Special Days
The Web Conference 2024 features three thematic special days: Online Trust & Safety Day, Large Language Model Day, and Health Day.
Online Trust & Safety Day
Date: May 13, 2024
In the digital age, the integrity of online spaces is paramount. The Online Trust & Safety Day is conceived as a pivotal initiative to address the growing concerns and challenges in maintaining secure and trustworthy digital environments. This event underscores the critical importance of fostering safe online interactions and protecting users from various cyber threats ranging from misinformation and hate speech to privacy breaches and cyberbullying.
As digital platforms increasingly become central to our personal, social, and professional lives, the responsibility to ensure these spaces are safe and inclusive has never been more significant. The advent of sophisticated technologies and algorithms has introduced complex challenges that require equally advanced solutions. Online Trust & Safety Day brings together leading experts and innovators from academia, industry, and government to share insights, discuss emerging trends, and showcase the latest research and technologies aimed at enhancing online safety.
The day’s sessions are designed to highlight the multifaceted nature of trust and safety issues and present collaborative efforts and solutions that span disciplines and borders. By facilitating an exchange of ideas and fostering partnerships, this event aims to catalyze the development of effective strategies and tools that empower users and elevate the standards of trust and safety across all online platforms.
This gathering is not just a conference; it is a call to action for all stakeholders involved in the digital ecosystem to commit to creating safer online environments. Through shared knowledge and collective endeavor, we can tackle the challenges of the digital world and ensure a safer, more trustworthy online experience for everyone.
Programme
| Time | Activity |
|---|---|
|
0900 - 0910 |
Welcome Message Dr. Roy Ka-Wei Lee Assistant Professor, Singapore University of Technology and Design Adjunct Senior Scientist, Centre for Advanced Technologies in Online Safety (CATOS) |
|
0910 - 0950 |
Keynote I: Factuality Challenges in the Era of Large Language Models Dr. Preslav Nakov, Professor and Department Chair of NLP, MBZUAI
We will discuss the risks, the challenges, and the opportunities that Large Language Models (LLMs) bring regarding factuality. We will then delve into our recent work on using LLMs to assist fact-checking (e.g., claim normalization, stance detection, question-guided fact-checking, program-guided reasoning, and synthetic data generation for fake news and propaganda identification), on checking and correcting the output of LLMs, on detecting machine-generated text (blackbox and whitebox), and on fighting the ongoing misinformation pollution with LLMs. Finally, we will discuss work on safeguarding LLMs, and the safety mechanisms we incorporated in Jais-chat, the world's best open Arabic-centric foundation and instruction-tuned LLM. |
|
0950 - 1030 |
Keynote II: Building Trust and Safety on Facebook Lluís Garcia Pueyo Director of Engineering, Meta
This talk from the Facebook Integrity organization within Meta will highlight its mission to enhance safety and trust on Facebook by addressing three critical areas: reducing harmful content, minimizing negative user experiences, and increasing user control over their social media interactions. We'll discuss innovations in machine learning models to handle sparse positive labels and biased classifiers, as detailed in our Semi Supervised Monotonic Regression research. Furthermore, we will explore approaches to mitigate negative experiences from non-violating content through personalized recommendation techniques, as described in our recent study presented at The Web Conference. Additionally, we will cover efforts to empower users with better insights and management tools for their Facebook experience, exemplified by our work on Informative Integrity Frictions. The session will wrap up with opportunities for further engagement and showcasing our continuous collaboration with the research community. |
|
1030 - 1100 |
Coffee Break |
|
1100 - 1230 |
Invited Presentations
The Dynamics of (Not) Unfollowing Misinformation Spreaders Joshua Ashkinaze (Presenter), Eric Gilbert, Ceren Budak
Unmasking the Web of Deceit: Uncovering Coordinated Activity to Expose Information Operations on Twitter Luca Luceri (Presenter), Valeria Pantè, Keith Burghardt, Emilio Ferrara
Modularized Networks for Few-shot Hateful Meme Detection Rui Cao (Presenter), Roy Ka-Wei Lee, Jing Jiang
Fake Resume Attacks: Data Poisoning on Online Job Platforms Michiharu Yamashita (Presenter), Thanh Tran, Dongwon Lee
Analyzing Ad Exposure and Content in Child-Oriented Videos on YouTube Emaan Bilal Khan, Nida Tanveer, Aima Shahid, Mohammad Jaffer Iqbal, Haashim Ali Mirza, Armish Javed, Ihsan Ayyub Qaz (Presenter)i, Zafar Ayyub Qazi
CapAlign: Improving Cross Modal Alignment via Informative Captioning for Harmful Meme Detection Junhui Ji, Lin Xuanrui and Usman Naseem (Presenter) |
|
1230 - 1330 |
Lunch |
|
1330 - 1340 |
Online Safety Prize Challenge - Welcome Brief Dr. See Kiong Ng Professor of Practice, National University of Singapore Director, AI Technology - AI Singapore |
|
1340 - 1500 |
Online Safety Prize Challenge - Finalist Presentations |
|
1500 - 1530 |
Coffee Break |
|
1530 - 1545 |
Online Safety Prize Challenge - Award Ceremony Prize presentation by special guest: Dr. Bernice Khoo, Director, Technology Office and Director (Research), National AI Group, Ministry of Communications and Information |
|
1545 - 1600 |
Closing Remarks Dr. Yinping Yang Senior Principal Scientist, A*Star Director, Centre for Advanced Technologies in Online Safety (CATOS) |
Organizing Chairs:
- Roy Lee (Singapore University of Technology and Design)
- See-Kiong Ng (AI Singapore)
- Wee Siong Ng (AI Singapore)
- Yinping Yang (Centre for Advanced Technologies in Online Safety)
Large Language Model Day
Date: May 14, 2024
Large Language Models (LLMs), such as ChatGPT, Gemini, GLM-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?
