Monday, May 15, 2017

LDC May 2017 Newsletter

In this newsletter:

Recent Collaborations

New publications:
______________________________________________________________

Recent Collaborations
Collaborations play an important role in many LDC activities. Over the past twenty-five years, LDC has partnered, consulted, and otherwise “collaborated” with a variety of organizations to advance research community goals. Recently, LDC partnered with Oxford Wave Research to integrate its latest speech technology into data collection and annotation processes. LDC also supports the Hearables Challenge sponsored by the National Science Foundation by creating and distributing training and test corpora. Finally, LDC Executive Director Chris Cieri is working with international colleagues to plan LREC2018 as a member of the Conference Programme Committee.
LDC welcomes new collaborations. Let us know what interests you and how we can work together. Contact LDC to begin the conversation.

New publications:

(1) IARPA Babel Lao Language Pack IARPA-babel203b-v3.1a was developed by Appen for the IARPA (Intelligence Advanced Research Projects Activity) Babel program. It contains approximately 207 hours of Lao conversational and scripted telephone speech collected in 2013 along with corresponding transcripts.

The Babel program focuses on underserved languages and seeks to develop speech recognition technology that can be rapidly applied to any human language to support keyword search performance over large amounts of recorded speech.

The Lao speech in this release represents that spoken in the Vientiane dialect region in Laos. The gender distribution among speakers is approximately equal; speakers' ages range from 16 years to 60 years. Calls were made using different telephones (e.g., mobile, landline) from a variety of environments including the street, a home or office, a public place, and inside a vehicle.

IARPA Babel Lao Language Pack IARPA-babel203b-v3.1a is distributed via web download.

2017 Subscription Members will receive copies of this corpus provided they have submitted a completed copy of the special license agreement. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*

(2) Multi-Language Conversational Telephone Speech 2011 -- Turkish  was developed by LDC and is comprised of approximately 18 hours of telephone speech in Turkish. The data was collected primarily to support research and technology evaluation in automatic language identification, and portions of these telephone calls were used in the NIST 2011 Language Recognition Evaluation (LRE).

Participants were recruited by native speakers who contacted acquaintances in their social network. Those native speakers made one call, up to 15 minutes, to each acquaintance. The data was collected using LDC's telephone collection infrastructure, comprised of three computer telephony systems. Human auditors labeled calls for callee gender, dialect type and noise. Demographic information about the participants was not collected.

LDC has also released the Multi-Language Conversation Telephone Speech 2011 -- Slavic Group (LDC2016S11)

Multi-Language Conversational Telephone Speech 2011 -- Turkish is distributed via web download.

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*

(3) Phrase Detectives Corpus was developed by the School of Computer Science and Electronic Engineering at the University of Essex and consists of approximately 19,012 words across 40 documents anaphorically-annotated by the Phrase Detectives game, an online interactive "game-with-a-purpose" designed to collect data about English anaphoric coreference.

The documents in the corpus are taken from Wikipedia articles and from narrative text in Project Gutenberg. Annotations are comprised of a gold standard version created by multiple experts, as well as a set created by a large non-expert crowd (via the Phase Detectives game).

The data was annotated according to a prevalent linguistically-oriented approach for anaphora used in several tasks, including OntoNotes Release 5.0 (LDC2013T19), SemEval-2010 Task 1 Ontonotes English: Coreference Resolution in Multiple Languages (LDC2011T01) and The ARRAU Corpus of Anaphoric Information (LDC2013T22).

Phrase Detectives Corpus is distributed via web download.

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data at no cost.

*

(4) The EventStatus Corpus was developed by researchers at Texas A&M University, Stanford University and The University of Utah. It consists of approximately 3,000 English and 1,500 Spanish news articles about civil unrest events annotated with temporal tags.

This corpus was designed to support the study of the temporal and aspectual properties of major events, that is, whether an event has already happened, is currently happening or may happen in the future. Since it focuses on a single domain (civil unrest events), it may be appropriate for tasks such as event extraction and temporal question answering.

The relevant news articles were sourced from English Gigaword Fifth Edition (LDC2017T09) and Spanish Gigaword Third Edition (LDC2011T12). The civil unrest events include protests, demonstrations, marches and strikes.

The EventStatus Corpus is distributed via web download.

