.png)
Workshop Format
SilkRoadNLP 2026 will be a half-day workshop, featuring approximately five long talks, one thematic panel, and an extended poster session to encourage interaction and feedback across communities.
We are also developing a dedicated GitHub repository for open-source datasets, benchmarks, annotation guidelines, and code contributions. This shared resource hub will help ensure that progress made through the workshop is both transparent and reusable across future research efforts.
About SilkRoadNLP
SilkRoadNLP is the first ACL-affiliated workshop dedicated to advancing Natural Language Processing (NLP) and Large Language Models (LLMs) for the Iranian linguistic family — a diverse network of languages spoken across Iran, Afghanistan, Central and South Asia, and the Caucasus.
Our mission is to bring together computational linguists, AI researchers, linguists, and community experts to foster open collaboration, share resources, and promote culturally aware, inclusive approaches to multilingual AI.
As the AI landscape is rapidly transformed by large-scale language models, SilkRoadNLP underscores a simple truth:
without intentional inclusion, linguistic diversity risks digital extinction.
Why Now
Languages of the Iranian linguistic family — including Persian (Farsi), Dari, Tajik, Kurdish (Sorani and Kurmanji), Balochi, Gilaki, Mazandarani, Luri, Pashto, and Ossetian — represent thousands of years of shared history, literature, and oral tradition. Yet most remain low-resource in NLP, lacking open datasets, benchmarks, and digital infrastructure.
The rise of LLMs amplifies both opportunity and risk. While these models promise new tools for cross-lingual communication and education, they also reinforce the data imbalances that can marginalize small or under-documented languages.
SilkRoadNLP responds to this moment by providing a collaborative, interdisciplinary venue for computational work that not only advances AI research but also safeguards linguistic and cultural heritage.
Our Goals
-
Advance research in NLP and LLMs for low-resource Iranian languages.
-
Develop and share open resources, benchmarks, and evaluation tools.
-
Foster interdisciplinary collaboration among linguists, technologists, and community organizations.
-
Promote responsible and culturally grounded AI that represents, rather than replaces, local voices.
-
Build continuity for future shared tasks, datasets, and regional collaboration networks.
Hybrid Participation
SilkRoadNLP 2026 will adopt a collaborative hybrid format to make participation accessible for researchers and practitioners across regions. In coordination with the EACL 2026 organizers, we aim to combine the strengths of in-person interaction with meaningful online engagement.
The hybrid setup will include live talks and panels streamed from Rabat, alongside virtual poster sessions and pre-recorded presentations that can be viewed asynchronously. Remote participants will be able to take part in interactive Q&A sessions, join networking spaces, and share their own work digitally.
Our goal is to ensure that every participant, whether joining from the conference venue or from across the globe, can contribute, collaborate, and be represented. Final technical details will be shared once the hybrid configuration is confirmed with EACL.
A Growing Global Movement
SilkRoadNLP builds on the momentum of AfricaNLP, AmericasNLP, and SIGTYP, extending this global dialogue to the Iranian linguistic region. Together, these initiatives represent a growing commitment within the ACL community to ensure that every language, regardless of resource level, has a voice in the era of artificial intelligence.
The rise of LLMs amplifies both opportunity and risk. While these models promise new tools for cross-lingual communication and education, they also reinforce the data imbalances that can marginalize small or under-documented languages.