< Go Back

AMTA 2023 Virtual | 2nd Call for Proposals

by | June 13, 2023
read

Association for Machine Translations in the Americas

 

Second Call for Proposals for 

Generative AI and the Future of Machine Translation

A virtual one-day event under the auspices of AMTA

Wednesday, 8 November 2023

 

The Board of Directors of AMTA (the Association for Machine Translation in the Americas) is pleased to announce the first call for proposals for Generative AI and the Future of Machine Translation, a one-day virtual event to be held on Wednesday, 8 November 2023. 

Generative AI and the Future of Machine Translation will focus on advances in cross-lingual technology and processes that leverage Large Language Models. Moving beyond the explosion of hype that began with the general availability of ChatGPT in November 2022, this event is intended as a forum for a thoughtful exchange of ideas informed by a year of successes and failures in applying Large Language Models to the challenges of cross-lingual communication and data processing.

Like all AMTA events, Generative AI and the Future of Machine Translation will bring together researchers, practitioners, and providers of MT and related cross-lingual technology from academia, industry, and government and will include a keynote address by Christian Federmann, Principal Researcher Manager, Microsoft, expert panel discussions, individual presentations, succinct tutorials for both beginners and more experienced practitioners, and demonstrations from technology providers.

The organizing committee of Generative AI and the Future of Machine Translation is seeking proposals for presentations and tutorials on all topics related to research, development, application, and evaluation of cross-lingual technology(ies). Our goal is to have a program that appeals to the various constituents of the MT community (researchers, developers, users, and language professionals). Therefore we welcome not only proposals on technical research and development topics but also on, for instance, the collection and curation of training data, best practices in training models, human/computer interaction among translators, interpreters, and other users of generative output in cross-lingual use cases, and the evolution of translation automation in the commercial translation production pipeline. 

Developers, practitioners, and analysts from industry, the language industry, and government are encouraged to submit proposals that cover leading-edge R&D and practical applications of LLMs as they relate to Machine Translation and the creation of processing of cross-lingual content.

Members of the research community are encouraged to submit proposals to this non-archival event that treat recent research directions, trends and concerns about the use of generative AI from the perspective of researchers, or suggestions for research projects, including for industry players.

We seek submissions for:

  • 20-minute talks (15 minute presentations, plus 5 minutes for questions), 
  • 40-minute tutorials (practical description or exercises concerning the use of Large Language Models, generative AI, Machine Translation, and/or related tools, processes and technologies for cross-lingual tasks to support real-world goals

Topics of interest might include, but are not limited to, the following:

  • Open-source Large Language Models for Translation, Transcreation, and other cross-lingual use cases: development, deployment, and adaptation to specific use-cases.
  • Adaptation and customization of large language models for cross-lingual use cases: case studies on applying few-shot learning, prompt-tuning and fine-tuning, comparison of methodologies used to adapt and customize foundation models for specific tasks with respect to business, technical and linguistic requirements.
  • Combining narrow and large models to improve performance of specific tasks, such as translation, OCR, ASR and others.
  • Augmenting MT systems with generative AI: including approaches to leveraging TM and end-user feedback, classification, context awareness, content moderation, sentiment analysis.
  • Training Data: data sources, processing, cleaning of data, terminology, data augmentation, multimodal data, etc.
  • Output quality and confidence scoring for cross-lingual tasks: tools, methods and metrics, such as human evaluations, automatic scoring, reference-based and source-only scoring.
  • Challenges to adopting LLMs in cross-lingual use cases: Responsible deployment and regulatory considerations of generative AI (such as technical, ethical, social, legal, and environmental challenges). 
  • Generative AI use for professional translation: Approaches, success and failure stories, applicability to content-types, fair pricing models, and working with customers, buyers or providers.
  • Business Cases: making the business case for or against adopting generative AI to drive business requirements.
  • Future research directions: open problems that researchers are/should be considering in this area.

Important dates  

  • Submission deadline: Extended to 15 August 2023
  • Notification of acceptance: 15 September 2023

Submission Instructions

Proposals should be submitted in PDF format by 15 August 2023, to submissions@amtaweb.org and include: 

  • Title of proposed session
  • Type of proposed session (20-minute presentation or 40-minute tutorial) 
  • A 250-500 word description of the proposed session
  • Name and e-mail address of the person who is submitting the proposal
  • Name(s) and e-mail address(es) of proposed speakers
  • A short (<100 words) biographical introduction to the proposed presenter(s)
  • Any special technical requirements you may have

Publication and recording

This is a non-archival event. Print proceedings of the event will not be produced. However, video recordings will be made of each presentation and will be made available to association members on the AMTA website. Submission of a proposal will be understood by the Organizing Committee as your tacit permission for AMTA to create an audio and video recording of your presentation and make it available to conference attendees and, through archiving, to members of AMTA, EAMT, and AAMT.

Background

AMTA, the Association for Machine Translation in the Americas, is well known for its biennial conferences, held since 1994, and for its sponsorship and organization of the MT Summit in 1991, 1997, 2003, 2009, 2015, and 2021. Although the next full AMTA conference will not be held until 2024, the board of directors unanimously agreed that the pace of development in and rapid adoption of large language models and generative AI warranted organizing and hosting an out-of-cycle virtual event.

 

We look forward to receiving your proposal!

 

AMTA Board of Directors

 

Contacts:  Jay Marciano, President (president@amtaweb.org)

                   Alex Yanishevsky, Vice President (vicepresident@amtaweb.org)

                   Janice Campbell, Secretary (secretary@amtaweb.org)