Overview

LLM 2023 is the first edition of the Symposium on Advances and Open Problems in Large Language Models (LLM@IJCAI'23). Pre-Trained Large Language Models (LLMs), such as ChatGPT and GPT-4, have revolutionized the field of AI with their remarkable capabilities in natural language understanding and generation. LLMs are widely used in various applications such as voice assistants, recommender systems, content-generation models like ChatGPT and text-to-image models like Dall-E. However, these powerful models also pose significant challenges for their safe and ethical deployment. How can we ensure that LLMs are fair, safe, privacy-preserving, explainable, and controllable?

This symposium aims to provide a platform for academic researchers and industry practitioners to discuss the latest advances and open problems in LLMs. The symposium will cover two main themes: a) to present recent progress of foundational LLMs and their applications in different domains; b) to address open issues and challenges for building trustworthy LLMs. We hope that this symposium will foster interdisciplinary collaboration and exchange of ideas among the participants, and contribute to the development and use of LLMs that benefit humanity.

Schedule

Date: Aug 21, 2023
Room: Kokand 6302+6303
The parallel LLM Session on Medical Large Models organized by Profs. Hongbin Liu, Qifeng Liu, and Zhen Chen is located at Room: Kokand 6307. Check it at https://www.cair-cas.org.hk/article/ijcai-medical-large-models.

Keynotes

Accepted Papers

    Call for Papers

    LLM@IJCAI'23 will be a non-archival symposium; we are excited to host this edition in person on IJCAI 2023 at Macao, China. At least one author of each accepted paper is expected to represent it offline at the symposium.

    We welcome submissions on recent advances and applications of large language models (LLMs) together with an emphasis on enhancing trust in the use of LLMs. The symposium will focus on two main themes: a) to report recent advances of foundation LLMs and their applications in various domains; b) to address open issues and build trustworthiness in LLMs. Topics including (but not limit to):

    Techniques:
    • Advanced model architectures for LLMs, e.g., Transformer architectures and attention mechanisms.
    • Advanced algorithms for improving performance, cost, robustness, and complexity of LLMs.
    • Model transfer and compression techniques for LLMs.
    • Federated Learning for LLMs.
    • Prompt Engineering for LLMs.
    Applications:
    • Innovative applications of LLMs in various domains, e.g., psychotherapy, elderly care, etc.
    • Educational technologies based on LLMs such as chat-bots, content generation, feedback systems, etc.
    • Natural language understanding and generation tasks using LLMs, e.g., storytelling, marketing copywriting, etc.
    • LLMs for health care, protein synthesis, etc.
    Challenges
    • Ethics, social economics, and trustworthiness of LLMs.
    • Data labeling and quality issues for training LLMs.
    • Privacy and security risks of models and data used by LLMs.
    • Potential bias and unfairness in the output of LLMs
    • Human oversight and intervention mechanisms for controlling LLMs.
    • Hallucination detection and alleviation for LLMs.
    • Emergent behavior in LLMs

    Important Dates

    All deadlines are in Anywhere on Earth (AoE) time zone.

    • Submission deadline: June 10, 2023
    • Acceptance notification: June 24, 2023
    • Camera-ready version due: July 10, 2023
    • Symposium: Aug. 21, 2023
    • Venue: Sheraton Grand Macao Hotel, Macau

    Submission

    Each submission can be up to 7 pages of contents plus up to 2 additional pages of references and acknowledgements. The submitted papers must be written in English and in PDF format according to the IJCAI'23 template. All submitted papers will be under a single-blinded peer review for their novelty, technical quality and impact. The submissions can contain author details. Submission will be accepted via the Easychair submission website: https://easychair.org/conferences/?conf=llmijcai23.

    Based on the requirement from IJCAI'23, at least one author of each accepted paper must travel to the IJCAI venue in person. In addition, multiple submissions of the same paper to more than one IJCAI symposium/workshop are forbidden.

    For inquiries, please email to llm-ijcai23@easychair.org

    Organizers

    General Co-Chairs

    • Lei Chen, Hong Kong University of Science and Technology (Guangzhou)
    • Jie Tang, Tsinghua University
    • Qiang Yang, Hong Kong University of Science and Technology & WeBank AI

    Program Co-Chairs