Keynote Speakers

  1. Home
  2. Keynote Speakers

Evolutionary Deep Learning and Applications to Image Classification
Mengjie Zhang

Image classification problems occur in our everyday life. Recognising faces in digital images and diagnosing medical conditions from X-Ray images are just two examples of the many important tasks for which we need computer based image classification systems. Since the 1980s, many image analysis algorithms have been developed. Among those algorithms, deep learning particularly deep convolutional neural networks have received very good success and attracted attentions to industry people and researchers in computer vision and image processing, neural networks, and machine learning. However, there are at least three major limitations in deep convolutional neural networks: (1) the learning architecture including the number of layers, the number of feature maps in each layer and the number of nodes in each feature map are still very much determined manually via "trial and error", which requires a large amount of hand-crafting/trial time and good domain knowledge. However, such experts are hard to find in many cases, or using such expertise is too expensive. (2) Almost all the current deep learning algorithms need a large number of examples/instances (e.g. AlphaGo used over 30 million instances) that many problems do not have. (3) Those algorithms require a huge computational cost that big companies such as Google, Baidu, and Microsoft can cope well but most universities and research institutions cannot.

To address these limitations, evolutionary computation techniques start playing a significant role for automatically determining deep structures, transfer functions and parameters to tackle image classification tasks, and have great potential to advance the developments of deep structures and algorithms. This talk will provide an extended view of deep learning, overview the state-of-the-art work in evolutionary deep learning using Genetic Algorithms (GAs), Particle Swarm Optimisation (PSO) and Differential Evolution (DE). In addition, we will discuss some recent developments using Genetic Programming (GP) to automatically evolving deep structures and feature construction for image recognition with a highlight of the interpretation capability and visualisation of constructed features.

Mengjie Zhang Mengjie Zhang is a Fellow of Royal Society of New Zealand, a Fellow of IEEE, an IEEE Distinguished Lecturer, currently Professor of Computer Science at Victoria University of Wellington, where he heads the interdisciplinary Evolutionary Computation Research Group. He is a member of the University Academic Board, a member of the University Postgraduate Scholarships Committee, a member of the Faculty of Graduate Research Board at the University, Associate Dean (Research and Innovation) in the Wellington Faculty of Engineering, and Chair of the Research Committee of the Faculty of Engineering and School of Engineering and Computer Science.

His research is mainly focused on artificial intelligence (AI), machine learning and big data, particularly in evolutionary computation and learning (using genetic programming, particle swarm optimisation and learning classifier systems), feature selection/construction and big dimensionality reduction, computer vision and image processing, job shop scheduling and resource allocation, multi-objective optimisation, classification with unbalanced data and missing data, and evolutionary deep learning and transfer learning. Prof Zhang has published over 500 research papers in refereed international journals and conferences in these areas. He has been serving as an associated editor or editorial board member for over ten international journals including IEEE Transactions on Evolutionary Computation, IEEE Transactions on Cybernetics, IEEE Transactions on Emergent Topics in Computational Intelligence, ACM Transactions on Evolutionary Learning and Optimisation, the Evolutionary Computation Journal (MIT Press), Genetic Programming and Evolvable Machines (Springer), Applied Soft Computing, Natural Computing, and Engineering Applications of Artificial Intelligence, and as a reviewer of over 30 international journals. He has been involving major AI and EC conferences such as GECCO, IEEE CEC, EvoStar, AAAI, PRICAI, PAKDD, AusAI, IEEE SSCI and SEAL as a Chair. He has also been serving as a steering committee member and a program committee member for over 100 international conferences. Since 2007, he has been listed as one of the top ten (currently No. 4) world genetic programming researchers by the GP bibliography (http://www.cs.bham.ac.uk/~wbl/biblio/gp-html/index.html).

Prof Zhang is the (immediate) past Chair of the IEEE CIS Intelligent Systems Applications, the IEEE CIS Emergent Technologies Technical Committee and the IEEE CIS Evolutionary Computation Technical Committee, a vice-chair of the IEEE CIS Task Force on Evolutionary Feature Selection and Construction, a vice-chair of the IEEE CIS Task Force on Evolutionary Computer Vision and Image Processing, and the founding chair of the IEEE Computational Intelligence Chapter in New Zealand.

School of Engineering and Computer Science, Victoria University of Wellington

https://people.wgtn.ac.nz/Mengjie.Zhang

Mengjie.Zhang@ecs.vuw.ac.nz

Cinematic XR – Teleport into the Video
Taehyun Rhee

New Zealand is well known for beautiful real nature contributed to live action background for many movies and commercials. A strong post processing pipeline for Computer Graphics and Visual Effect has been built, attract global talents and investment into the country.

Recent advancement of hardware and software technology changed the traditional movie watching experience with immersion and interaction toward presence at the story. Convergence of the immersive, interactive, and intelligent technology produces new opportunity to foster the future digital media.

This talk will present recent research in immersive, interactive, and intelligent media technology. We will introduce research in image-based lighting, mixed reality rendering, image reconstruction, and realistic composition, which has been contributed to high fidelity mixed reality in films, games, and XR applications.

Convergence of the solutions provides a novel Cinematic XR platform allowing the user’s illusion to teleport and interact with the scene objects in the video due to the coherent illumination, shadows, and seamless blending between real and virtual scene computed in real-time. Inverse rendering technology to estimate scene environment using machine learning will be further presented to provide solutions for the mobile platform.

