1. Machine Learning for Computer Graphics An brief introduction By Dr. Zhang Hongxin State Key Lab of CAD&CG, ZJU

2. Outline Background Background What is Machine Learning? What is Machine Learning? Is it really useful for computer graphics? Is it really useful for computer graphics? Our plan Our plan

3. The largest challenge of Today’s CG&A The tedious effort required to create digital worlds and digital life. The tedious effort required to create digital worlds and digital life. “Finding new ways to communicate and new kinds of media to create.” “Finding new ways to communicate and new kinds of media to create.” Filmmakers, scientists, graphic designers, fine artists, and game designers. Filmmakers, scientists, graphic designers, fine artists, and game designers.

6. Pure procedural synthesis vs. Pure data Creating motions for a character in a movie Creating motions for a character in a movie Pure procedural synthesis. Pure procedural synthesis. compact, but very artificial, rarely use in practice. compact, but very artificial, rarely use in practice. “By hand” or “pure data”. “By hand” or “pure data”. higher quality but lower flexibility. higher quality but lower flexibility. the best of both worlds: hybrid methods?!? the best of both worlds: hybrid methods?!?

7. Bayesian Reasoning  Principle modeling of uncertainty.  General purpose models for unstructured data.  Effective algorithm for data fitting and analysis under uncertainty.  But currently it is always used as an black box.

8. Data driven modeling

9. What is machine learning? Data mining (KDD) Data-base Computer Vision Multi-mediaBio-informatics Artificial Intelligence Computer Graphics Machine Learning Statistics and Bayesian methods Control and information Theory

10. What is machine learning? (Cont.) Definition by Mitchell, 1997 A program learns from experience E with respect to some class of tasks T and performance measure P, if its performance at task T, as measured by P, improves with experience E. Learning systems are not directly programmed to solve a problem, instead develop own program based on: examples of how they should behave from trial-and-error experience trying to solve the problem Different than standard CS: want to implement unknown function, only have access to sample input-output pairs (training examples) Hertzmann, 2003 For the purposes of computer graphics, machine learning should really be viewed as a set of techniques for leveraging data(???).

11. Why Study Learning? Develop enhanced computer systems Develop enhanced computer systems automatically adapt to user, customize often difficult to acquire necessary knowledge discover patterns offline in large databases (data mining) Improve understanding of human, biological learning computational analysis provides concrete theory, predictions explosion of methods to analyze brain activity during learning Timing is good growing amounts of data available cheap and powerful computers suite of algorithms, theory already developed

12. Main class of learning problems Learning scenarios differ according to the available information in training examples Supervised: correct output available Classification: 1-of-N output (speech recognition, object recognition,medical diagnosis) Regression: real-valued output (predicting market prices, temperature) Unsupervised: no feedback, need to construct measure of good output Clustering : Clustering refers to techniques to segmenting data into coherent “clusters.” Reinforcement: scalar feedback, possibly temporally delayed

13. And more … Time series analysis. Time series analysis. Dimension reduction. Dimension reduction. Model selection. Model selection. Generic methods. Generic methods. Graphical models. Graphical models.

14. Is it really useful for computer graphics? Con: Everything is machine learning or everything is human tuning? Con: Everything is machine learning or everything is human tuning? Sometimes, this may be true. Sometimes, this may be true. Pro: more understanding of learning, but yields much more powerful and effective algorithms. Pro: more understanding of learning, but yields much more powerful and effective algorithms. Problem taxonomy. Problem taxonomy. General-purpose models. General-purpose models. Reasoning with probabilities. Reasoning with probabilities.  I believe the mathematic magic.

15. Mesh Processing– clustering/segmentation Hierarchical Mesh Decomposition using Fuzzy Clustering and Cuts. Hierarchical Mesh Decomposition using Fuzzy Clustering and Cuts. By Sagi Katz and Ayellet Tal, SIGGRAPH 2003

16. Texture synthesis and analysis - Hidden Markov Model "Texture Synthesis over Arbitrary Manifold Surfaces", by Li-Yi Wei and Marc Levoy. In Proceedings of SIGGRAPH 2001. "Texture Synthesis over Arbitrary Manifold Surfaces", by Li-Yi Wei and Marc Levoy. In Proceedings of SIGGRAPH 2001. "Fast Texture Synthesis using Tree-structured Vector Quantization", by Li- Yi Wei and Marc Levoy. In Proceedings of SIGGRAPH 2000. "Fast Texture Synthesis using Tree-structured Vector Quantization", by Li- Yi Wei and Marc Levoy. In Proceedings of SIGGRAPH 2000.

