View source for 2020Q3 Reports: SIGMORPHON

This is the 2019-2020 annual report for SIGMORPHON, ACL's special interest group for computational morphology, phonology, and phonetics.

== Membership == 
The SIG membership is documented in the following table, observing 15% growth from 150 to 176 members in the past 5 years.  We are planning to hold a membership drive to further promote growth of the SIG.

{|
! Year !! Membership
|-
| 2020 || 176
|-
| 2019 || 176
|-
| 2018 || 157
|-
| 2017 || 167
|-
| 2016 || 157
|-
| 2015 || 150
|-
| 2014 || 144
|-
| 2013 || 142
|-
| 2012 || 137
|-
| 2011 || 120
|-
| 2010 || 105
|-
| 2009 || 96
|-
| 2008 || 90
|-
| 2007 || 73
|-
| 2006 || 55
|}

== Workshop == 

The 17th SIG workshop will be held at ACL 2020, chaired by Garrett Nicolai, Kyle Gorman and Ryan Cotterell. There is a push now to hold the workshop every year to increase interest in the SIG’s target areas among the wider ACL community. This is especially important now as our community is undergoing rapid growth. 

The 2020 Workshop had 14 submissions from which the program committee selected 8 for publication (57% acceptance rate).  There will also be invited talks.  This will be the first SIGMORPHON workshop to be held entirely online. 

== Shared Tasks ==

In 2020, SIGMORPHON looked to build off the success of its previous shared tasks, and expanded its shared task offerings to three.  Each task was separately proposed by an organizing committee, and separately approved by the SIG Executive Committee, who worked with the organizing committee in some cases.  This structure was inspired by other groups (SIGSEM and SIGNLL) that run multiple shared tasks, and we expect to keep it in future years. 

=== Task 0 ===
SIGMORPHON's fifth installment of its inflection generation shared task focuses on languages that are typologically diverse from languages in our previous tasks. Many of these languages are extremely low-resource. This edition of the task is particularly interested in inflection generation systems' ability to generalize to new languages, including languages that are typologically distinct. 10 teams submitted 22 systems, of which 19 were neural.  Data hallucination proved to be a popular tactic this year, following the methods of the best system from last year’s task.

=== Task 1 ===
This new task, the first of its kind at SIGMORPHON, focuses on grapheme-to-phoneme conversion. This technology is a key component of speech recognition and synthesis engines, but much of the existing published research is either limited to a small number of closely related languages/scripts, or uses proprietary data sets, limiting replicability. The training and development data consists of words and corresponding IPA pronunciations extracted from Wiktionary, a free online encyclopedia, in 15 languages and scripts. 9 teams submitted a total of 23 different systems.

=== Task 2 ===
Task 2 fills the gap between recent SIGMORPHON shared tasks on morphological inflection learned from limited training data and completely unsupervised morphological generation by proposing the task of unsupervised morphological paradigm completion. The goal was to generate complete inflection tables exclusively from raw text and a lemma list for a known part of speech. 3 teams submitted a total of 7 different systems to tackle this new task.

== Other Activities (Online) ==

=== Migration of group mailing list ===
This year saw the migration of the SIG’s mailing list from an internal server at Johns Hopkins University to a Google Group. This migration serves to facilitate coordination between the group and several other SIG functions, such as elections which were conducted through Google in 2019. The SIG also created a similar Google Group to replace the Exec mailing list. 

=== Mentorship program ===
This year saw the initiation of a mentorship program, which will take place at ACL. Due to the fact that ACL 2020 will be held online, the mentorship program is virtual; if they take place, future iterations should also be in person. The program was spearheaded by Ryan Cotterell, but at the suggestion of Kyle Gorman who forwarded a request from a SIGMOPHON member inquiring if such a program is possible to the SIGMORPHON Exec. Interested mentees were advised to sign up online on a Google form [https://docs.google.com/forms/d/e/1FAIpQLSfjEixHNavRJKr2pY-dkq3Y1ZR70OvwbMs6sYmoNgF5zb4pAg/viewform?usp=send_form], and were paired with more senior researchers who will serve as mentors to them. At the time of writing, 10 people have signed up; all who signed up were either graduate students or prospective graduate students; some did not have papers at the conference. One mentee who signed up early has been paired so far; we are working on pairing the rest. The success of the program will not be known until after the workshop in July, but this early success suggests that we should run the program again next year. 

=== Membership drive ===
With the successful migration of the mailing list to Google groups, we are expecting to conduct the planned membership drive in the coming year.

== Elections ==

An election of the executive committee of the SIG was held in January, 2019. The elected officers of the SIG are listed at [https://sigmorphon.github.io/], and are serving a 2-year term. The next elections are scheduled to take place in January, 2021.