# AFNLP

The Asian Federation of Natural Language Processing (AFNLP) is a regional academic body coordinating NLP research and conferences across Asia, founded in 2000 to foster collaboration among national societies.

The Asian Federation of Natural Language Processing (AFNLP) is a regional umbrella organization that coordinates research, education, and professional activities in natural language processing (NLP) across Asia. Established in 2000, it brings together national and regional NLP societies, aiming to advance the scientific study of human language computation and its applications in machine translation, speech recognition, and text analysis. AFNLP is best known for sponsoring the biennial International Joint Conference on Natural Language Processing (IJCNLP) and supporting the Asian Language Resources workshop series, which have become key venues for the Asian NLP community.

As a federation, AFNLP does not conduct direct research but facilitates exchange among member societies, organizes joint conferences, and promotes open standards and shared resources for underserved Asian languages. Its membership has grown to include societies from China, Japan, Korea, India, Singapore, Thailand, and other countries, reflecting the region's diversity and rapid growth in computational linguistics.

## History and founding

AFNLP was founded in 2000 in response to the growing need for a unified platform for NLP researchers in Asia, who previously relied on ad hoc gatherings and links to Western conferences. The founding initiative came from senior scholars in China, Japan, and Korea, who saw the value of pooling resources and expertise. The first formal statute was adopted in 2001, establishing a secretariat and an executive committee with rotating chairmanship among member countries.

Early milestones included the launch of the first IJCNLP in Sanya, China, in 2003, which attracted over 200 participants. The federation also began publishing the journal "Journal of Natural Language Processing" (in cooperation with member societies) and later the "Asian Journal of Natural Language Processing" as an open-access outlet. Membership expanded steadily through the 2000s, with Southeast Asian and South Asian societies joining by 2008.

## Objectives and governance

The primary objective of AFNLP is to promote the development of NLP as a scientific discipline in Asia, emphasizing both theoretical rigor and practical applications for regional languages. Its charter lists four goals: supporting high-quality research, facilitating the sharing of linguistic data and tools, organizing educational events and summer schools, and fostering international collaboration with non-Asian bodies.

Governance is structured around a general assembly of member society representatives, an elected executive committee, and a secretariat that rotates among host institutions. The president serves a two-year term, assisted by vice-presidents from different subregions. Decisions are made by consensus where possible, with voting on budget and major policy changes. Day-to-day operations are managed by a permanent secretariat office, currently located in Japan, with support from the members' annual dues.

## Major conferences and activities

The flagship activity of AFNLP is the International Joint Conference on Natural Language Processing (IJCNLP), held every two years. IJCNLP features invited talks, paper sessions, tutorials, and shared tasks that attract researchers globally. Recent editions have been held in Taipei (2017), Hong Kong (2019), and online in 2021 due to the COVID-19 pandemic, with the 2023 edition in Bali, Indonesia, drawing over 600 participants.

In addition to IJCNLP, AFNLP sponsors the Asian Language Resources Workshop, which focuses on corpus building, annotation standards, and tools for low-resource Asian languages such as Thai, Vietnamese, and Mongolian. The federation also co-sponsors region-specific workshops like the Workshop on Asian Translation (WAT), held annually, which evaluates machine translation systems for languages including Japanese, Chinese, and Korean.

Educational activities include biennial summer schools on computational linguistics, typically held in collaboration with local universities. For example, the 2019 summer school in Seoul covered deep learning and transformer-based methods, reflecting the field's shift toward [deep learning](https://www.wikiprompt.org/wiki/deep-learning). These events also serve as recruitment pipelines for graduate students and early-career researchers.

## Membership and regional impact

AFNLP's membership includes national academic societies as well as research groups and industry affiliates. As of 2024, the federation counts 12 full member societies and several associate members. Notable members include the Chinese Information Processing Society of China, the Association for Natural Language Processing of Japan, the Korea Association of Computational Linguistics, and the Indian Association for Computational Linguistics. Industry partners include major tech firms like [Samsung](https://www.wikiprompt.org/wiki/samsung-electronics) and [Fujitsu](https://www.wikiprompt.org/wiki/fujitsu), which contribute to shared tasks and sponsor workshops.

The federation's impact is most visible in the standardization of evaluation practices for Asian languages. Its shared tasks on named entity recognition and sentiment analysis have produced benchmark datasets used by hundreds of teams, including those from [Alibaba Damo Academy](https://www.wikiprompt.org/wiki/alibaba-damiao-academy) and academic labs. AFNLP also maintains a directory of language resources, listing over 300 corpora and dictionaries, which has reduced duplication and accelerated research in low-resource settings.

## Relationship with global NLP community

While AFNLP focuses on Asia, it maintains formal ties with global organizations. It has a standing cooperation agreement with the Association for Computational Linguistics (ACL), allowing joint sponsorship of workshops and cross-listing of conferences. Many AFNLP members also serve on ACL committees, and IJCNLP submissions often overlap with those to ACL venues, reflecting the close scientific community.

The federation also coordinates with the International Committee on Computational Linguistics (ICCL) for the COLING conference, ensuring that Asian venues are not scheduled in conflict. In recent years, AFNLP has pushed for greater inclusion of under-represented languages in mainstream NLP benchmarks, advocating for datasets beyond English and Chinese. This aligns with broader trends in [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) research, where multilingual models are becoming central to [large language models](https://www.wikiprompt.org/wiki/large-language-model) developed by [OpenAI](https://www.wikiprompt.org/wiki/openai) and [Google DeepMind](https://www.wikiprompt.org/wiki/google-deepmind).

## Challenges and future directions

One persistent challenge is the diversity of Asian languages, which vary in writing systems, morphology, and available digital resources. AFNLP has responded by forming working groups on script-specific issues, such as Thai tokenization and Hindi word segmentation. Another challenge is the brain drain of top researchers to Western institutions and industry labs, though the federation's networking reduces its impact.

Looking forward, AFNLP is focusing on AI-driven applications and ethical considerations. It has initiated a task force on responsible NLP, examining biases in training data and the implications of automated translation for cultural preservation. The 2025 IJCNLP, planned for Seoul, will include a special track on multilingual and multimodal learning, aiming to connect linguistic research with progress in [neural networks](https://www.wikiprompt.org/wiki/neural-network) and [transformers](https://www.wikiprompt.org/wiki/transformer). The federation also plans to expand its digital infrastructure, offering a virtual repository for datasets and code to support replication and comparability.

## Significance in the field

AFNLP occupies a unique niche as a regional catalyst rather than a research institute. Its conferences have launched influential careers, and its shared tasks have set standards for Asian languages that global benchmarks often overlook. For researchers in Asia, it provides a home-like community that balances rigorous science with practical support for under-served languages. As computational linguistics becomes more globalized, AFNLP's model of federation and cooperation offers a template for other regions seeking to amplify their voice in the field.

---
Source: https://www.wikiprompt.org/wiki/afnlp
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-09-14T04:15:15.877067+00:00