Machine Learning in Computer Graphics

You are here

Credits
3
Types
Elective
Requirements
This subject has not requirements
Department
UB
Mail
The application of Machine Learning (ML) techniques in Computer Graphics (CG) is rapidly increasing. The massive computations required in many CG applications together with the ability of ML to identify and explore coherence in these computations lead to a natural symbiosis of the two fields. This course will provide an overview of recent applications of ML to solve CG problems with a focus in photo-realistic rendering applications, allowing the students to link different ML-based approaches with practical application cases in CG.

Teachers

Person in charge

  • Ricardo Jorge Rodrigues Sepúlveda Marques ( )

Weekly hours

Theory
1
Problems
0
Laboratory
1
Guided learning
0
Autonomous learning
3.8

Competences

Generic Technical Competences

Generic

  • CG2 - Capability to lead, plan and supervise multidisciplinary teams.
  • CG3 - Capacity for modeling, calculation, simulation, development and implementation in technology and company engineering centers, particularly in research, development and innovation in all areas related to Artificial Intelligence.

Technical Competences of each Specialization

Academic

  • CEA3 - Capability to understand the basic operation principles of Machine Learning main techniques, and to know how to use on the environment of an intelligent system or service.
  • CEA12 - Capability to understand the advanced techniques of Knowledge Engineering, Machine Learning and Decision Support Systems, and to know how to design, implement and apply these techniques in the development of intelligent applications, services or systems.
  • CEA13 - Capability to understand advanced techniques of Modeling , Reasoning and Problem Solving, and to know how to design, implement and apply these techniques in the development of intelligent applications, services or systems.

Professional

  • CEP1 - Capability to solve the analysis of information needs from different organizations, identifying the uncertainty and variability sources.
  • CEP3 - Capacity for applying Artificial Intelligence techniques in technological and industrial environments to improve quality and productivity.
  • CEP4 - Capability to design, write and report about computer science projects in the specific area of ??Artificial Intelligence.

Transversal Competences

Teamwork

  • CT3 - Ability to work as a member of an interdisciplinary team, as a normal member or performing direction tasks, in order to develop projects with pragmatism and sense of responsibility, making commitments taking into account the available resources.

Information literacy

  • CT4 - Capacity for managing the acquisition, the structuring, analysis and visualization of data and information in the field of specialisation, and for critically assessing the results of this management.

Reasoning

  • CT6 - Capability to evaluate and analyze on a reasoned and critical way about situations, projects, proposals, reports and scientific-technical surveys. Capability to argue the reasons that explain or justify such situations, proposals, etc..

Analisis y sintesis

  • CT7 - Capability to analyze and solve complex technical problems.

Basic

  • CB6 - Ability to apply the acquired knowledge and capacity for solving problems in new or unknown environments within broader (or multidisciplinary) contexts related to their area of study.
  • CB8 - Capability to communicate their conclusions, and the knowledge and rationale underpinning these, to both skilled and unskilled public in a clear and unambiguous way.
  • CB9 - Possession of the learning skills that enable the students to continue studying in a way that will be mainly self-directed or autonomous.

Objectives

  1. Acquire an overview of the field of Computer Graphics, and of Physically-Based Rendering techniques in particular.
    Related competences: CEA13, CG2, CT3, CT6, CT7, CB6, CB9,
    Subcompetences:
    • Overview of the Computer Graphics field and the main current challenges.
    • Details about the open problem of Physically-Based Rendering (PBR) and the Light Transport Equation (LTE) on which we will focus during this course.
  2. Achieve an in-depth understanding of Monte Carlo Methods for Physically-Based Rendering
    Related competences: CEA13, CG2, CG3, CEP1, CT3, CT4, CT6, CT7, CB6, CB8, CB9,
    Subcompetences:
    • Understand how to improve the performance of Monte Carlo Methods through variance reduction techniques, and the main limitations of the typical approaches
    • Details on the use of Monte Carlo methods for PBR.
    • Understand why Monte Carlo methods are needed and ubiquitous in photo-realistic image synthesis.
  3. Learn and experiment with Machine Learning (ML) techniques for Boosting Monte Carlo Methods in PBR.
    Related competences: CEA3, CEA12, CEA13, CG2, CG3, CEP1, CEP3, CEP4, CT3, CT4, CT6, CT7, CB6, CB8, CB9,
    Subcompetences:
    • Analyis of different ML-based approaches to overcome some of the limitations of Monte Carlo methods for PBR.

