The 2nd Computational Linguistics Scientific Document Summarization Shared Task
You are invited to participate in the CL-SciSumm 2016 Shared Task, as part of the Joint Workshop of Bibliometric-enhanced IR and NLP for Digital Libraries (BIRNDL) at JCDL 2016 on 23 June 2016
As a part of the Joint workshop on Bibliometric-enhanced Information Retrieval and Natural Language Processing for Digital Libraries (BIRNDL) at JCDL 2016, we are pleased to announce the 2nd CL-SciSumm Shared Task on scientific paper summarization. This task follows up on the successful CL Pilot Task conducted as a part of the BiomedSumm Track at the Text Analysis Conference 2014 (TAC 2014). Nine teams from four countries expressed an interest in participating in the shared task; three teams submitted system descriptions and findings.
The current shared task will be on automatic paper summarization in the Computational Linguistics (CL) domain. The output summaries will be of two types: faceted summaries of the traditional self-summary (the abstract) and the community summary (the collection of citation sentences ‘citances’). We also propose to group the citances by the facets of the text that they refer to.
=== The Task ===
Given: A topic consisting of a Reference Paper (RP) and up to ten Citing Papers (CPs) that all contain citations to the RP. In each CP, the text spans (i.e., citances) have been identified that pertain to a particular citation to the RP.
Task 1a: For each citance, identify the spans of text (cited text spans) in the RP that most accurately reflect the citance. These are of the granularity of a sentence fragment, a full sentence, or several consecutive sentences (no more than 5).
Task 1b: For each cited text span, identify what facet of the paper it belongs to, from a predefined set of facets.
Evaluation: Task 1 will be scored by overlap of text spans in the system output vs the gold standard created by human annotators.
=== The Corpus ===
The CL-SciSumm corpus is created by randomly sampling documents from the ACL Anthology corpus and selecting their citing papers. For CL-SciSumm 2016, we have selected three portions of this source collection to be annotated and serve as training, development and test collections. The training set of 10 articles is available for download at GitHub https://github.com/WING-NUS/scisumm-corpus and can be used by participants to pilot their systems. Watch for updates to the GitHub repository, as we are still updating a few training files. We will finalise the training set by 29 February. The development set of 10 articles, an additional part of the same corpus, will be released in April, which participants can add to the training set to tune their system parameters. Finally the test set of 10 articles will be released in May. The system outputs from the test set should be submitted to the task organizers, for the collation of the final results to be presented at the workshop.
We thank Microsoft Research Asia for their support of the shared task in underwriting the project's cost, and IIT Hyderabad for providing annotation support.
=== Registration ===
Organizations wishing to participate in the CL Shared Task track at BIRNDL 2016 are invited to register on EasyChair: https://easychair.org/conferences/?conf=birndl2016 by 30 March 2016. Participants are advised to register as soon as possible in order to receive timely access to evaluation resources, including training development and testing data. Registration for the task does not commit you to participation - but is helpful to know for planning. All participants who submit system runs are welcome to present their system at the BIRNDL Workshop.
Dissemination of CL-SciSumm work and results other than in the workshop proceedings is welcomed, but the conditions of participation specifically preclude any advertising claims based on these results. Any questions about conference participation may be sent to the organizers mentioned below.
=== Important Dates ===
February 2016: Training set posted March 30, 2016: Deadline for expression of interest and short system descriptions due April 8, 2016: Development set posted April 22, 2016: Notification of acceptance of presentation proposals April 29, 2016: Test set posted May 20, 2016: System reports and system runs from the test set due June 3, 2016: Camera ready contributions due June 23 2016: Participants present at BIRNDL 2016 workshop in Newark, New Jersey, USA
The CLSciSumm16 Task is expected to be of interest to a broad community including those working in computational linguistics and natural language processing, text summarization, discourse structure in scholarly discourse, paraphrase, textual entailment and text simplification.