Workshop Goals

We are excited to announce the AI4Bio Workshop on Artificial Intelligence for Biology. Organized by the Machine Learning department of MBZUAI, this workshop aims to explore the intersection of artificial intelligence and biology, with a focus on the latest advancements in AI techniques applied to biological data. The event is designed to foster collaboration and accelerate progress in AI-driven research in biology. We will bring together leading experts to discuss state-of-the-art methodologies, share groundbreaking research findings, and identify key challenges and opportunities in the AI for Biology field.

Click here for more information

There is significant debate around the future of Human Computer Interaction (HCI) in the emerging world of AI. Some say that the field of human-computer interaction has been rendered irrelevant by new kinds of AI systems, such as LLM chatbots, designed to directly interact with people. Many researchers in HCI, however, have shown the ways in which these systems may ignore people’s needs, trample on their rights, and demonstrate a very narrow vision of what it means for AI to be human-centered. We have gathered some of the most important voices in the field to discuss the future of HCI in this new era of AI. The attendees come from HCI fields as diverse as learning technologies, XR (virtual and enhanced reality), accessibility and health, multimodal technologies, body-centered devices, mobile devices, human-robot interaction, and search and information retrieval. Participants will debate what aspects of HCI deserve investment of time and other resources, to contribute and shape new AI initiatives, and what HCI education must look like in order to support the next generation of scholars and contributors in the area of HCI in AI and AI in HCI.

• Sat 9: Presentations & discussion
• Sun 10: Presentations & discussion
• Mon 11: Round-up and cultural activity

Organizers:

• Michael I. Jordan, UC Berkeley
• Sai Praneeth Karimireddy, UC Berkeley
• Yaodong Yu, UC Berkeley

Local committee:

• Samuel Horvath, MBZUAI
• Karthik Nandakumar, MBZUAI
• Praneeth Vepakomma, MIT, MBZUAI

The pursuit of sustainable development has become a global priority as societies strive to ensure a harmonious coexistence between economic growth, social well-being, and environmental preservation. Achieving the ambitious sustainable development goals (SDGs) outlined by the United Nations requires a comprehensive understanding of complex systems and the availability of high-quality data to guide informed decision-making.

However, much of the data required for this purpose is sensitive and private, belonging to individuals or organizations that are understandably reluctant to share it due to concerns about privacy, security, and potential misuse. In fact, the UN “Global Pulse” initiative has identified lack of data access to be the key bottleneck for unleashing AI for achieving SDGs. This poses a significant hurdle to harnessing the full potential of data-driven approaches for sustainable development. Fortunately, collaborative learning may present a solution to this predicament.

In addition to addressing data privacy concerns, collaborative learning has the potential to embody a more inclusive approach to sustainable development. Traditional centralized approaches to data analysis often exclude local stakeholders from the decision-making process, limiting their ability to actively participate in shaping interventions that directly impact their communities. Collaborative learning, on the other hand, empowers these stakeholders by requiring them to actively contribute to the model training process.

The decentralized nature of collaborative learning can also lead to more scalable and efficient systems, particularly in scenarios where data sources are geographically dispersed or resource-constrained. By leveraging local computational resources, we can reduce the reliance on centralized infrastructure, making it a more viable option for initiatives in resource-limited settings.

Watch the Workshops

Research Session – Thursday, October 13, 10 a.m. to 4 p.m.

1. AI-Guided Experiments for ML Systems, AutoML, and Science, Willie Neiswanger, Stanford University
2. Standard Model and Algorithm for ML, Zhiting Hu, Assistant Professor at UC San Diego
3. Accelerated Deep Learning via Efficient, Compressed and Managed Communication, Marco Canini, Associate Professor at KAUST
4. Algorithms and Software for Text Classification, Chih-Jen Lin, Distinguished Professor at National Taiwan University
5. Causal Representation Learning: Advances and Perspective, Kun Zhang, Associate Professor at MBZUAI and Director of the Center for Integrative Artificial Intelligence (CIAI)
6. Systems and Applications of Multilevel Optimization, Sang Choe, Carnegie Mellon University
7. Accelerating Drug Discovery via Deep Learning + Molecular Simulation, Maruan Al-Shedivat, Principal Research Scientist at Genesis Therapeutics
8. AI Ecosystems: The Case of Natural Language Processing, Timothy Baldwin, Professor at MBZUAI
9. Grounded Multi-modal Pretraining and Applications, Xiaodan Liang, Associate Professor at Sun Yat-sen University
10. Fair and Accurate Federated Learning under Heterogeneous Targets, Samuel Horvath, Assistant Professor at MBZUAI
11. On the Utility of Gradient Compression in Distributed Training Systems, Hongyi Wang, Carnegie Mellon University
12. Scaling Deep Learning Training With Compiler Optimizations, Jinliang Wei, Google

Industry Session (open for public) – Friday, October 14, 10 a.m. to 4 p.m.

1. Opening remarks: AI at Scale
2. Growing a workforce of talents for AI at Scale, Co-Founder at Petuum Inc., Qirong Ho, Assistant Professor at MBZUAI
3. Panel discussion: AI at Scale in the Middle East
4. Improving Correctness, Performance, and Energy Efficiency of ML Applications, Shan Lu, Professor at University of Chicago
5. History and Systems forTraining AI at Scale, Hao Zhang, UC Berkeley
6. Fantastic Large Models andWhat We Should Do, Hector Liu, Head of Engineering at Petuum Inc.
7. AI System for Drug Design, Prof. Le Song, Department Chair of Machine Learning, and Professor of Machine Learning