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee. 

Monday, April 17, 2017

LDC April 2017 Newsletter

LDC celebrates 25 years

LDC data and commercial technology development

New publications:
_________________________________________________________________________

LDC celebrates 25 years
April 2017 marks the beginning of LDC’s 25th year as the leader in language resource development and distribution. Founded in 1992, the Consortium has grown from a data repository to a vibrant data center that creates, shares and archives language resources. The Catalog continues to grow, boasting over 700 titles in more than 90 languages. With the support of members, licensees, sponsors and collaborators, LDC has distributed over 120,000 copies of data to more than 3,500 organizations worldwide. Our heartfelt thanks for your support as we continue our mission to provide large quantities of diverse data, research program support and high quality member services.

LDC data and commercial technology development
Any organization wishing to use LDC data to develop or test products for commercialization or use LDC data in any commercial product or for any commercial purpose, must first license the data as a For-Profit Member. Once the data is licensed under the For-Profit Membership, the organization retains perpetual rights to use the data for commercial technology development. LDC data users should consult corpus-specific license agreements for limitations on the use of certain corpora. Visit our Licensing page for more information. 

New Corpora
(1) 2010 NIST Speaker Recognition Evaluation Test Set was developed by LDC and NIST (National Institute of Standards and Technology). It contains 2,255 hours of American English telephone speech and interview speech recorded over a microphone channel used as test data in the NIST-sponsored 2010 Speaker Recognition Evaluation (SRE).

The telephone speech segments include two-channel excerpts of approximately 10 seconds and 5 minutes. There are also summed-channel excerpts in the range of 5 minutes. The microphone excerpts are 3-15 minutes in duration. As in prior evaluations, intervals of silence were not removed.

The 2010 evaluation includes not only conversational telephone speech (CTS) recorded over ordinary telephone channels for the core training and test conditions, but also CTS and conversational interview speech recorded over a room microphone channel. Unlike prior evaluations, some of the conversational telephone style speech was collected in a manner to produce particularly high, or particularly low, vocal effort on the part of the speaker of interest. In addition to evaluation data, this package also consists of answer keys, trial and train files, development data and evaluation documentation.

2010 NIST Speaker Recognition Evaluation Test Set is distributed via hard drive.

2017 Subscription Members will receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.
*
(2) BOLT Egyptian Arabic SMS/Chat and Transliteration was developed by LDC and consists of naturally-occurring Short Message Service (SMS) and Chat (CHT) data collected through data donations and live collection involving native speakers of Egyptian Arabic. The corpus contains 5,691 conversations totaling 1,029,248 words across 262,026 messages. Messages were natively written in either Arabic orthography or romanized Arabizi. A total of 1,856 Arabizi conversations (287,022 words) were transliterated from the original romanized Arabizi script into standard Arabic orthography and then reviewed, corrected and normalized by LDC annotators according to "Conventional Orthography for Dialectal Arabic" (CODA).

The BOLT (Broad Operational Language Translation) program developed machine translation and information retrieval for less formal genres, focusing particularly on user-generated content. LDC supported the BOLT program by collecting informal data sources -- discussion forums, text messaging and chat -- in Chinese, Egyptian Arabic and English. The collected data was translated and annotated for various tasks including word alignment, treebanking, propbanking and co-reference.

BOLT Egyptian Arabic SMS/Chat and Transliteration is distributed via web download.
2017 Subscription Members will receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.
*
(3) CHiME2 Grid was developed as part of The 2nd CHiME Speech Separation and Recognition Challenge and contains approximately 120 hours of English speech from a noisy living room environment. The CHiME Challenges focus on distant-microphone automatic speech recognition (ASR) in real-world environments.

CHiME2 Grid reflects the small vocabulary track of the CHiME2 Challenge. The target utterances were taken from the Grid corpus and consist of 34 speakers reading simple 6-word sequences. The Data is divided into training, development and test sets.

CHiME2 Grid is distributed via web download.

2017 Subscription Members will receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

Friday, March 17, 2017

LDC March 2017 Newsletter


New publications:

__________________________________________________________________

New publications:
(1) BOLT Chinese Discussion Forum Parallel Training Data was developed by LDC and consists of 1,876,799 tokens of Chinese discussion forum data collected for the DARPA BOLT program along with their corresponding English translations.

