11 pages max with references and ethics statement). The Long and Short papers will be published in ICWSM Workshop proceedings (http://workshop-proceedings.icwsm.org/). The papers have to follow the AAAI format, as outlined here: https://www.overleaf.com/latex/templates/aaai-2023-author-kit/wxnmhzcrjbpc Submit here: https://easychair.org/conferences/?conf=dwmv23 Contact Feel free to contact organizers with any questions: Yelena Mejova <yelena.mejova(a)isi.it> ISI Foundation, Italy https://yelenamejova.com Kyriaki Kalimeri <kalimeri(a)ieee.org> ISI Foundation, Italy https://www.linkedin.com/in/kalimeri/ Daniela Paolotti <daniela.paolotti(a)isi.it> ISI Foundation, Italy https://www.isi.it/en/people/daniela-paolotti

1 0

Call for Papers: MATCHING WORKSHOP @ ACL 2023
by Estevam Rafael Hruschka Junior 16 Feb '23

16 Feb '23

MATCHING WORKSHOP @ ACL 2023 https://megagon.ai/matching-2023/ *I**mportant Dates* direct submission via OpenReview: April 24, 2023 submission via ARR - commitment deadline: May 8th, 2023 Notification of Acceptance: May 22, 2023 Camera-ready papers due: June 6, 2023 Pre-recorded video due: June 12, 2023 MATCHING Workshop @ACL2023 (Toronto and remote): July 13 Matching Entities from structured and unstructured sources is an important task in many domains and applications such as HR and E-commerce. For example, in HR platforms/services, it is important to match resumes to job descriptions and job seekers to companies. Similarly in web platforms/services, it is important to match customers to businesses such as hotels and restaurant, among others. In such domains, it is also relevant to match “textual customer reviews” to customers queries, and sentences (or phrases) as answers to customer questions. Recent advances in Natural Language Processing, Natural Language Understanding, Conversational AI, Language Generation, Machine Learning, Deep Learning, Data Management, Information Extraction, Knowledge Bases/Graphs, (MultiSingle Hop/Commonsense) Inference/Reasoning, Recommendation Systems, and others, have demonstrated promising results in different Matching tasks related (but not limited) to the previously mentioned domains. We believe that there is tremendous opportunity to further exploit and explore the use of advanced NLP (and language related) techniques applied to Matching tasks. Therefore, the goal of this workshop is to bring together the research communities (from academia and industry) of these related areas, that are interested in the development and the application of novel natural-language-based approaches/models/systems to address challenges around different Matching tasks. *Confirmed Invited Speakers* Alan Ritter - Georgia Tech Ndapa Nakashole - UC San Diego William Cohen - Google We invite submissions of long and short papers on original and unpublished work that address challenges around the specific task of matching information from heterogeneous sources spanning structured (e.g., databases, resume) and unstructured (e.g., online reviews, job advertisements, social media posts) data. The topics include but are not limited to: * Entity Alignment; * Entity Matching; * Entity Linking; * Language-Model-based Matching; * Data-driven Matching; * Knowledge-graph-based matching; * Rule-based Matching; * Temporal Information Matching; * Human-Centric AI for Matching; * Trust, Explainability, and Fairness in Matching; * Information extraction from user-generated text; * Robustness to noise * Matching applications: recommendations, question answering, fact checking, news and social media content identification/suggestion, etc. All machine learning, text mining, and natural language processing techniques are welcome. All regular papers and short papers should follow the ACL 2023 style guidelines: “Both long and short papers must follow the ACL 2023 two-column format, using the supplied official style files. The templates can be downloaded in Style Files and Formatting. Please do not modify these style files, nor should you use templates designed for other conferences. Submissions that do not conform to the required styles, including paper size, margin width, and font size restrictions, will be rejected without review.” The maximum length of a regular paper is 8 pages plus an unlimited number of pages for references. The maximum length of a short paper is 4 pages plus an unlimited number of pages for references. At least one author of every accepted paper is expected to attend the workshop. MATCHING is using a hybrid submission process, Authors can submit their papers using the OpenReview platform ( https://openreview.net/group?id=aclweb.org/ACL/2023/Workshop/MATCHING&refer…). Alternatively, authors can also commit papers and reviews from ARR. We allow parallel commitment to the EACL 2023 conference and our workshop, with the requirement that if the paper is accepted at EACL 2023, it will be withdrawn from archival publication at the workshop. We ask the authors to notify MATCHING if they have also committed to EACL or other workshops. All accepted short and long papers must be presented as a talk/poster/demo at the workshop, depending on the workshop schedule. At least one author of each accepted paper must register for ACL 2023 and attend the workshop. Mentorship Program We will be hosting a mentorship program to facilitate exchange between authors and experts working in areas relevant to the workshop. The goal of this program is to foster collaborations and increase the quality and impact of the submissions. Mentors are expected to guide mentees during the mentorship period as they prepare submissions for this workshop. Their interactions may include an in-depth discussion of related work to ensure submissions are well contextualized, discussions on writing and presentation of the submission, and discussions about possible extensions of the submissions. Mentees are expected to prepare the submissions by February 28, 2023 and contact their assigned mentor. Mentors and mentees are encouraged to dedicate at least 4hrs over the course of the program to maximize the benefits of the program. They can meet virtually within the first week after the mentor-mentee matching is made and set expectations for subsequent meetings. Their efforts should culminate in a final version of the paper that should be submitted by the deadline. Applications to the mentorship program are due February 28, 2023 Application to be a mentor/mentee: Apply Here: *Organizers:* Estevam Hruschka - Megagon Labs Tom Mitchell - Carnegie Mellon University Dunia Mladenić - Jozef Stefan Institute (JSI) Marko Grobelnik - Jozef Stefan Institute (JSI) Sajjadur Rahman - Megagon Labs *Contact* If you have any questions or inquiries regarding the workshop or need further information, please do not hesitate to send an email to matching-workshop(a)megagon.ai -- Estevam Hruschka Lab Director and Staff Research Scientist Megagon Labs - www.megagon.ai -- Sent from my iPhone

