MODSIM2023 banner

Workshops

Workshops will be held on Friday 14 July. Please contact the workshop organiser for more information. Venues and exact times are still being confirmed and will be updated as soon as possible.

 


Bayesian ideas in pictures

Time: 9:00 am to 3:00 pm
Location: TBA
Organisers: Samantha Low-Choy and Daniela Vasco
Contact: s.low-choy@griffith.edu.au

This workshop has three parts, leading you through the ideas of Bayesian statistics, relying heavily on pictures. At first the pictures are used to illustrate core concepts of probability -- conditional probability, priors, and posteriors. This provides a foundation to understand the conceptual diagrams that can be used to communicate Bayesian models, as an alternative to mathematical equations.

1. Bayesian for Babies. Following Chris Ferrie's boardbook (ostensibly for parents to read to small children), we introduce the ideas behind Bayes' Theorem using cookies and candy (or biccies with smarties, in the local slang).

2. Bayesian for Toddlers. The baby steps of Bayes Theorem (part 1) are increased to toddler-sized steps, explaining how Bayes' theorem underlies Bayesian statistical modelling, which updates prior information to posterior information.

3. Bayesian via diagrams. Dynamic acyclic graphs (affectionately known as DAGs) provide the foundation for several statistical modelling frameworks. We show how to interpret models using DAGs, and understand the conceptual differences between different model choices. In addition, we consider how DAGs are used to represent uncertainty in different ways: for Bayesian networks (or Conditional Probability Networks) analysed using deterministic methods; or for Bayesian statistical models, analysed using Bayesian (prior-posterior) inference.

The workshop will include activities to help participants to discuss and explore concepts. It will suit beginners who are new to these ideas. It may also suit researchers who have used Bayesian statistical models or Bayesian networks before, but wish to deepen their conceptual understandings and ability to communicate complex ideas via pictures and diagrams.


How to build a spatial causal network

Time: 9:00 am to 12:00 pm
Location: TBA
Organisers: Kate Holland, Luk Peeters and Cameron Huddlestone-Holmes
Contact: kate.holland@csiro.au

The Geological and Bioregional Assessment (GBA) Program developed a new methodology based on causal networks (CN) to assess potential impacts due to resource development on water and the environment at a regional scale. The method provides a transparent and consistent way to understand and evaluate impacts, and outlines priorities for management, mitigation, and monitoring requirements. We published the methodology as a journal paper (Science of the Total Environment 802. Doi: 10.1016/j.scitotenv.2021.149845) and as a short technical report available from the GBA website (https://www.bioregionalassessments.gov.au/gba/introduction-causal-networks).

GBA Explorer (https://gba-explorer.bioregionalassessments.gov.au/) is an online interactive graphical presentation of the spatial causal networks that provides immediate access to node descriptions, link evaluations and an overall assessment summary. Users can visualise the entire causal network or simplify it by selecting specific pathways. Spatial data are also presented via interactive maps that include the spatial information used to inform the assessment and spatial impact maps.

This interactive workshop will be presented by the CN authors Kate Holland, Luk Peeters and Cameron Huddlestone-Holmes from CSIRO. After briefly reviewing the main points in the CN methodology, they will demonstrate the various tools and methods, illustrated by worked examples to show how to interpret results and acknowledge constraints.


The En-ROADS Interactive Climate Simulation Game

Time: 9:00 am to 1:00 pm
Location: TBA
Organisers: Emiliya Suprun, Oz Sahin and Russell Richards
Contact: e.suprun@griffith.edu.au

The Climate Action Simulation is a highly interactive, role-playing game that engages a wide range of participants in exploring key technology and policy solutions for addressing climate change. It uses the cutting-edge simulation model En-ROADS, created by Climate Interactive and MIT Sloan. Participants will experience what it’s like to negotiate a climate deal to address one of the greatest human and environmental challenges of this century.

Format: The game is conducted as a simulated emergency summit organized by the United Nations that convenes global stakeholders to establish a concrete plan that limits warming to Paris Agreement goals. Participants propose climate solutions such as energy efficiency, carbon pricing, fossil fuel taxes, reducing deforestation, and carbon dioxide removal.

Target audience: policymakers, educators, businesses, and the public willing to explore, for themselves, the likely consequences of energy, economic growth, land use, and other policies and uncertainties, with the goal of improving their understanding.


Modular Assessment of Rainfall-Runoff Models Toolbox (MARRMoT) tutorial

Time: 9:00 am to 12:00 pm
Location: TBA
Organisers: Luca Trotter
Contact: l.trotter@unimelb.edu.au

The Modular Assessment of Rainfall-Runoff Models Toolbox (MARRMoT) is a flexible modelling framework designed to reproduce the behaviour of 47 published rainfall-runoff models in a consistent way. The tutorial will familiarise participants with the use of MARRMoT and will include a general introduction to the toolbox as well as practical application examples covering:

(1) formatting of input data
(2) how to run simulations with MARRMoT
(3) model calibration
(4) how to edit MARRMoT models.

Depending on the number of participants, we will allocate time to specific questions or applications.

Participants are expected to have their own computers for the workshop, running a licensed version of MATLAB.


Navigating a Career in Science: A professional development workshop for Early Career Researchers and Young Professionals

Time: 9:30 am to 12:30 pm
Location: TBA
Organisers: Anna Lintern, Danlu Guo, Fiona Tang, Pallavi Goswami, Clare Stephens and Nevenka Bulovic
Contact: anna.lintern@monash.edu
Danlu.guo@anu.edu.au
n.bulovic@uq.edu.au

This is a professional development workshop for Early Career Researchers and Young Professionals starting out their career in academia, industry or government.

The workshop has two parts.
 - Part 1 comprises of several presentations from senior researchers and professionals from diverse backgrounds, who will discuss their career journeys, key decisions they have made, challenges, and how they have overcome them.
 - Part 2 is an open discussion where participants can discuss issues in small groups with the senior researchers and professionals.


Simulation-based Optimization: A tutorial with Python

Time: 10:00 am to 1:00 pm
Location: TBA
Organisers: Hasan Turan and Sanath Kahagalage
Contact: h.turan@adfa.edu.au
Registration: https://forms.gle/UtYNS4sNvrFHN5j89

Tutorial Overview

Real-world systems and decision problems are often too complex to model and rarely analytically tractable to solve. Simulation modeling is a prevailing technique to alleviate these difficulties by capturing intricate relationships and uncertainties associated with such systems. With the recent advances in computational resource power, the use of simulation-based optimization algorithms for tackling decision problems (especially when closed-form analytical models provide poor estimations or do not exist at all) has been drastically increased.

In this tutorial, we will demonstrate how to implement and couple simulation models and optimization algorithms in Python programming language to provide hands-on experience for participants.

Tutorial Outline

  • A fast and furious introduction to Python and Jupyter notebook
  • A simple simulation model development in Python
  • Analytical versus simulation-based approaches
  • Developing and coupling optimization algorithms with simulation models

Who should attend

This introductory level tutorial is for PhD and Master students or Early Career Researchers who are new to simulation-based optimization and interested in using Python for their research. Even though the content in this course can be applied to a range of decision problems, the presented case study during the tutorial will be on supply chain management, particularly inventory optimization problems.

Pre-requests and required software

Some familiarity with Monte Carlo simulation modelling and optimization algorithms in particular meta-heuristics is required.

  • Python (Anaconda distribution)
  • Python libraries including Matplotlib, Seaborn, Numpy, etc.
  • Source codes will be shared over GitHub during the tutorial with registered attendees