top of page
DALL-E-ALL4Health-Hindi.png

The First Workshop on Applying LLMs in LMICs for Healthcare Solutions

@ ICHI
3 June, 2024
Orlando, Florida, USA

Agenda

1:00pm-1:05 pm: Workshop chairs. Welcome remarks


1:05pm-1:25 pm: David Restrepo, Luis Filipe Nakayama, Robyn Gayle Dychiao, Chenwei Wu, Liam G. McCoy, Jose Carlo Artiaga, Marisa Cobanaj, João Matos, Jack Gallifant, Danielle S. Bitterman, Vincenz Constantine Ferrer, Yindalon Aphinyanaphongs and Leo Anthony Celi. Seeing Beyond Borders: Evaluating LLMs in Multilingual Ophthalmological Question Answering


1:25pm-1:45 pm: Shuyue Stella Li, Vidhisha Balachandran, Shangbin Feng, Emma Pierson, Pang Wei Koh and Yulia Tsvetkov. Beyond the Stethoscope: Operationalizing Interactive Clinical Reasoning in Large Language Models via Proactive Information Seeking


1:45pm-2:20 pm: Mark Dredze. Keynote: "Large Language Models in Medicine: Opportunities and Challenges"


2:20pm-2:50 pm: Coffee Break (20 minutes)


2:50pm-3:10 pm: Walelign Sewunetie, Assefa Beza, Tesfamariam Abuhay, Hailemariam Endalamaw, Wasyihun Admass, Hayat Hassen, Hana Hailemariam, Nathnael Bekele, Mahlet Mammo, Lydia Debebe, Mahlet Berta, Tsion Haile, Surafel Luleseged, Nurilign Moges, Seid Yimam and Laszlo Kovacs. Large Language Models for Sexual, Reproductive, and Maternal Health Rights


3:10pm-3:30 pm: Amelia Taylor and Kai Lab. Self-Directed Learning for Community Health Workers in Malawi Through Generative AI


3:30pm-3:50 pm: Paulina Boadiwaa Mensah, Nana Serwaa Quao, Sesinam Dagadu, James Kwabena Mensah, Emmanuel Kwabena Asiedu, Noreen Irena Ghartey, Kwame Konadu Akowuah, Kwasi Antwi Donkor, Jude Domfeh Darkwah, Kwabena Agyei Asiedu Asante, Seth Makafui Passah, Bismark Yeboah Dankwa, Nene Christine Osa-Afiana and Esther Poku Dwumah. All You Need Is Context: Clinician Evaluations of various iterations of a Large Language Model-Based First Aid Decision Support Tool in Ghana


3:50pm-4:10 pm: Grady McPeak, Anja Sautmann, George Ohia, Adham Hallal, Eduardo Arancón, Ravi Nirmal and Robert Pless. An LLM's medical testing recommendations in a Nigerian clinic: potential and limits of prompt engineering for clinical decision support


4:10pm-4:30 pm: Delta-Marie Lewis, Brian DeRenzi, Amos Misomali, Themba Nyirenda, Everlisto Phiri, Lyness Chifisi, Charles Makwenda and Neal Lesh. Human review for post-training improvement of low-resource language performance in large language models


4:30pm-4:40 pm: Workshop Chairs. Closing remarks
 

Introduction

ALL 4 Health 2024 – The First Workshop on Applying LLMs in LMICs for Healthcare Solutions will be held at the University of Florida on June 3rd in conjunction with the IEEE International Conference on Healthcare Informatics (ICHI 2024). There has been substantial and growing interest and funding from the development sector in applying Large Language Model (LLM) technologies in Low- and Middle-Income Countries (LMICs) to address healthcare and other social good challenges.[1] Simultaneously, there have been acknowledgements from the software industry and from NLP researchers that state of the art LLMs are heavily influenced by Western / developed world data and have significant capability gaps between high- and low-resource languages.[2,3,4] Additional research and collaboration is required to bridge this gap.

​

The goal of this workshop is to bring together researchers and practitioners from diverse disciplinary backgrounds to discuss challenges and opportunities for applying LLMs for health applications in low-resource settings, and to share findings on gaps, pitfalls, best practices, and opportunities for impact.

​

References:

  1. R. Shrivastava. “Gates Foundation Funds Nearly 50 Generative AI Projects In Low And Middle Income Countries.” Forbes, 10 August 2023, https://www.forbes.com/sites/rashishrivastava/2023/08/10/gates-foundation-funds-nearly-50-generative-ai-projects-in-low-and-middle-income-countries/

  2. Viet Dac Lai, et al. "Chatgpt beyond english: Towards a comprehensive evaluation of large language models in multilingual learning." arXiv preprint arXiv:2304.05613 (2023).

  3. J. Dodge, et al. "Documenting large webtext corpora: A case study on the colossal clean crawled corpus." arXiv preprint arXiv:2104.08758 (2021).

  4. N.R. Robertson, et al. "ChatGPT MT: Competitive for High- (but not Low-) Resource Languages." arXiv preprint arxiv:2309.07423 (2023).

Call for Submissions

We invite novel approaches, works in progress, comparative analyses of tools, and advancing state-of-the-art work relevant to applying LLMs for health applications in low-resource languages and settings. Specific topics of interest include, but are not limited to: 

  • Evaluations of LLMs in contexts with substantial code-switching

  • Comparisons of LLM accuracy/suitability between high- and low-resource languages

  • Approaches to localizing the health information processing of LLMs in the context of the laws, culture, service availability, and public health realities in specific LMICs

  • Data sources for training or tuning LLMs for use on low-resource languages or in LMIC contexts

  • Studies demonstrating the health or health knowledge impact of LLM applications in low-resource language and/or LMIC contexts

  • Equity- and Diversity-based evaluations of LLM performance on health domain tasks

  • Evidence-based position papers on best practices 

We will accept full papers (4-6 pages, including references) and abstracts (2 pages, including references).

Best Paper Award

Full papers will be eligible for a Best Paper Award with a $300 (USD) prize sponsored by MSD for Mothers.

​

Instructions for Authors

Authors are invited to submit their work through EasyChair (https://easychair.org/conferences/?conf=ieeeichi2024) -- please select the "All 4 Health" track. Submissions should adhere to the IEEE Proceedings format. Accepted papers will be published in ICHI  conference proceedings on IEEE Xplore and will have an oral presentation slot in the workshop.

 

ALL4Health uses a double-blind two-layer peer reviewing process. Reviewers will be drawn from academia, industry, and the development sector. Before a submission is sent to reviewers, the program chairs will perform an assessment to determine the best fit for the submission. The final decision will be made by the program committee based on at least two reviews.

​

We ask that at least one author of each submission participate in the review process by reviewing two other submissions.

bottom of page