1 0

First CFP: AmericasNLP 2023 @ ACL
by Katharina Kann 15 Feb '23

15 Feb '23

The Third Workshop on NLP for Indigenous Languages of the Americas (AmericasNLP 2023) First Call for Papers The Third Workshop on NLP for Indigenous Languages of the Americas (AmericasNLP) will be co-located with the 61st Annual Meeting of the Association for Computational Linguistics (ACL 2023 <https://2023.aclweb.org/>), which is scheduled to be held in Toronto, Canada, between July 9-14, 2023. The goal of the workshop is to encourage and increase the visibility of work on the Indigenous languages of the Americas. It aims to encourage research on NLP, computational linguistics, corpus linguistics and speech for Indigenous languages, to connect researchers and professionals from underrepresented communities and native speakers of endangered languages with the ACL community, and, more generally, to promote machine learning approaches suitable for low-resource languages. We invite the submission of - Long papers (8 pages) and short papers (4 pages) on substantial, original, and unpublished research - Non-archival extended abstracts (2 pages), technical reports (8 pages), and work which has been presented at other venues (in the format of the original publication) Submissions do not need to describe work on native languages directly, as long as it is clear why those can benefit from the described approaches. Areas of interest include but are not limited to: - Creation of datasets for NLP applications - Incorporation of external knowledge into neural systems - Linguistic typology and the use of typological features for NLP - Transfer learning, meta-learning, and active learning - Weakly supervised, semi-supervised, and unsupervised learning - Machine translation of low-resource languages - Morphology and phonology of low-resource languages - NLP applications for Indigenous languages of the Americas Important dates: - Start of the anonymity period: March 15, 2023 - Submission deadline: April 15, 2023 - Notification of acceptance: May 15, 2023 - Camera ready papers due: May 26, 2023 - Workshop: July 14, 2023 All deadlines are 11.59 pm UTC -12h (anywhere on earth). Link to submission portal: https://softconf.com/acl2023/AmericasNLP2023/ The workshop also includes: - An open-ended machine translation shared task on truly low-resource languages - A mentoring program to support students and newcomers from underrepresented communities (application form: https://forms.gle/afBWauDfDQijXHTy9) We also have a diverse set of invited speakers, focused on bridging the gap between linguists, NLP, and machine learning research! - Steven Bird (linguistics; ethics) - Emiliana Cruz Cruz (linguistics; anthropology; education) - Angela Fan (NLP; machine translation) - Kristine Stenzel (field linguistics; American Indigenous languages) Organizing Committee - Manuel Mager, AWS AI Labs - Arturo Oncevay, University of Edinburgh - Enora Rice, University of Colorado Boulder - Abteen Ebrahimi, University of Colorado Boulder - Shruti Rijhwani, Google Research - Alexis Palmer, University of Colorado Boulder - Katharina Kann, University of Colorado Boulder More information and contact information can be found at http://turing.iimas.unam.mx/americasnlp/. -- Dr. Katharina Kann Assistant Professor of Computer Science University of Colorado Boulder Personal page: https://kelina.github.io Group page: https://nala-cub.github.io