The BOLT (Broad Operational Language Translation) program developed machine translation and information retrieval for less formal genres, focusing particularly on user-generated content. LDC supported the BOLT program by collecting informal data sources -- discussion forums, text messaging and chat -- in Chinese, Egyptian Arabic and English. The collected data was translated and annotated for various tasks including word alignment, treebanking, propbanking and co-reference.

The source data in this release consists of discussion forum threads harvested from the Internet by LDC using a combination of manual and automatic processes. The full source data collection is released as BOLT Chinese Discussion Forums (LDC2016T05). Word-aligned and tagged data is released as BOLT Chinese-English Word Alignment and Tagging - Discussion Forum Training (LDC2016T19).

BOLT Chinese Discussion Forum Parallel Training Data is distributed via web download.

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*

(2) IARPA Babel Swahili Language Pack IARPA-babel202b-v1.0d was developed by Appen for the IARPA (Intelligence Advanced Research Projects Activity) Babel program. It contains approximately 200 hours of Swahili conversational and scripted telephone speech collected from 2012-2014 along with corresponding transcripts.

The Babel program focuses on underserved languages and seeks to develop speech recognition technology that can be rapidly applied to any human language to support keyword search performance over large amounts of recorded speech.

The Swahili speech in this release represents that spoken in the Nairobi dialect region of Kenya. The gender distribution among speakers is approximately equal; speakers' ages range from 16 years to 65 years. Calls were made using different telephones (e.g., mobile, landline) from a variety of environments including the street, a home or office, a public place, and inside a vehicle.

Transcripts are encoded in UTF-8.

IARPA Babel Swahili Language Pack IARPA-babel202b-v1.0d is distributed via web download.

2017 Subscription Members will receive copies of this corpus provided they have submitted a completed copy of the special license agreement. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*

(3) Noisy TIMIT Speech was developed by the Florida Institute of Technology and contains approximately 322 hours of speech from the TIMIT Acoustic-Phonetic Continuous Speech Corpus (LDC93S1) modified with different additive noise levels. Only the audio has been modified; the original arrangement of the TIMIT corpus is still as described by the TIMIT documentation.

The additive noise are white, pink, blue, red, violet and babble noise with levels varying in 5 dB (decibel) steps, ranging from 5 to 50 dB. The color noise types were generated artificially using MATLAB. The babble noise was selected from a random segment of recorded babble speech scaled relative to the power of the original TIMIT audio signal.

Noisy TIMIT Speech is distributed via web download.

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*

(4) GALE English-Chinese Parallel Aligned Treebank -- Training was developed by LDC and contains 196,123 tokens of word aligned English and Chinese parallel text with treebank annotations. This material was used as training data in the DARPA GALE (Global Autonomous Language Exploitation) program.

Parallel aligned treebanks are treebanks annotated with morphological and syntactic structures aligned at the sentence level and the sub-sentence level. Such data sets are useful for natural language processing and related fields, including automatic word alignment system training and evaluation, transfer-rule extraction, word sense disambiguation, translation lexicon extraction and cultural heritage and cross-linguistic studies. With respect to machine translation system development, parallel aligned treebanks may improve system performance with enhanced syntactic parsers, better rules and knowledge about language pairs and reduced word error rate.

The English source data was translated into Chinese. Chinese and English treebank annotations were performed independently. The parallel texts were then word aligned. The material in this release corresponds to portions of the treebanked data in OntoNotes 3.0 (LDC2009T24) and OntoNotes 4.0 (LDC2011T03).

This release consists of English source broadcast programming (CNN, NBC/MSNBC) and web data collected by LDC in 2005 and 2006.

GALE English-Chinese Parallel Aligned Treebank – Training is distributed via web download

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.