The recent extension to adapt high-speed network and live streaming, provides a novel platform for immersive telecollaboration. Finally, we will introduce a novel asymmetric system, Augmented Virtual Teleportation (AVT), which converge virtual reality (VR) and augmented reality (AR) into a high-fidelity mixed reality (MR) space in live. It will be able to realise telepresence into the video and immersive interaction to connect people in distance. Possible applications and extensions will be further discussed at the talk.

A/Prof. Taehyun James (TJ) Rhee is a director of the Computational Media Innovation Centre (CMIC) at Wellington Faculty of Engineering, and Associate Professor at Victoria University of Wellington, New Zealand.

He is a founder of the Victoria Computer Graphics Programme, founder/director of the Victoria Computer Graphics Research Lab, and a founder of the Mixed Reality startup, DreamFlux.

His current research activities are focused on developing future media technology and platform; cinematic XR including real-time lighting, rendering, composition in virtual, augmented, and mixed reality; virtual teleportation; immersive remote collaboration; immersive visualization and interaction; and human digital content interaction.

He is highly interested in prototyping the research outcome into potential commercial products and platforms; a winner of 2018 Researcher Entrepreneur Award by Kiwinet.

He is serving for the Computer Graphics community as a conference chair of Pacific Graphics 2020 and 2021, executive committee of Asia Graphics Association, and SIGGRAPH Asia 2018 Virtual and Augmented Reality programme chair.

Before joining Victoria in 2012, he was a principal researcher and senior manager in the Mixed Reality Group, Future IT Centre at Samsung (2008-2012). He was also a senior researcher/researcher of Research Innovation Center at Samsung Electronics (1996-2003).

Computational Media Innovation Centre (CMIC), Victoria University of Wellington

https://people.wgtn.ac.nz/Taehyun.Rhee

taehyun.rhee@ecs.vuw.ac.nz

Expressive Facial Modeling and Animation
Karan Singh

Humans are hard-wired to see and interpret minute facial detail. The rich signals we extract from facial expressions imposes high expectations for computer generated facial imagery. This talk focuses on the science and art of expressive facial animation. Specifically, aspects of facial anatomy, biomechanics, linguistics and perceptual psychology will be used to motivate and describe the construction of geometric face rigs, and techniques for the animator-centric creation of emotion, expression and speech animation from input images, audio and video.

Karan Singh is a Professor of Computer Science at the University of Toronto. He co-directs a globally reputed graphics and HCI lab, DGP, has over 100 peer-reviewed publications, and has supervised over 40 MS/PhD theses. His research interests lie in interactive graphics, spanning art and visual perception, geometric design and fabrication, character animation and anatomy, and interaction techniques for mobile, Augmented and Virtual Reality (AR/VR). He has been a technical lead for the Oscar award winning software Maya and was the R&D Director for the 2004 Oscar winning animated short Ryan. He has co-founded multiple companies including Arcestra (architectural design), JALI (facial animation), and JanusVR (Virtual Reality).

ACM involvement:
  • ACM SIGGRAPH student volunteer in 1992
  • Pioneer member of ACM SIGGRAPH and SIGCHI since 2002
  • Program or conference chair of ACM sponsored conferences Symposium of Computer Animation (SCA 2002), Symposium of Applied Perception (SAP 2014), and Sketch-Based Interfaces and Modeling (SBIM 2008 and SBIM 2012)

Department of Computer Science, University of Toronto

https://www.dgp.toronto.edu/~karan/

karan@dgp.toronto.edu

One  Ring  Representation to Rule Them All
Andrea Tagliasacchi

Synthesizing photorealistic images has historically been the core objective of Computer Graphics. Yet, thanks to deep learning, vision researchers have recently shown that it is possible to synthesize visual content of quality that is sufficient to fool the human eye. The pivotal point has been to discount the role of polygonal meshes in learning – the 3D representation that has dominated the last 30 years of computer graphics R&D. This has been made possible by marrying ideas from traditional computer graphics (volume rendering) with modern deep learning (functions stored within networks). These representations are not only amenable to unsupervised end-to-end learning from raw imagery (easy to differentiate), but also do not suffer the curse of dimensionality (easy to store). However, the computational toolbox to operate on such representations is still rather limited, therefore restricting their usage in practical applications. My research addresses this issue by bridging the gap between deep learning and contemporary Computer Graphics, developing hybrid representations that inherit benefits from both worlds. At the same time, I look at this problem from a Computer Vision perspective, as we seek representations that are structured, where the concept of objects within scenes should be learnt autonomously. Overall, by jointly tackling the problem from a Graphics and Vision perspective, my hope is to develop 3D deep representations where one is not forced to choose between artistic controllability and visual fidelity.

Andrea Tagliasacchi is a staff research scientist at Google Brain and an adjunct faculty in the computer science department at the University of Toronto. His research focuses on 3D perception, which lies at the intersection of computer vision, computer graphics and machine learning. In 2018, he was invited to join Google Daydream as a visiting faculty and eventually joined Google full time in 2019. Before joining Google, he was an assistant professor at the University of Victoria (2015-2017), where he held the "Industrial Research Chair in 3D Sensing". His alma mater include EPFL (postdoc) SFU (PhD, NSERC Alexander Graham Bell fellow) and Politecnico di Milano (MSc, gold medalist).

Google Brain and Adjunct Professor for the Department of Computer Science, University of Toronto

https://taiya.github.io/

taglia@google.com

Copyright © 2020 Victoria University of Wellington. All Rights Reserved.

Log In