[CPM-SPIRE-L] Competition on Applied Text Searching (CATS'2014) - Call for Tasks
Simone Faro
faro at dmi.unict.it
Tue Oct 1 01:35:51 PDT 2013
===========================================================================
CALL FOR TASKS:
FIRST COMPETITION ON APPLIED TEXT SEARCHING, CATS'2014
Prague, Czech Republic, September 2014
http://www.cats-series.org
===========================================================================
The CATS committee cordially invites proposals for tasks to be run
as part of the "Competition on Applied Text Searching 2014 (CATS'14)"
which is intended to be an ongoing series of evaluations of text
searching algorithms. The main goal of the CATS series is to explore
the potentialities of the most recent advances in text searching on
real world problems. We particularly aim to provide a fruitful
platform, where the practical demands would be better sensed by the
academia, and the advances in algorithm design and engineering would
be more apparent to the industry.
SUBMISSION DEADLINE: DECEMBER 1 2013
Pdf version of the Call for Tasks:
http://www.cats-series.org/CATS14-CallForTasks.pdf
COMPETITION TASKS
Tasks are typically focused on the design and implementation of
efficient, correct algorithms, in the field of text searching,
including exact, approximate, single and multiple strings.
Any problem definitions extending these fundamental topics as well
as introducing new concepts or dimensions are surely very welcome.
We especially encourage tasks on diverse data sets that are relevant
in (but not restricted to) searching biological/DNA sequences, musical
sequences, binary data, encrypted texts, natural language texts,
compressed data, numerical sequences, dictionaries, images, specialized
texts, etc. Tasks based on purely theoretical or open-ended problems,
i.e., tasks that do not necessarily have a known efficient or optimal
solution may also be considered. In typical setting, the tasks are
expected to evaluate solutions in terms of correctness and speed
(CPU running times). However, tasks with alternative measures of
solution effectiveness other than CPU time consumption will be also
taken into account. It will help us a lot if tasks are designed in such
a way that they can be run and scored automatically from a central
website or from a freely available tool. We will be happy to help with
task design, data formatting, and so on. Task proposals must
be submitted
in English and must include an email address and information
on the task
author(s) including name, affiliation, and country.
To ensure a fair and interesting competition for all participants,
proposed tasks should ideally contain:
1. Abstract
The abstract will contain a short description of the task and of the
data which will be used for the evaluation.
2. Introduction and Motivation
This section will present an introduction to the problem
containing some
historical informations and the most significant references.
It will include also applications and motivations of the problem
proposed in the task.
3. Task Description
This section will present a detailed description and a formal
definition
of the problem involved in the proposed task. It will include technical
details about input and output data.
4. Evaluation Procedure and Environment
This section will describe very clearly the procedure to evaluate the
submitted solutions. It is expected that all submissions for the task
will be checked for correctness and then compared in terms of speed via
an automatic system. The proposers may consider to use the
smart research
tool (http://www.dmi.unict.it/~faro/smart), which is a platform to
automatically score and compare string matching algorithms. It might be
also possible, and we will be happy, to extent and tune the smart tool
according to the needs of the tasks. However, the proposers are surely
free to offer their own evaluation procedure with their own
tools according
to their specific needs. It is also vital to announce the
details of the
computing device that will be used throughout the evaluation,
such as the
CPU information, memory, operating system, or any other task specific
properties.
5. Data Format
This section will explain the data format which will be used for the
evaluation, and the description of the text and of the set of patterns.
6. Basic Solution
While not strictly required, it is strongly preferred that the
submission
also includes at least one implementation of the desired
solution, attached
as a separate file. It will be considered as the basic solution to be
outperformed in the final evaluations process.
7. Reward Proposal
Although not mandatory, we strongly suggest the task authors to offer a
reward for the winner, which will help to receive more attention to the
specific task. In this case description of the reward and the
constraint
to comply with the reward must be given. The rewards may be
considered in
the form of cash money, grant for attending the conference or
even a job
contract. However, the reward proposal is at author's complete
discretion.
Submitted tasks must be kept strictly confidential until the end of
CATS 2014. After this time, authors are free to do whatever
they wish with
the tasks. A set of task proposals will be selected by the
CATS Committee
on the basis of their interest to the community, peculiarity
and novelty.
The set of final selected tasks will be announced by December10, 2013.
All authors will be acknowledged at the close of CATS 2014.
SUBMISSION
Please submit proposals as soon as possible, preferably in a pdf file.
Task materials must be placed together in a single file (use
.zip or .tgz
for multiple files) and submitted via email to the address:
tasksubmission at cats-series.org.
Task proposals must be submitted by December 1, 2013
WHAT HAPPENS NEXT
Receipt of submissions will be confirmed via email.The CATS Scientific
Committee will carefully review all submissions and will select a list
of tasks to be run at the CATS 2014 competition. The authors
of the tasks
selected for competition will be invited to attend the CATS 2014 as
Invited Guests. Following the announcement of the selected
tasks, the CATS
Committee will publish corresponding Call for Participation to invite
solutions for the selected tasks.The authors of the selected
tasks will be
asked to provide training data to competitors in order to test
and evaluate
their own solutions. Training data will be released to competitors by
December 25 2013.At the end of the Call for Participation a
set of evaluation
data will be provided to the CATS committee by the task
authors. This data
will be similar (but not equal) to training data provided in the first
phase of the process.Task authors could eventually ask to keep
confidential
the evaluation data and/or to participate during the testing process of
the submitted solutions. In this case authors should
explicitly state their
interest in keeping confidential the evaluation data. In any
case training
data must be released to the participants.Solutions will be
checked first
for correctness, and will be evaluated in terms of running
times or in terms
of some other measures of effectiveness. The authors of tasks
included in the
actual competition will also be recognized by listing their
name, affiliation
and country in the official CATS 2014 proceedings (unless they
specifically
decline this).
IMPORTANT DATES
1 December 2013 : Task submission deadline
1 January 2014 : Call for solutions
May 2014 : Registration deadline
June 2014 : Winners announcement
September 2014 : CATS Workshop
COMMITTEE
Simone Faro
Dipartimento di Matematica e Informatica, University of Catania, Italy
email: faro at dmi.unict.it
M. Oguzhan Kulekci
TUBITAK, National Research Institute of Electronics and
Cryptology, Turkey
email: oguzhan.kulekci at tubitak.gov.tr
More information about the CPM-SPIRE-L
mailing list