Contents

  1. Block 1: Introduction to Computer Graphics and Rendering Techniques
    This first block provides an overview of the Computer Graphics field and the main current challenges. It will also provide details about the open problem of Physically-Based Rendering (PBR) and the Light Transport Equation (LTE) on which we will focus during this course.
  2. Block 2: Monte Carlo Methods for Physically-Based Rendering
    This block presents the use of Monte Carlo methods for PBR. We will see why Monte Carlo methods are needed and ubiquitous in photo-realistic image synthesis, how to improve their performance through variance reduction techniques, and the main limitations of the typical approaches.
  3. Block 3: Machine Learning (ML) for Boosting Monte Carlo Methods in PBR
    In this third block we will cover different ML-based approaches to overcome some of the limitations identified in the previous block.

Activities

Activity Evaluation act




Students' Presentation

Students' Presentation
Objectives: 3
Contents:
Theory
1h
Problems
0h
Laboratory
1h
Guided learning
0h
Autonomous learning
3h

Teaching methodology

The weekly schedule of in-person activities is distributed in two hours of class that includes theory and practice.

As far as possible, the gender perspective will be incorporated in the development of the subject. In addition, teachers will be attentive to those specific gender needs that students may raise, such as being able to choose a partner of the same gender if group work is carried out or being able to pose challenges against the gender gap.

Evaluation methodology

The course will follow a continuous evaluation consisting of:

Practical Project (60%) + Presentation and Report on a Research Paper (40%).

Students will work in groups. Marks for oral presentations, project development and submitted reports will be awarded on an individual basis.

Bibliography

Basic:

  • Physically based rendering : from theory to implementation - Pharr, Matt; Jakob, Wenzel; Humphreys, Greg, Morgan Kaufmann Publisher, 2016. ISBN: 9780128006450
    http://cataleg.upc.edu/record=b1521142~S1*cat
  • Advanced Global Illumination (2nd Edition) - Dutré, Philip; Bala, Kavita; Beckaert, Philippe, A. K. Peters, Ltd., 2006. ISBN: 978-1568813073
  • Pattern recognition and machine learning - Bishop, Christopher M, Springer, cop. 2006. ISBN: 978-0-387-31073-2
    http://cataleg.upc.edu/record=b1298151~S1*cat
  • Gaussian processes for machine learning - Rasmussen, Carl Edward, The MIT Press, cop. 2006. ISBN: 978-0-262-18253-9
    http://cataleg.upc.edu/record=b1294718~S1*cat
  • Efficient Quadrature Rules for Illumination Integrals: From Quasi Monte Carlo to Bayesian Monte Carlo - Marques, Ricardo; Bouville, Christian; Santos, Luís Paulo; Bouatouch, Kadi, Morgan & Claypool Publishers, 2015. ISBN: 978-1627057691

Complementary:

  • Machine Learning and Rendering - Keller, Alexander; Křivánek, Jaroslav; Novák, Jan; Kaplanyan, Anton; Salvi, Marco, ACM SIGGRAPH 2018 Courses (SIGGRAPH '18) , 2018.

Addendum

Contingency plan

In case mixed teaching is required by the health situation (this is the expected model): * If the health situation allows it and the necessary conditions are met, we expect to have between 50% and 70% of in-person activities. In general, when having an occupancy rate of 50%, students will attend in-person for a week and will follow class on streaming for the following week. In case on-line teaching is required by the health situation: * The time ranges of mixed teaching are maintained but all teaching will be carried out in an online format, prioritizing synchronous sessions for the subject dynamization.