1 0

CfP 4th Workshop on Patent Text Mining and Semantic Technologies (PatentSemTech2023) @ SIGIR 2023
by Krestel, Ralf 15 Feb '23

15 Feb '23

4th Workshop on Patent Text Mining and Semantic Technologies (PatentSemTech2023) ================================================================================ The PatentSemTech2023 workshop aims to establish a long-term collaboration and a two-way communication channel between the IP industry and academia from relevant fields such as natural language processing (NLP), text and data mining (TDM), and semantic technologies (ST) in order to explore and transfer new knowledge, methods, and technologies for the benefit of industrial applications as well as support research in applied sciences for the IP and neighbouring domains. Call for Contributions ====================== PatentSemTech2023 will be held as a full-day event in conjunction with SIGIR 2023. Workshop website: http://ifs.tuwien.ac.at/patentsemtech/ Important Dates =============== Submission deadline: April 25, 2023 Notification: May 23, 2023 SIGIR PatentSemTech2023 workshop: July 27, 2023 Topics of Interest ================== We encourage submissions of high quality research papers on all topics related to the IP domain. Topics of interest include (but are not limited to): * Text mining and retrieval from patents, legal documents, or other scientific-technical information sources * Machine learning methods, in particular deep learning methods for - Representation learning (word and document embeddings) - Query expansion - Clustering and classification - Recommendation - IPC/CPC prediction - Trend detection - Entity extraction * Semantic approaches for - Linking semantic information - Integrating external knowledge sources - Semantic enrichment * Methods and applications for retrieving, mining, and analysing, including - Patent landscaping - Hot spot / White spot analysis - Multi-modal analysis - Technology trend analysis - Innovative user interfaces - Visual user interface concepts Contributions ============= We solicit two types of submissions: full papers and short papers for three tracks: research, demo, and summarization task. Full papers will be limited to 8 pages (including references); short papers will be 4 pages (including references). The submissions will be peer-reviewed by at least two program committee members and evaluated based on innovativeness, novelty, interestingness, and impact. We plan for three tracks: *Research Track* For this track, we solicit contributions from academia that present * Novel applications of existing state of the art methods for the IP domain * Novel methods or tasks in the IP domain * Novel user interfaces for the IP domain * Novel evaluation or analysis insights in the IP domain * Novel benchmark datasets or other resources of interest * A survey or overview related to a particular task in the IP domain *Demo/System Track* We solicit demos, case study, insights, or novel ideas from industry that present * Focused case studies making use of semantic technologies or machine learning * Interesting IP-related task descriptions or best practices for patent analysis * In-use systems or prototype implementations of semantic technologies * Demos on processing or analysing data from the IP domain, or user interfaces * In-use resources related to patents or external resources, e.g., linked open data. *Summarization Task Track* Within the patent text mining community, especially from the industry, there is an interest in developing text mining tools targeting text summarization. * Participants are free to use publicly available data sets to train their models. We recommend exploring US Patents, which many contain the text section SUMMARY OF THE INVENTION. * We will also publish a small training and test data set on the 23rd of February. The provided data set is composed of patents within the field of Green Plastics Technology. * Participants are asked to submit a short (4 pages) scientific paper, which will be peer-reviewed by the workshop organizers. The most interesting submissions will be invited to present their solution at the workshop. * Furthermore, we will have an additional interactive evaluation to reflect a more real-life scenario at the workshop, making it possible to evaluate not only the performance in terms of F1, ROUGE, recall, precision etc., but also efficiency. Therefore the invited participants will be asked to set-up their solutions as a service and provide a REST API. Input will be a patent document (PDF,DOCX), and output should be a summary of not more than 700 words. Submission Guidelines ===================== Submissions must be in English, in PDF, and in the current ACM two-column conference format. Suitable LaTeX, Word, and Overleaf templates are available from the ACM Website: https://www.acm.org/publications/proceedings-template ("sigconf" template for LaTeX; Interim Template for Word). Submissions should be at most 8 (full) or 4 (short) pages (including figures and references) in length. Submissions should be submitted electronically via EasyChair: https://www.easychair.org/conferences/?conf=patentsemtech2023. At least one author of each accepted paper is required to register for, and present the work in person at the workshop. Publication =========== Accepted papers will be published as CEUR proceedings. Selected contributions will be invited to submit extended, full papers to Elsevier's World Patent Information (WPI) journal: https://www.journals.elsevier.com/world-patent-information/ Organizers ========== Ralf Krestel (ZBW & CAU Kiel, Germany), Hidir Aras (FIZ Karlsruhe, Germany), Linda Andersson (Artificial Researcher, Austria), Florina Piroi (Data Science Studio, RSA FG, Austria), Allan Hanbury (TU Wien, Austria), Dean Alderucci (CMU, USA) All questions about submissions should be emailed to: r.krestel(a)zbw.eu and hidir.aras(a)fiz-karlsruhe.de

