Speakers

Bio:  Alla Sheffer is a Professor of Computer Science at the University of British Columbia and a Scholar at Amazon Inc. She received her BSc (1991), MSc (1995), and PhD (2000) from Hebrew University, Jerusalem, Israel. She investigates algorithms for geometry processing, focusing on computer graphics applications. She is particularly interested in leveraging connections between geometry and perception to enable users to create and manipulate geometric content, including garments and 3D printable artifacts. Prof. Sheffer regularly publishes at selective computer graphics venues and has co-authored over 50  papers published in ACM Transactions on Graphics, including numerous papers in SIGGRAPH and SIGGRAPH Asia proceedings. Sheffer is a Fellow of IEEE, a Fellow of ACM, a Fellow of the Royal Society of Canada, a Fellow of Eurographics, and a Member of the SIGGRAPH Academy. She is the recipient of the 2018 Canadian Human Computer Communications Society Achievement Award; a UBC Killam Research Award; multiple faculty awards from Adobe, Google and IBM; an NSERC Discovery Accelerator award; and an Audi Production Award. She was the Technical Papers Committee Chair for SIGGRAPH’23 and co-chaired the program committees for Eurographics’18, 3DV’18, PG’19, SGP’06 and IEEE SMI’13. She has served on the editorial boards of ACM TOG, IEEE TVCG, Computer Graphics Forum, Graphical Models, Computers & Graphics, and CAGD.

Title: Explainable Geometric Algorithms

Abstract: Supporting artists and designers creating 2D and 3D content requires solving a myriad geometric problems. In this talk I will focus on problems, which once well-understood can be solved using fully explainable (no AI inside) methods. I will include examples across a wide range of 2D and 3D geometry processing tasks, including clip art and 3D geometry reshaping, VR interfaces for shape modeling,  3D reconstruction from sketches, and real-time atlasing for texture space shading. The common thread in our approach  to these problems is the formulation of the desired outcomes as the solution of discrete-continuous optimization problems followed by derivation of efficient, tailored optimization algorithms for solving them.

Bio: Parmit Chilana is an Associate Professor in the School of Computing Science at Simon Fraser University, where she holds the Ebco-Eppich Research Chair in Human-Computer Interaction. Her research focuses on designing and deploying interactive systems that improve software learnability, particularly in complex domains. She is widely recognized for pioneering new approaches to help retrieval and in-context learning, and is passionate about using interdisciplinary methods to design for user diversity and empower people of all backgrounds and skill levels to use, learn, and program emerging technologies. Parmit’s work has been recognized with numerous accolades, including the CS-Can|Info-Can Early Career Researcher Award, an NSERC Discovery Accelerator Supplement, and multiple Best Paper and Honorable Mention awards at top venues such as ACM CHI. At SFU, she co-directs the cross-departmental Visual and Interactive Computing Institute (VINCI) and regularly serves on program and organizing committees for international conferences including CHI, UIST, and VL/HCC. Before joining SFU, Parmit was a faculty member at the University of Waterloo. She earned her PhD from the University of Washington, where she co-founded AnswerDash, a venture-funded startup that commercialized her award-winning dissertation on crowdsourced contextual help retrieval.

Title: Empowering Users in the Age of AI: Beyond One-Size-Fits-All Approaches to Software Help and Learning

Abstract: As software interfaces become increasingly complex and powered by artificial intelligence (AI) and automation, the expectation is that these technologies will streamline knowledge work and boost productivity. Yet, this promise is often hindered by a persistent gap in software learnability. In this talk, I will discuss how end users—particularly those working with feature-rich applications—continue to struggle with discovering functionality, resolving issues, and learning effectively, despite an abundance of tutorials, forums, and online documentation. Drawing on over a decade of research in human-computer interaction, I will present novel interactive systems that deliver context- and task-specific help, learning materials, and examples to support users as they navigate complex software environments. Through studies conducted in both lab and real-world settings, I highlight the diverse ways people learn and seek help, and why one-size-fits-all solutions fail to support their nuanced workflows. While generative AI offers significant potential for reshaping knowledge work, I argue that adopting a human-centered approach is more important than ever—one that augments user agency rather than replaces it. Ultimately, the goal is to design software that adapts to users’ evolving goals, experiences, and expertise—empowering them to work more effectively and creatively, while allowing them to remain in control.

Bio: Sowmya Somanath is an Associate Professor in the Department of Computer Science at the University of Victoria (UVic) in Canada. At UVic, she co-leads the VIXI lab, which focuses broadly on Human-Computer Interaction and Information Visualization research. Her research focuses on studying and designing software and hardware-based systems to foster creativity and happiness within specific domain contexts such as making, digital fabrication, wearables, knowledge work, and accessibility. Before joining UVic in 2020, Sowmya was an Assistant Professor at Ontario College of Art and Design University in Toronto for two years. She completed her PhD in Computer Science from the University of Calgary in 2017.

Title: Designing Tools for Fostering Creativity and Happiness

Abstract: Engaging in creative activities, such as making physical computing projects, fabricated objects, or craft can bring joy, meaning, personalized solutions, and purpose to people’s lives. However, making artifacts can be challenging, often requiring people to develop technical skills new to them, get access to necessary material resources and tools, and cultivate specific problem-solving approaches. In this talk, I will present novel interactive systems that my students and I have developed, which show how people can be supported when making artifacts in different settings, such as at makerspaces, educational environments, or in collaboration with other stakeholders. I will also present some of our new and ongoing projects focused on designing digital tools to foster happiness within the context of creative activities and knowledge work. Through these projects I hope to highlight the role digital tools can play in empowering people to address personal and societal needs.

Bio:  Dr. Ali Mahdavi-Amiri is an assistant professor at Simon Fraser University working on visual computing applications. His research has resulted in several publications in major journals and venues including SIGGRAPH, SIGGRAPH Asia, TOG, CVPR, ICCV, ECCV, ICLR, NeurIPS, CGF, TVCG, etc. He is also the Research Director at MARZ VFX, where he leads a team focused on developing cutting-edge tools for visual effects applications, including facial aging and de-aging, as well as lip synchronization for dubbing.

Title: Leveraging diffusion models beyond text-based generation

Summary:  Generative models, particularly diffusion models, have demonstrated remarkable abilities in generating images, videos, and 3D shapes. This talk will explore their application across various domains, including the creation of artistic works like creative logos, and their potential to assist artists and users in producing their envisioned results. I will discuss how to enhance control over the results produced by diffusion models by properly conditioning them with various types of constraints to specify the desired outputs. Additionally, these models have underlying knowledge about objects' semantics and compositions in a scene that can be leveraged to address vision tasks such as semantic segmentation and object counting. I will show how to benefit from such knowledge to avoid costly annotations.