We are pleased to announce that, for the first time at The Web Conference, we are organizing a dedicated full-day event---the "Large Language Model Day (LLM Day)." This event aims to provide a platform for academic researchers and industry practitioners to discuss the latest advances and open problems in the field of large language models (LLMs). The event will focus on two main themes:
Recent Progress of Foundational LLMs: This theme will highlight the latest advancements in foundational LLMs and their applications across different domains. It will provide attendees with insights into the cutting-edge developments and potential applications of LLMs.
Building Trustworthy LLMs: This theme will address the open issues and challenges associated with building trustworthy LLMs. Topics of discussion will include fairness, safety, privacy preservation, explainability, and controllability. By tackling these challenges, we aim to ensure that LLMs are developed and deployed in an ethical and responsible manner.
During the Large Language Model Day, we are excited to host a diverse range of renowned invited speakers from world-class research institutions. These experts will share their knowledge and perspectives, fostering interdisciplinary collaboration and promoting the exchange of ideas among participants. Our goal is to contribute to the development and use of LLMs that not only push the boundaries of AI but also benefit humanity as a whole.
Join us at TheWebConf 2024 for this special day dedicated to exploring the vast potential and challenges of large language models. Together, we can shape the future of AI and ensure its responsible and impactful deployment.
Programme
| Time | Activity |
|---|---|
| 1100 | Opening remarks of LLM Day |
| 1100 - 1145 | Safety and Trustworthiness Challenges for LLMS. Mohan Kankanhalli, National University of Singapore, Singapore |
| 1145 - 1230 | Enable LLMs to Reason: Narrowing the gap between human intelligence and machine intelligence. XuezhiWang, Google Brain, US |
| 1230 - 1330 | Lunch Break |
| 1330 - 1415 | Intuitions on large language models. Jason Wei, OpenAI, US |
| 1415 - 1500 | Building frontier LLMs from scratch in the wilderness. Yi Tay, Reka.ai, Singapore |
| 1500 - 1530 | Coffee Break |
| 1530 - 1615 | ChatGLM: A LLM for Building Agents. Xiaotao Gu, Zhipu AI, China |
| 1615 - 1700 | The Geopolitics of AI. Dame Wendy Hall, Southampton University, UK |
Organizing Chair:
- Yuxiao Dong (Tsinghua University)
Health Day
Date: May 14, 2024
The emergence of novel Artificial Intelligence (AI) and Machine Learning (ML) technologies is poised to significantly transform clinical care practices. These advancements are opening new avenues in healthcare delivery and medical research, ultimately enhancing patient outcomes. AI and ML have already demonstrated substantial potential in catalyzing the digital health revolution. This includes early prediction of disease risks, improved forecasting of patient prognosis, and the development of personalized treatment plans to optimize healthcare delivery. Significant achievements have been made in the application of computer vision models for medical imaging, illustrating the robust capabilities of these technologies. Additionally, the use of reinforcement learning models for intervention recommendations presents promising opportunities for further innovation. Moreover, the rapid evolution of Large Language Models (LLMs) has sparked extensive discussions about their potential to reimagine various healthcare practices. This underscores the pivotal role of AI and ML in shaping the future of healthcare, offering a glimpse into a more efficient and effective medical landscape.
The Health Day is a forum to foster interactions among researchers and practitioners in the Healthcare AI and ML field by allowing them to present their new and innovative work. In addition, the track will enable conference attendees to learn novel ongoing research projects through informal interactions.
Programme
| Time | Activity |
|---|---|
| 0900 - 0915 | Opening Address by Organising Chairs |
| 0915 - 1000 | Opening Keynote: Stories Over Snapshots: Optimising Health with Digital Platforms by Prof Dean Ho (Director of N.1 Institute, National University of Singapore) |
| 1000 - 1030 | Keynote: Unleashing the power of AI for Health: the vital role of physiological data by Prof Xiao Hu (Professor and Asa Griggs Candler Chair of Nursing Data Science, Emory University) |
| 1030 - 1100 | Coffee Break |
| 1100 - 1230 | Oral Presentation Session #1 |
| 1230 - 1330 | Lunch Break |
| 1330 - 1430 | Afternoon Opening Keynote: Generative AI for Healthcare by Prof Kee Yuan Ngiam (Group Chief Technology Officer, National University Health System) |
| 1500 - 1530 | Keynote: AI-ECG: Datasets, Methods, Wearable Devices and Clinical Applications by Prof Shenda Hong (National Institute of Health Data Science, Peking University) |
| 1530 - 1645 | Oral Presentation Session #2 |
| 1645 - 1700 | Closing Remarks |
Organizing Chairs:
- Mengling Feng (National University of Singapore)
- Xia "Ben" Hu (Rice University)