1 0

February 2023 Newsletter - LDC
by Penn LDC 15 Feb '23

15 Feb '23

In this newsletter: LDC membership discounts expire March 1 30th Anniversary Highlight: Arabic Treebank New publications: 2019 NIST Speaker Recognition Evaluation Test Set - Audio-Visual<https://catalog.ldc.upenn.edu/LDC2023V01> LORELEI Tagalog Representative Language Pack<https://catalog.ldc.upenn.edu/LDC2023T02> ________________________________ LDC membership discounts expire March 1 Time is running out to save on 2023 membership fees. Renew your LDC membership, rejoin the Consortium, or become a new member by March 1 to receive a discount of up to 10%. For more information on membership benefits and options, visit Join LDC<https://www.ldc.upenn.edu/members/join-ldc>. 30th Anniversary Highlight: Arabic Treebank The Penn/LDC Arabic Treebank (ATB) project began in 2001 with support from the DARPA TIDES program and later, the DARPA GALE and BOLT programs. The original focus was on Modern Standard Arabic (MSA), not natively spoken and not homogenously acquired across its writing and reading community. In addition to the expected issues associated with complex data annotation, LDC encountered several challenges unique to a highly inflected language with a rich history of traditional grammar. LDC relied on traditional Arabic grammar, as well as established and modern grammatical theories of MSA -- in combination with the Penn Treebank approach to syntactic annotation -- to design an annotation system for Arabic. (Maamouri, et al., 2004<https://www.ldc.upenn.edu/sites/www.ldc.upenn.edu/files/nemlar2004-penn-ara…>). LDC was innovative with respect to traditional grammar when necessary and when other syntactic approaches were found to account for the data. LDC also developed a wide-coverage MSA morphological analyzer, LDC Standard Arabic Morphological Analyzer (SAMA) Version 3.1 (LDC2010L01<https://catalog.ldc.upenn.edu/LDC2010L01>), which greatly benefited ATB development. Revisions to the annotation guidelines during the DARPA GALE program (principally related to tokenization and syntactic annotation) improved inter-annotator agreement and parsing scores. ATB corpora were annotated for morphology, part-of-speech, gloss, and syntactic structure. Data sets based on MSA newswire developed under the revised annotation guidelines include Arabic Treebank: Part 1 v 4.1 (LDC2010T13<https://catalog.ldc.upenn.edu/LDC2010T13>), Arabic Treebank: Part 2 v 3.1 (LDC0211T09<https://catalog.ldc.upenn.edu/LDC2011T09>), and Arabic Treebank: Part 3 v 3.2 (LDC2010T08<https://catalog.ldc.upenn.edu/LDC2010T08>). Other genres are represented in Arabic Treebank - Broadcast News v 1.0 (LDC2012T07<https://catalog.ldc.upenn.edu/LDC2012T07>) and Arabic Treebank - Weblog (LDC2016T02<https://catalog.ldc.upenn.edu/LDC2016T02>). LDC's later work on Egyptian Arabic treebanks in the DARPA BOLT program benefited from the strides in its MSA treebank annotation pipeline. As for the challenges presented by informal, dialectal material, collaborator Columbia University provided a normalized Arabic orthography to account for instances of Romanized script (Arabizi) in the data and developed a morphological analyzer (CALIMA) in parallel, working in a tight feedback loop with LDC's annotation team. SAMA and CALIMA were synchronized in the Egyptian Arabic treebanks, the former used for MSA tokens and the