International Conference on the Computational Processing of Portuguese
September 24-26, 2018 – Canela, Brazil
POP – Por Outras Palavras
POP @ PROPOR 2018
In Other Words
1st Workshop on Linguistic Tools and Resources for Paraphrasing in Portuguese
The workshop is a venue to present and discuss novel research on linguistic tools and resources for paraphrasing in Portuguese
CALL FOR PAPERS
- Full and short paper submission deadline: Jul 31, 2018 – 23:59 (GMT-11)
- Notification of paper acceptance or rejection: Aug 20, 2018
- Camera-ready papers due: Aug 30, 2018
- Workshop: Sep 24, 2018
Submission Web page:
=== WORKSHOP DESCRIPTION
“POP – Por Outras Palavras: Linguistic Tools and Resources for Paraphrasing in Portuguese” aims to bring together linguists and NLP researchers who are interested in discussing new ideas on the development and use of linguistic resources in real-world, paraphrase-driven applications for Portuguese.
Paraphrases are extremely important in human communication, both in language production and understanding, and they take on a greater role in ongoing and future research activities and projects. Several linguistic experiments have shown the feasibility of using paraphrastic resources in a wide variety of software applications, as they allow to recognize and generate equivalent forms of expressing the same content, enabling systems to provide the user suggestions to say and write the same thing/idea in other words, adding fluency, creativity and diversity to style. In the current stage of development, paraphrasing systems require linguistic knowledge and context-aware intelligence to “understand” and recognize a wide variety of targeted expressions. These include semantico-syntactic units, both simple words and idiomatic, frozen and semi-frozen expressions, in order to transform them into semantically equivalent constructions or expressions. Paraphrases also have a significant impact on technological and applicational solutions related to e-learning, information retrieval, question-answering and machine translation, among others. For Portuguese, the usefulness of paraphrastic capabilities has already been explored in application scenarios, such as a dialogue system, to increase the linguistic knowledge of an intelligent virtual agent, in summarization and simplification tools, and also in tools aiming at achieving higher quality machine translation. However, more research is needed for a long-term viability and success of a paraphrasing system in the areas of text production and revision, namely in developing and improving online authoring platforms, developing interactive programs to help second language learners produce different, but equivalent sentences, or even aiming at Portuguese-speaking students, as an aid in the tasks of producing and revising texts.
Industrial paraphrase applications require a large number of linguistic resources of a very high quality, such as extensive dictionaries, transformational grammars, etc. to generate paraphrases. Relevant public resources have been made available to establish relationships between semantically related elements, for morphological, syntactic and semantic analysis, disambiguation, identification of multiword units, parsing discontinuous expressions (e.g., multiword units that involve insertions, which are useful for paraphrasing purposes). Some efforts have been undertaken to integrate and improve linguistic resources available for Portuguese, which were previously dispersed and lacked organic articulation among them, such as lexicon-grammar tables, among others. The availability of these resources makes it possible to paraphrase constructions involving support verbs (for example, paraphrasing the verbal construction with the nominal construction and vice-versa, or alternation between the support verb and other stylistic or aspectual verbs), active and passive constructions, adverbial compounds, relatives and participles, among others. The recent collaboration aims at creating synergies between dispersed researches and enhancing their potential to address more complex challenges, testing systems with larger volumes of data studied and validated by linguists. Even though the initial implementation efforts are beginning to bear fruit, there are still challenges to face and problems to address, which should be discussed in the workshop, such as adaptation to Portuguese variants, paraphrases of technical versus non-expert expressions, evaluation of paraphrasing systems, extending and improving paraphrase-oriented resources and achieving enhanced collective quality sets of resources for advancing the state of the art in Portuguese paraphrasing.
=== WORKSHOP OBJECTIVES
The objectives of the POP workshop are:
- Gather researchers and students with an interest in the field of paraphrasing with a special focus on Portuguese to teach, learn and share information on the topic;
- Collate a set of high-quality papers discussing the latest trends in the field, and contributing to improving the state of the art of paraphrasing in Portuguese;
- Exchange ideas and disseminate best practices to help foster research in the field;
- Foster the convergence of research efforts for a consensual common definition of the scientific methods, and encourage international co-operation in order to achieve common strategies to respond to current technological needs;
- Discuss new methodologies, such as neural nets, etc., and how to combine those methodologies with linguistic efforts;
- Discuss the challenges ahead and exchange information on scientific and technological aspects;
- Encourage and bolster the creation of parallel corpora of paraphrases for Portuguese as datasets for the gathering of paraphrastic alignment resources, and training and testing of paraphrasing systems;
- Locate sources of public funding boost further research, support innovation, and development on this enabling technology.