Wednesday, February 15, 2017

LDC February 2017 Newsletter

LDC Director Mark Liberman receives the IEEE James L. Flanagan Speech and Audio Processing Award
Only two weeks left to enjoy 2017 membership discounts
Spring 2016 LDC Data Scholarship recipients
New publications:
______________________________________________________________

LDC Director Mark Liberman receives the IEEE James L. Flanagan Speech and Audio Processing Award

LDC Director Mark Liberman is the 2017 recipient of the IEEE James L. Flanagan Speech and Audio Processing Award. Established in 2002, this annual award recognizes an individual for his or her outstanding contribution to the advancement of speech and/or audio processing. Liberman’s pioneering contributions and continued leadership in robust, replicable, and data-driven speech and language science and engineering have fueled the development and advancement of human language technologies including speech and speaker recognition, machine translation, and semantic analysis. As LDC’s founder, Mark has shepherded the Consortium from a small organization to the largest developer of shared language resources, distributing more than 120,000 copies of over 2,000 databases covering 91 different languages to more than 3,600 organizations in over 70 countries. 

Liberman will receive the award at ICASSP 2017 in New Orleans (March 5-9). LDC will be an exhibitor at Booth 43. Please stop by and say hello. We hope to see you there.   

Only two weeks left to enjoy 2017 membership discounts
There is still time to save on 2017 membership fees. Through March 1, all organizations receive a discount on the 2017 membership fee (up to 10%) when they choose to join or renew.  

For more information on membership benefits, visit Join LDC.

Spring 2016 LDC Data Scholarship recipients

Congratulations to the recipients of LDC's Spring 2017 data scholarship:

Umad Ul Hassan and Muhammad Awais Zulfiqar: National University of Sciences and Technology (Pakistan); BS Computer Science. Hassan and Zulfiqar are awarded copies of CSLU: Kids’ Speech Version 1.1 and The CMU Kids Corpus for their research in speech recognition for children with learning difficulties.

For information about the program, visit the Data Scholarship page.

New publications:
(1) First-Year Law Students' Court Memoranda consists of 197 English law student writing samples of legal briefs annotated for certain characteristics along with accompanying survey responses by student writers.

The briefs were created in a law school writing class at two law schools in the US Midwest during the 2011-12 academic year. Students who agreed to participate in this study uploaded their briefs to an online survey instrument and answered questions regarding their age, gender, level of education, most recent writing course and method of learning English. The study's purpose was to apply natural language processing approaches to determine any differences in the briefs' language attributable to the students' self-reported genders.

The samples were imported into the General Architecture for Text Engineering (GATE) and annotated by two human coders who identified large text segments specific to the legal genre in which the students wrote, such as text headings, citations, block quotes and footnotes.

Writing samples are presented as MS Word documents and annotations and survey responses are presented in XML format. The data has been anonymized to remove names and other identifying information about the student participants.

First-Year Law Students' Court Memoranda is distributed via web download.

2017 Subscription Members will receive copies of this corpus provided they have submitted a completed copy of the special license agreement. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*
(2) IARPA Babel Haitian Creole Language Pack IARPA-babel201b-v0.2b was developed by Appen for the IARPA (Intelligence Advanced Research Projects Activity) Babel program. It contains approximately 203 hours of Haitian Creole conversational and scripted telephone speech collected in 2012 and 2013 along with corresponding transcripts.

The Babel program focuses on underserved languages and seeks to develop speech recognition technology that can be rapidly applied to any human language to support keyword search performance over large amounts of recorded speech.

The Haitian Creole speech in this release represents that spoken in the Northern, Western and Southern dialect regions in Haiti. The gender distribution among speakers is approximately equal; speakers' ages range from 16 years to 75 years. Calls were made using different telephones (e.g., mobile, landline) from a variety of environments including the street, a home or office, a public place, and inside a vehicle.

Transcripts are encoded in UTF-8.

IARPA Babel Haitian Creole Language Pack IARPA-babel201b-v0.2b is distributed via web download.

2017 Subscription Members will receive copies of this corpus provided they have submitted a completed copy of the special license agreement. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*
(3) GALE Phase 3 Arabic Broadcast News Speech Part 2 was developed by LDC and is comprised of approximately 128 hours of Arabic broadcast conversation speech collected in 2007 by LDC, MediaNet, Tunis, Tunisia and MTC, Rabat, Morocco during Phase 3 of the DARPA GALE (Global Autonomous Language Exploitation) program.

Corresponding transcripts are released as GALE Phase 3 Arabic Broadcast News Transcripts Part 2 (LDC2017T04).