17. Image processing and synthesis by graphical Model Image Quilting for Texture Synthesis and Transfer. Alexei A. Efros and William T. Freeman. SIGGRAPH 2001. Image Quilting for Texture Synthesis and Transfer. Alexei A. Efros and William T. Freeman. SIGGRAPH 2001. Graphcut Textures: Image and Video Synthesis Using Graph Cuts. Vivek Kwatra, Irfan Essa, Arno Schödl, Greg Turk, Aaron Bobick. SIGGRAPH 2003. Graphcut Textures: Image and Video Synthesis Using Graph Cuts. Vivek Kwatra, Irfan Essa, Arno Schödl, Greg Turk, Aaron Bobick. SIGGRAPH 2003.

19. BTF – reflectance texture synthesis Synthesizing Bidirectional Texture Functions for Real- World Surfaces. Xinguo Liu, Yizhou Yu and Heung- Yeung Shum. More recent papers…

20. Style machines - Time series analysis By Matthew Brand (MERL) and Aaron Hertzmann. SIGGRAPH 2000 By Matthew Brand (MERL) and Aaron Hertzmann. SIGGRAPH 2000

21. Motion texture - linear dynamic system Yan Li, Tianshu Wang, and Heung-Yeung Shum. Motion Texture: A Two-Level Statistical Model for Character Motion Synthesis. Yan Li, Tianshu Wang, and Heung-Yeung Shum. Motion Texture: A Two-Level Statistical Model for Character Motion Synthesis.

22. Video Textures – Reinforce Learning Arno Schödl, Richard Szeliski, David H. Salesin, and Irfan Essa. Video textures. Proceedings of SIGGRAPH 2000, pages 489-498, July 2000.Video textures

23. Human shapes – Dimension Reduction The Space of Human Body Shapes: Reconstruction and Parameterization From Range Scans. Brett Allen, Brian Curless, Zoran Popović. SIGGRAPH 2003. The Space of Human Body Shapes: Reconstruction and Parameterization From Range Scans. Brett Allen, Brian Curless, Zoran Popović. SIGGRAPH 2003. A Morphable Model for the Synthesis of 3D Faces. Volker Blanz and Thomas Vetter. SIGGRAPH 1999. A Morphable Model for the Synthesis of 3D Faces. Volker Blanz and Thomas Vetter. SIGGRAPH 1999.

24. Our Plan The SIG-ML4CG The SIG-ML4CG http://www.cad.zju.edu.cn/home/z hx/ML4CG/ http://www.cad.zju.edu.cn/home/z hx/ML4CG/ http://www.cad.zju.edu.cn/home/z hx/ML4CG/ http://www.cad.zju.edu.cn/home/z hx/ML4CG/

25. Schedule

26. Textbooks Information Theory, Inference, and Learning Algorithms, Information Theory, Inference, and Learning Algorithms, by David MacKay. by David MacKay. Machine Learning Machine Learning by Tom Mitchell. by Tom Mitchell. Data mining: Concepts and Techniques Data mining: Concepts and Techniques By Jiawei Han and Micheline Kamber By Jiawei Han and Micheline Kamber Pattern Classification (2nd ed.) Pattern Classification (2nd ed.) by Richard O. Duda, Peter E. Hart and David G. by Richard O. Duda, Peter E. Hart and David G.

27. Reference Links Links ML4CG by Hertzmann http://www.dgp.toronto.edu/~hertzman/mlcg2003/hert zmann-mlcg2003.pdf ML4CG by Hertzmann http://www.dgp.toronto.edu/~hertzman/mlcg2003/hert zmann-mlcg2003.pdf http://www.dgp.toronto.edu/~hertzman/mlcg2003/hert zmann-mlcg2003.pdf http://www.dgp.toronto.edu/~hertzman/mlcg2003/hert zmann-mlcg2003.pdf DDM concepts http://datamining.ihe.nl/symposium/intro.htm DDM concepts http://datamining.ihe.nl/symposium/intro.htm http://datamining.ihe.nl/symposium/intro.htm

28. Q & A Thanks Thanks