latter used for Egyptian Arabic tokens. Resulting corpora include BOLT Egyptian Arabic Treebank - Discussion Forum (LDC2018T23<https://catalog.ldc.upenn.edu/LDC2018T23>), Conversational Telephone Speech (LDC2021T12<https://catalog.ldc.upenn.edu/LDC2021T12>), and SMS/Chat (LDC2021T17<https://catalog.ldc.upenn.edu/LDC2021T17>). ATB corpora and its related releases are available for licensing to LDC members and nonmembers. For more information about licensing LDC data, visit Obtaining Data<https://www.ldc.upenn.edu/language-resources/data/obtaining> ________________________________ New publications: 2019 NIST Speaker Recognition Evaluation Test Set - Audio-Visual<https://catalog.ldc.upenn.edu/LDC2023V01> contains approximately 64 hours of English audio-visual data for development and test, answer keys, enrollment, trial files, and documentation from the NIST-sponsored 2019 Speaker Recognition Evaluation (SRE)<https://www.nist.gov/itl/iad/mig/nist-2019-speaker-recognition-evaluation>. The 2019 evaluation task was speaker detection, that is, to determine whether a specified target speaker was speaking during a segment of speech. The evaluation was conducted in two parts: (1) a leaderboard-style challenge based on conversational telephone speech and (2) a separate evaluation using audio-visual data. This release relates to the audio-visual evaluation. The source audio-visual data was collected by LDC for the VAST (Video Annotation for Speech Technology) project. That collection focused on amateur video recordings from various online media hosting services. The recordings vary in duration from 17.5 seconds to 13 minutes; most have two audio channels (stereo), but some are monophonic (one channel). 2023 members can access this corpus through their LDC accounts. Non-members may license this data for a fee. * LORELEI Tagalog Representative Language Pack<https://catalog.ldc.upenn.edu/LDC2023T02> was developed by LDC and is comprised of approximately 4.8 million words of Tagalog monolingual text, 341,000 words of found Tagalog-English parallel text, and 124,000 Tagalog words translated from English data. Approximately 78,000 words were annotated for named entities and over 26,000 words were annotated for entity discovery and linking and situation frames (identifying entities, needs and issues). Data was collected from discussion forum, news, reference, social network, and weblogs. The LORELEI (Low Resource Languages for Emergent Incidents) program was concerned with building human language technology for low resource languages in the context of emergent situations. Representative languages were selected to provide broad typological coverage. The knowledge base for entity linking annotation is available separately as LORELEI Entity Detection and Linking Knowledge Base (LDC2020T10)<https://catalog.ldc.upenn.edu/LDC2020T10>. 2023 members can access this corpus through their LDC accounts. Non-members may license this data for a fee. To unsubscribe from this newsletter, log in to your LDC account<https://catalog.ldc.upenn.edu/login> and uncheck the box next to "Receive Newsletter" under Account Options or contact LDC for assistance. Membership Coordinator Linguistic Data Consortium<ldc.upenn.edu> University of Pennsylvania T: +1-215-573-1275 E: ldc(a)ldc.upenn.edu<mailto:ldc@ldc.upenn.edu> M: 3600 Market St. Suite 810 Philadelphia, PA 19104