The recordings in this corpus feature news broadcasts focusing principally on current events from various broadcast programmers including Abu Dhabi TV, Al Alam News Channel, Al Arabiya, Al Iraqiyah, Aljazeera, Al Ordiniyah, Dubai TV, Kuwait TV, Lebanese Broadcast Corporation, Nile TV, Saudi TV and Syria TV.

This release contains 175 audio files presented in FLAC-compressed Waveform Audio File format (.flac), 16000 Hz single-channel 16-bit PCM. Each file was audited by a native Arabic speaker.

GALE Phase 3 Arabic Broadcast News Speech Part 2 is distributed via web download.

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.
*

(4) GALE Phase 3 Arabic Broadcast News Transcripts Part 2 was developed by LDC and contains transcriptions of approximately 128 hours of Arabic broadcast news speech collected in 2007 by LDC, MediaNet, Tunis, Tunisia and MTC, Rabat, Morocco during Phase 3 of the DARPA GALE (Global Autonomous Language Exploitation) program.

Corresponding audio data is released as GALE Phase 3 Arabic Broadcast News Speech Part 2 (LDC2017S02). 

The transcript files are in plain-text, tab-delimited format (TDF) with UTF-8 encoding, and the transcribed data totals 721,846 tokens. The transcripts were created with the LDC tool, XTrans, which supports manual transcription and annotation of audio recordings.

The files in this corpus were transcribed by LDC staff and/or by transcription vendors under contract to LDC. Transcribers followed LDC's quick transcription guidelines (QTR) and quick rich transcription specification (QRTR) both of which are included in the documentation with this release.

GALE Phase 3 Arabic Broadcast News Transcripts Part 2 is distributed via web download.

2017 Subscription Members will automatically receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.



Thursday, January 19, 2017

LDC January 2017 Newsletter

LDC Membership Discounts for MY2017 Still Available

New publications:
___________________________________________________________________
LDC Membership Discounts for MY2017 Still Available
Join LDC now while membership savings are still available. 2016 members receive a 10% discount when renewing before March 1, 2017, or a 5% discount when renewing any time in 2017. Non-consecutive members and new members receive a 5% discount when renewing before March 1, 2017.  Membership remains the most economical way to access LDC releases.  This year’s planned publications include 2010 NIST Speaker Recognition Evaluation data set, Multilanguage Conversational Telephone Speech, Noisy TIMIT, IARPA Babel Language Packs, RATS Keyword Spotting, BOLT parallel and word-aligned data in all languages and more. Browse the Members pages for details on membership options and benefits.

New Corpora

(1) Arabic Speech Recognition Pronunciation Dictionary was developed by the Qatar Computing Research Institute. It contains approximately two million pronunciation entries for 526,000 Modern Standard Arabic words, for an average of 3.84 pronunciations for each grapheme word. The dictionary was developed from news archive resources, including the Arabic news website Aljazeera.net. The selected words were those that occurred more than once in the news collection. 
Arabic Speech Recognition Pronunciation Dictionary is distributed via web download.
2017 Subscription Members will receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.
*
(2) IARPA Babel Vietnamese Language Pack IARPA-babel107b-v0.7 was developed by Appen for the IARPA (Intelligence Advanced Research Projects Activity) Babel program. It contains approximately 201 hours of Vietnamese conversational and scripted telephone speech collected in 2012 along with corresponding transcripts.

The Babel program focuses on underserved languages and seeks to develop speech recognition technology that can be rapidly applied to any human language to support keyword search performance over large amounts of recorded speech.

The Vietnamese speech in this release represents that spoken in the North, North-Central, Central and Southern dialect regions in Vietnam. The gender distribution among speakers is approximately equal; speakers' ages range from 16 years to 64 years. Calls were made using different telephones (e.g., mobile, landline) from a variety of environments including the street, a home or office, a public place, and inside a vehicle.

Transcripts are encoded in UTF-8.

IARPA Babel Vietnamese Language Pack IARPA-babel107b-v0.7 is distributed via web download.