=== WORKSHOP GOALS
The goals of the POP workshop are:
- Create a forum for discussion where participants meet and share their experience as developers, researchers and teachers;
- Present the latest linguistic resources and NLP applications developed for paraphrasing in Portuguese, as well as discuss future developments;
- Offer researchers and graduate students broader opportunities for development and introduce a broader range of useful instruments (tools and resources);
- Build a community of “paraphrase researchers” for the Lusophone world;
- Help researchers build NLP applications using paraphrases;
- Provide an occasion to present and discover the recent developments in the field.
=== NON-EXHAUSTIVE LIST OF TOPICS OF INTEREST
The workshop invites papers on a range of topics, including (but not limited to):
— Importance of paraphrases in NLP, paraphrase functions and software applications that can parse sentences and produce paraphrases
— Development of new linguistic resources for paraphrasing purposes
* Collaborative linguistic resources and resource integration
* Expert-crafted resources, such as large-coverage dictionaries and grammars with semantico-syntactic features aiming the identification of paraphrases in corpora
* Complex and dynamic libraries of resources
* Transformational analysis and local grammars for paraphrase generation
* Lexicon-grammar tables in paraphrasing
— Analysis, identification, and alignment of paraphrases in corpora
— Methods to collect paraphrastic data
— Paraphrastic multiword unit alignment
— Practical, text-processing applications using paraphrases
— Architecture, design, implementation and management of paraphrasing systems
— Practical experiences and testbeds related to the use of paraphrases
— Paraphrases involving distinct styles, language domains, expert/non-expert language, etc.
— Industrial experiences with paraphrases
=== PAPER SUBMISSION
Submissions are accepted in PROPOR conference style in two formats:
– short papers (no more than 4 pages + references)
– full papers (no more than 8 pages + references)
Submissions should begin with an abstract of 250 words or less.
Documents must be submitted in PDF format, following the same style of PROPOR 2018 (Springer LNCS).
Authors of accepted full papers will have 30 minutes for presentation and discussion during the workshop, while authors of short papers will have 15 minutes. At least one author of an accepted paper must register at the workshop.
Submission website will be announced soon
=== PAPER PROCEEDINGS AND PUBLICATION OF A SPECIAL ISSUE
Accepted papers for the POP Workshop will be made available in an online open-access publication repository containing the conference proceedings on the PROPOR 2018 website at the time of the Conference.
Selected papers will be considered for publication (in English) in a Special Issue in a yet-to-be-decided International Journal in connection with this workshop. We will issue an open call after the workshop, submissions will go through a separate peer review process.
=== LANGUAGES OF THE WORKSHOP
The Workshop will be in Portuguese and English. Papers can be written in Portuguese or English.
=== REGISTRATION FEES
The participation in the workshop, for authors or non-authors of papers alike, is free of charge for participants of the main conference. The fees for attendees who are not participants of the main conference is to be arranged with PROPOR’s organizers.
=== PROGRAM COMMITTEE
|Alberto Simões, 2Ai Lab/IPCA (Portugal)|
|Amanda Pontes Rassi, Lionbridge (Brazil)|
|Ariani Di Felippo, Universidade Federal de São Carlos (Brazil)|
|Belinda Maia, Centro de Linguística da Universidade do Porto (Portugal)|
|Cláudia Freitas, PUC-Rio (Brazil)|
|Diana Santos, Linguateca (Portugal) & ILOS, UiO (Norway)|
|Éric Laporte, Université Paris-Est Marne-la-Vallée (France)|
|Horacio Rodríguez, Universitat Politècnica de Catalunya – UPC (Spain)|
|Hugo Gonçalo Oliveira, Universidade de Coimbra (Portugal)|
|Lucília Chacoto, Universidade do Algarve and CLUL (Portugal)|
|Luísa Coheur, IST and INESC-ID Lisboa (Portugal)|
|M. Antonia Martí, Universitat de Barcelona (Spain)|
|Maria José Bocorny Finatto, Universidade Federal do Rio Grande do Sul (Brazil)|
|Max Silberztein, Université de Franche-Comté (France)|
|Nuno Mamede, INESC-ID Lisboa (Portugal)|
|Oto Vale, Universidade Federal de São Carlos (Brazil)|
|Paolo Rosso, Universitat Politècnica de València (Spain)|
|Paulo Gamallo, Universidade de Santiago de Compostela (Spain)|
|Ricardo Ribeiro, ISCTE-IUL and INESC-ID Lisboa (Portugal)|
|Sandra Maria Aluísio, Universidade de São Paulo, ICMC (Brazil)|
|Thiago Alexandre Salgueiro Pardo, Universidade de São Paulo (Brazil)|
=== WORKSHOP ORGANIZATION
Anabela Barreiro, L2F/INESC-ID Lisboa, Portugal (Workshop Main Chair)
Jorge Baptista, Universidade do Algarve and INESC-ID Lisboa, Portugal
Paulo Quaresma, Universidade de Évora, Portugal
Renata Vieira, PUC-RS, Brazil
Questions about POP 2018 may be sent to [email protected]