1 0

2nd Deadline extension to 31/03: JLM spec. issue on Computational Approaches to Morphological Typology
by Sacha Beniamine 15 Feb '23

15 Feb '23

Dear colleagues, We are further extending the deadline for this call to the 31st of March. Please find below the updated call: ------------- We invite researchers in the broad area of computational morphology to submit their recent, unpublished work to a special issue of the Journal of Language Modelling <https://jlm.ipipan.waw.pl/index.php/JLM><https://jlm.ipipan.waw.pl/index.php/JLM>. Motivation: Computational techniques have a long history of use in the study of morphology, where they have been used both for practical tasks such as the analysis and production of complex word forms and for theoretical ones such as structural and informational analysis of morphological systems. As both systems and datasets improve, these techniques are increasingly developed and evaluated on a typologically diverse array of languages, including many which are endangered or lack large-scale resources. Detailed comparisons across languages can help to reveal typological biases or assumptions within existing computational techniques [1, 2]. Alternatively, computational methods and analyses can also shed light on questions within linguistic typology [3, 4, 5, 6]. The goal of this special issue is to bring researchers from multiple communities together in exploring issues of linguistic typology across a wide range of different languages and phenomena. We encourage the submission of work on endangered or less-studied languages. The Journal of Language Modelling is a free (for readers and authors alike) open-access peer-reviewed journal. All articles are peer-reviewed by at least 3 reviewers, usually including at least one member of the Editorial Board. Topics of interest: - Typological clustering or classification of languages - Investigation of particular linguistic features which improve or detract from the performance of computational morphology tools - Comparison of morphological structures (e.g., inflection classes, implicative networks) across typologically different languages - Investigation of diachronic typological change using computational methods - Creation, curation or analysis of typological databases via computational methods Submissions: The submissions should be journal papers, not proceedings papers, totalling 25-50 pages, excluding references. Authors are advised to use the online manuscript submission for the journal. Make sure to select the special issue when asked to provide the article type. More information, including formatting instructions for authors can be found on the journal's webpage at: https://jlm.ipipan.waw.pl/index.php/JLM/about/submissions. An adaptation of the LaTeX template for overleaf can be found at: https://fr.overleaf.com/latex/templates/template-for-journal-of-language-mo…. Important dates: Call for papers issued: 15/7/2022 Submissions due: 15/1/2023 --- extended to 31/03/2023 Author notification: Spring 2023 Guest editors: Sacha Beniamine (University of Surrey) Micha Elsner (The Ohio State University) Katharina Kann (University of Colorado, Boulder) References [1] Ryan Cotterell, Christo Kirov, John Sylak-Glassman, David Yarowsky, Jason Eisner, and Mans Hulden. 2016a. The SIGMORPHON 2016 shared Task— Morphological reinflection. In Proceedings of the 14th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology, pages 10–22, Berlin, Germany. Association for Computational Linguistics. [2] Huiming Jin, Liwei Cai, Yihui Peng, Chen Xia, Arya McCarthy, and Katharina Kann. 2020. Unsupervised morphological paradigm completion. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 6696– 6707, Online. Association for Computational Linguistics. [3] Neil Rathi, Michael Hahn, and Richard Futrell. 2021. An Information-Theoretic Characterization of Morphological Fusion. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10115–10120, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. [4] Parker, J., Reynolds, R., & Sims, A. (2022). Network Structure and Inflection Class Predictability: Modeling the Emergence of Marginal Detraction. In A. Sims, A. Ussishkin, J. Parker, & S. Wray (Eds.), Morphological Diversity and Linguistic Cognition (pp. 247-281). Cambridge: Cambridge University Press. DOI: 10.1017/9781108807951.010 [5] Guzmán Naranjo, Matías and Becker, Laura. Statistical bias control in typology. Linguistic Typology, to appear, 2021. DOI: 10.1515/lingty-2021-0002 [6] Sacha Beniamine. 2021. One lexeme, many classes: Inflection class systems as lattices. In Berthold Crysmann & Manfred Sailer (eds.), One-to-many relations in morphology, syntax, and semantics, 23--51. Berlin: Language Science Press. DOI: 10.5281/zenodo.4729789