2017 Subscription Members will receive copies of this corpus provided they have submitted a completed copy of the special license agreement. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee..
*
(3) MWE-Aware English Dependency Corpus was developed by the Nara Institute of Science and Technology Computational Linguistics Laboratory and consists of English compound function words annotated in dependency format. The data is derived from the Wall Street Journal portion of OntoNotes Release 5.0 (LDC2013T19).
Compound function words are a type of multiword expression (MWE). MWEs are groups of tokens that can be treated as a single semantic or syntactic unit. Doing so facilitates natural language processing tasks such as constituency and dependency parsing.
MWE-Aware English Dependency Corpus is distributed via web download.
2017 Subscription Members will receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee..
*
(4) GALE Phase 3 and 4 Chinese Web Parallel Text was developed by LDC and contains Chinese source text and corresponding English translations selected from weblog and newsgroup data collected by LDC and translated by LDC or under its direction.
The data includes 88 source-translation document pairs, comprising 67,514 tokens of Chinese source text and its English translation.
GALE Phase 3 and 4 Chinese Web Parallel Text is distributed via web download.

2017 Subscription Members will receive copies of this corpus. 2017 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

Friday, December 16, 2016

LDC December 2016 Newsletter

Renew your LDC membership today
Spring 2017 LDC Data Scholarship Program - deadline approaching
LDC to close for Winter Break
New publications:




_____________________________________________________________________
Renew your LDC membership today
Membership Year 2017 (MY2017) is open for joining and discounts are available for those who keep their membership current and join early in the year. Now through March 1, 2017, current MY2016 members who renew before March 1, will receive a 10% discount off of the membership fee. New or returning organizations will receive a 5% discount through March 1. 

In addition to receiving new publications, current year LDC members also enjoy the benefit of licensing older data at reduced costs from our Catalog of almost 700 holdings; current year for-profit members may use most data for commercial applications.
Plans for MY2017 publications are in progress. Among the expected releases are:
  • 2010 NIST Speaker Recognition Evaluation data set
  • Multilanguage conversational telephone speech: developed to support language identification research in related languages
  • UCLA High Speed Laryngeal Database: audio recordings and high-speed video endoscopic images of the vocal folds while sustaining vowels
  • Noisy TIMIT: TIMIT with added artificial noise
  • CHiME shared task data: noisy read WSJ speech
  • First Year Law Students’ Memoranda: memos to a hypothetical court with annotations
  • IARPA Babel Language Packs: languages include Vietnamese, Haitian Creole, Zulu, Kazakh and Lithuanian
  • BOLT: source, parallel and word-aligned data in all languages
  • RATS Keyword Spotting data set
  • GALE Phases 3 and 4: all tasks and languages    

And don’t forget, MY2016 and MY2015 are still open for joining. MY2015 can be joined through December 31, 2016 and includes data such as RATS Speech Activity Detection and updates to Penn Treebank. MY 2016 will remain open through December 31, 2017 and includes data such as BOLT Chinese Discussion Forums, IARPA Babel Language Packs and Multi-Language Conversational Telephone Speech – Slavic Group. For full descriptions of these data sets, visit our Catalog.  
Visit Join LDC for details on membership, user accounts and payment.

Spring 2017 LDC Data Scholarship Program - deadline approaching
Students can apply for the Spring 2017 Data Scholarship Program now through January 16, 2017, 11:59PM EST. The LDC Data Scholarship program provides undergraduate and graduate students with access to LDC data at no cost.
For more information on application requirements and program rules, please visit LDC Data Scholarships. Students can email their applications to the LDC Data Scholarships program. Decisions will be sent by email from the same address.

LDC to close for Winter Break
LDC will be closed from Monday, December 26, 2016 through Monday, January 2, 2017 in accordance with the University of Pennsylvania Winter Break Policy. Our offices will reopen on Tuesday, January 3, 2017. Requests received for membership renewals and corpora during the Winter Break will not be processed until the week of January 3.

New Corpora

(1) Bamanankan Lexicon was developed by LDC and contains 5,978 entries of the Bamanankan language presented as a Bamanankan-English lexicon and a Bamanankan-French lexicon. It is the third publication in an LDC project to build an electronic dictionary of three Mandekan languages: Mawukakan, Maninkakan and Bamanankan. These are Eastern Manding languages in the Mande Group of the Niger-Congo language family. LDC released a Mawukakan Lexicon (LDC2005L01) in 2005 and a Maninkakan Lexicon (LDC2013L01) in 2013.

This lexicon is presented using a Latin-based transcription system because the Latin alphabet is familiar to the majority of Mandekan language speakers and it is expected to facilitate the work of researchers interested in this resource.

