Welcome to The Workshop on Sustainable AI for the Future Web at The Web Conference 2025!
Call for Papers
Web applications such as social media, e-commerce platforms, and search engines generate vast and diverse datasets that fuel AI advancements, enabling models to learn from real-world, dynamic information. Meanwhile, AI models are increasingly deployed within these web applications. Tools like ChatGPT, DALL·E, and sophisticated recommendation systems have seamlessly integrated into web-based services, delivering highly personalized user experiences, optimizing decision-making processes, and greatly enhancing productivity in areas like customer support, content creation, and data analysis. However, as AI and the web become more deeply intertwined, they also introduce significant challenges in building sustainable AI for the web.
This workshop will explore essential topics at the intersection of AI, sustainability, and the web, focusing on creating AI systems that are robust, energy-efficient, and ethically responsible within web applications. As AI models become increasingly integrated into platforms like social media, e-commerce, and search engines, their reliance on vast amounts of web data raises concerns around data quality, fairness, and computational resource demands. Key topics of discussion include, but are not limited to:
-
Sustainable AI Model Development: Methods to reduce the carbon footprint and computational demands of training and deploying AI models on the web.
-
Data Integrity and Filtering Mechanism: Strategies to handle the proliferation of low-quality and AI-generated content that risks polluting the web and undermining data quality.
-
Security, Robustness, and Fairness: Techniques to improve the resilience of AI systems trained on noisy, diverse web data, ensuring they operate securely and equitably across varied user demographics.
Important Dates
| Submission Open | December 5, 2024 |
| Submission Deadline | January 1, 2025 (AoE) |
| WWW-25 Fast-Track Submission Deadline | January 26, 2025 (AoE) |
| Decision Notification | January 27, 2025 (AoE) |
| Camera-Ready Deadline | Februray 7, 2025 (AoE) |
| Workshop Date | April 28 AM, 2025 |
Accepted Papers
-
GPT4Image: Large Pre-trained Models Help Vision Models Learn Better on Perception Task
Ning Ding (Peking University), Yehui Tang (Huawei), Zhongqian Fu (Huawei), Chao Xu (Peking University), Kai Han (Huawei), Yunhe Wang (Huawei)
-
HSF: Defending against Jailbreak Attacks with Hidden State Filtering
Cheng Qian (Beihang University), Hainan Zhang (Beihang University), Lei Sha (Beihang University)
-
KG-UQ: Knowledge Graph-Based Uncertainty Quantification for Long Text in Large Language Models
Yingqing Yuan (University of Sydney), Linwei Tao (University of Sydney), Haohui Lu (University of Sydney), Matloob Khushi (University of Suffolk), Imran Razzak (Mohamed bin Zayed University of Artificial Intelligence), Mark Dras (Macquarie University), Jian Yang (Macquarie University), Usman Naseem (Macquarie University)
-
Integrating Edge Computing-Based Ethical Decision-Making Framework for AI
Jimin Ryu (Sookmyung Women’s University), Hyeyoung Kim (Sookmyung Women’s University), Yong Ik Yoon (Sookmyung Women’s University)
-
OpenCarbonEval: How much $CO_2$ will your large model exhale during training?
Zhaojian Yu (Tsinghua University), Yinghao Wu (Tsinghua University), Zhuotao Deng (Tsinghua University), Xinchun Yu (Tsinghua University), Yansong Tang (Tsinghua University), Xiao-Ping Zhang (Tsinghua University)
-
Your Diffusion Classifier is Naturally a Robust Classifier
Yanxiang Ma (University of Sydney), Zixuan Huang (Beijing University of Posts and Telecommunications), Shan You (SenseTime Research)
Schedule
| 8:55-9:00 | Start | Opening Remarks |
| 9:00 - 9:30 | Invited Talk 1 | Dr. Jerone Andrews, Sony AI, “Ethical Considerations for Responsible Data Curation” |
| 9:30 - 10:00 | Invited Talk 2 | Dr. Tomasz Bednarz, NVIDIA, “Foundation Models and Microservices for Advancing Generative AI Research Running Efficiently on GPUs Anywhere” |
| 10:00 - 10:30 | Invited Talk 3 | Prof. Flora Salim, The University of New South Wales, “Data-efficient, robust, and continual time-series and multimodal open world learning” |
| 10:30 - 11:00 | Morning tea | |
| 11:00 - 11:15 | Paper presentation 1 | Cheng Qian, Beihang University, “HSF: Defending against Jailbreak Attacks with Hidden State Filtering” |
| 11:15 - 11:45 | Invited Talk 4 | Prof. Fengli Xu, Tsinghua University, “Efficient Reasoning and Optimization Frameworks for LLM Agents” |
| 11:45 - 12:15 | Invited Talk 5 | Prof. Abraham Bernstein, University of Zurich, TBD |
| 12:15 - 12:30 | Paper presentation 2 | Yingqing Yuan, The University of Sydney, “KG-UQ: Knowledge Graph-Based Uncertainty Quantification for Long Text in Large Language Models” |
| 12:30 | Closing |
Location: Room C4.10, ICC Sydney: International Convention & Exhibition Centre
Organizers
Chang Xu
The University of Sydney
Yunke Wang
The University of Sydney
Jianyuan Guo
The University of Sydney
Daochang Liu
The University of Western Australia
Minjing Dong
City University of Hong Kong
Yasmeen George
Monash University
Johan Barthélemy
NVIDIA
Yan Liu
University of Southern California
Ling Chen
University of Technology Sydney
Contact
Contact the organizers at workshopsai25@gmail.com.