1 0

Senior Research Associate at the Alan Turing Institute, Foundation Models and Commonsense Reasoning
by Eric Atwell 15 Feb '23

15 Feb '23

Senior Research Associate at the Alan Turing Institute, Foundation Models and Commonsense Reasoning Position In 2022, the Alan Turing Institute signalled its intention to establish a portfolio of foundational AI research, which would complement the strengths of the institute around applications of AI and AI policy. An initial portfolio of research in foundation models, game theory, and probabilistic programming will be launched in early 2023. Each of these areas is called a ‘Pillar’. It is intended that this portfolio will complement the UK’s current activity, rather than duplicating existing efforts, and aiming to promote emerging new areas that show promise for the future. FOUNDATION MODELS Foundation models are large ML models trained on large, broad data sets. Foundation models such as GPT-3 have been shown to have remarkable capabilities for generating realistic natural language, and, to some extent, capabilities for problem solving and common-sense reasoning. Developing a Turing Foundation Model is beyond our present capacity. Instead we therefore propose work aimed at developing Turing expertise around the problem of precisely understanding the capabilities of such models. The main issue we aim to address is that of *benchmarking* such models: although such models appear to be very capable in some respects, they fail on apparently simple tasks, in unpredictable ways. In short, we don't have a clear understanding of the capabilities and shortcomings of such systems - which raises concerns for their use. ROLE PURPOSE We are looking for a Senior Research Associate to support and enable the delivery of the Foundation Model theme, under the direction of Anthony (Tony) Cohn, and in collaboration with Michael Wooldridge and Nigel Shadbolt. The successful candidate will primarily focus on evaluating the extent to which, and the conditions under which existing Foundational Models can support common-sense reasoning (such as naive physics, spatial and temporal reasoning, the concept of agency, and causality). This work will be taken in parallel and in collaboration with researchers working on other aspects of the Foundation Models Pillar. The candidate will join a vibrant team of researchers and will have opportunities to engage with cutting-edge projects and experts at leading universities. The post can be based either at The Alan Turing Institute site in London, or at the University of Leeds. In either case you will need to travel to the other site when required (travel expenses will be paid as appropriate). Further details and the online application form can be found here: https://cezanneondemand.intervieweb.it/turing/jobs/senior-research-associat…<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcezanneon…>

1 0

2024

2023

2022

Corpora February 2023