Bamanankan Lexicon is distributed via web download.

2016 Subscription Members will receive copies of this corpus. 2016 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*

(2) IARPA Babel Tagalog Language Pack IARPA-babel106-v0.2g was developed by Appen for the IARPA (Intelligence Advanced Research Projects Activity) Babel program. It contains approximately 213 hours of Tagalog conversational and scripted telephone speech collected in 2012 along with corresponding transcripts.

The Babel program focuses on underserved languages and seeks to develop speech recognition technology that can be rapidly applied to any human language to support keyword search performance over large amounts of recorded speech.

The Tagalog speech in this release represents that spoken in the North, Central and South dialect regions in the Philippines. The gender distribution among speakers is approximately equal; speakers' ages range from 16 years to 65 years. Calls were made using different telephones (e.g., mobile, landline) from a variety of environments including the street, a home or office, a public place, and inside a vehicle.

Transcripts are encoded in UTF-8.

IARPA Babel Tagalog Language Pack IARPA-babel106-v0.2g is distributed via web download.

2016 Subscription Members will receive copies of this corpus provided they have submitted a completed copy of the special license agreement. 2016 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.
*

(3) TAC KBP Spanish Cross-lingual Entity Linking - Comprehensive Training and Evaluation Data 2012-2014 was developed by LDC and contains training and evaluation data produced in support of the TAC KBP Spanish Cross-lingual Entity Linking tasks in 2012, 2013 and 2014. It includes queries and gold standard entity type information, Knowledge Base links, and equivalence class clusters for NIL entities along with the source documents for the queries, specifically, English and Spanish newswire, discussion forum and web data. The corresponding knowledge base is available as TAC KBP Reference Knowledge Base (LDC2014T16).

More information about the TAC KBP Entity Linking task and other TAC KBP evaluations can be found on the NIST TAC website.

TAC KBP Spanish Cross-lingual Entity Linking - Comprehensive Training and Evaluation Data 2012-2014 is distributed via web download.

2016 Subscription Members will receive copies of this corpus. 2016 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

*

(4) GALE Phase 4 Arabic Newswire Parallel Sentences was developed by LDC and contains Modern Standard Arabic source text and corresponding English translations selected from newswire data collected by LDC in 2008 and translated by LDC or under its direction.

This release includes 393 source-translation document pairs drawn from six distinct newswire sources, comprising 62,669 tokens of Arabic source text and its English translation. Source data and translations are distributed in TDF format. All data is encoded in UTF-8.

GALE Phase 4 Arabic Newswire Parallel sentences is distributed via web download.

2016 Subscription Members will receive copies of this corpus. 2016 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.


Tuesday, November 15, 2016

LDC November 2016 Newsletter



In this newsletter:

Join LDC for Membership Year 2017
Commercial use and LDC data
Spring 2017 Data Scholarship Program
LDC closed November 24-25 for US Thanksgiving Holiday

New publications:






Join LDC for Membership Year 2017

Organizations engaged in language-related research, education and technology development are invited to join LDC for Membership Year (MY) 2017. Consortium members enjoy unparalleled access and continuing rights to new data releases and to an archive of close to 700 holdings.

Membership fees have not increased for 2017. In addition, discounts are available for organizations who keep their membership current and for those who join before March 1, 2017.

           • MY 2016 members receive a 10% discount if they renew their membership before March 1, 2017. After March 1, MY2016 members receive a 5% discount if they renew their membership any time in 2017.
           • New members and returning former members receive a 5% discount off the membership fee if they join/renew before March 1, 2017.

Plans for MY2017 publications are in progress. Among the expected releases are:

2010 NIST Speaker Recognition Evaluation data set
Multilanguage conversational telephone speech: developed to support language identification research in related languages
UCLA High Speed Laryngeal Database: audio recordings and high-speed videoendoscopic images of the vocal folds while sustaining vowels
Noisy TIMIT: TIMIT with added artificial noise
CHiME shared task data: noisy read WSJ speech
First Year Law Students’ Memoranda: memos to a hypothetical court with annotations
IARPA Babel Language Packs: languages include Vietnamese, Haitian Creole, Zulu, Kazakh and Lithuanian
BOLT: source, parallel and word-aligned data in all languages
RATS Keyword Spotting data set
GALE Phases 3 and 4: all tasks and languages   

Visit Join LDC for details on membership, user accounts and payment.

Commercial use and LDC data

For-profit organizations are reminded that an LDC membership is a pre-requisite for obtaining a commercial license to almost all LDC databases. Non-member organizations, including non-member for-profit organizations, cannot use LDC data to develop or test products for commercialization, nor can they use LDC data in any commercial product or for any commercial purpose. LDC data users should consult corpus-specific license agreements for limitations on the use of certain corpora. Visit the Licensing page for further information. 

Spring 2017 Data Scholarship Program

Applications are now being accepted through January 15, 2017 for the Spring 2017 LDC Data Scholarship program which provides university students with no-cost access to LDC data. Consult the LDC Data Scholarship page for further information about program rules and submission requirements.

LDC closed November 24-25 for US Thanksgiving Holiday

LDC will be closed on Thursday, November 24, 2016 and Friday, November 25, 2016 in observance of the US Thanksgiving Holiday. The office will reopen on Monday, November 28, 2016.

New Corpora

(1) JANA: A Human-Human Dialogues Corpus for Egyptian Dialect was developed by researchers at Cairo University. This is a special release in addition to the LDC scheduled corpora for membership year 2016, available under separate terms.

This corpus consists of 82 transcribed dialogues from call center inquiries annotated for dialogue acts. Data was collected from call centers for banks, airlines and mobile network providers in the form of spontaneous spoken telephone dialogues (52) and instant messaging dialogues (30) amounting to over 20,000 words.

Not-for-profit organizations may license this data set for a fee under the LDC Not-for-Profit Membership Agreement or under the LDC User Agreement for Non-Members for use in linguistic research, education and non-commercial technology development. For-profit organizations may license this data for a fee under a commercial license.

(2) Multi-Language Conversational Telephone Speech 2011 – Slavic Group was developed by LDC and is comprised of approximately 60 hours of telephone speech in Polish, Russian and Ukrainian. The data was collected to support research and technology evaluation in automatic language identification, specifically language pair discrimination for closely related languages/dialects.

Call were made using LDC’s telephone collection infrastructure. Human auditors labeled calls for gender, dialect type and noise.  Audio data is presented in FLAC-compressed MS-WAV (RIFF) file format. Each uncompressed file is two channels, recorded at 8000 samples/second with samples stored as 16-bit signed integers.

Multi-Language Conversational Telephone Speech 2011 – Slavic Group is distributed via web download.

2016 Subscription Members will receive copies of this corpus. 2016 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

(3) IARPA Babel Georgian Language Pack IARPA-babel404b-v1.0a was developed by Appen for the IARPA (Intelligence Advanced Research Projects Activity) Babel program. It contains approximately 190 hours of Georgian conversational and scripted telephone speech collected in 2014-2015 along with corresponding transcripts.

The Babel program focuses on underserved languages and seeks to develop speech recognition technology that can be rapidly applied to any human language to support keyword search performance over large amounts of recorded speech.

The Georgian speech in this release represents that spoken in the Eastern and Western dialect regions in Georgia. The gender distribution among speakers is approximately equal; speakers' ages range from 16 years to 73 years. Calls were made using different telephones (e.g., mobile, landline) from a variety of environments including the street, a home or office, a public place, and inside a vehicle.

Transcripts are encoded in UTF-8.

IARPA Babel Georgian Language Pack IARPA-babel404b-v1.0a is distributed via web download.

2016 Subscription Members will receive copies of this corpus provided they have submitted a completed copy of the special license agreement. 2016 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.

(4) GALE Phase 3 and 4 Chinese Newswire Parallel Text was developed by LDC and contains Chinese source text and corresponding English translations selected from newswire data collected by LDC in 2007-2008 and translated by LDC or under its direction.

This release includes 367 source-translation document pairs drawn from five distinct newswire sources, comprising 210,048 tokens of Chinese source text and its English translation. Source data and translations are distributed in TDF format. All data is encoded in UTF-8.

GALE Phase 3 and 4 Chinese Newswire Parallel Text is distributed via web download.

2016 Subscription Members will receive copies of this corpus. 2016 Standard Members may request a copy as part of their 16 free membership corpora. Non-members may license this data for a fee.