Meeting Notes¶

6th of May 2019¶

• INFO (Hannes): Potential psychology student found: Hulda Sveindís (recommended by Berglind).
• INFO (Hannes): Has asked Paloma to come on Slack.
• INFO (Michelangelo): Is currently reading the papers suggested last week.
• INFO (Claudio): Now has a Trello Board up
• INFO (Claudio): Items, such as files (e.g. papers), can be "pinned" in Slack so we don't lose track of them.
• THIS WEEK (Now moved into Trello!):
  • TODO (Michelangelo, Hannes): Talk to Kári, Gylfi and Jacky about issues with the compute shaders. Who has been using them here?
  • TODO (Michelangelo): Look into the evaluation methods (see paper from last week and the Predicting perceived level of disturbance... and try finding more).
  • TODO (Michelangelo): Look at supporting multiple simultaneous spawners in the simulation.

29th of April 2019¶

• DEADLINE: Student Innovation Fund grant contracts to be signed by 14th of May, maybe start on June 1st? Focus on the validation through data gathering (see note from last meeting).
• INFO (Claudio): Has been improving the AI code a lot in the simulation.
• THIS WEEK:
  • TODO (Hannes): Follow up with Paloma.
  • TODO (Hannes): Locate potential NSN students.
  • TODO (Michelangelo): Start looking into the opportunities to publish a review paper on evaluation/validation methods in crowd/group simulations with a social behavior focus. Look at possible publication venues for such a review (e.g. AAMAS). An example of an evaluation paper is this one: A Statistical Similarity Measure for Aggregate Crowd Dynamics by Guy et al. in Transactions on Graphics
  • TODO (Claudio): Pick a "post-it" task board for the project.
  • TODO (Michelangelo): Get Compute Shader going.
  • TODO (Michelangelo): Look at supporting multiple simultaneous spawners in the simulation.
• KEEP EYE ON:
  • Ray Tracing in the new NVidia cards, new way to approach things. Great opportunities for simulation.
• EXTRA MEETING THIS WEEK
  • Thursday at 11 am.

15th of April 2019¶

• INFO (CP): Talked to the Technology Transfer Office, Einar and Susan, they were in love with the project idea and immediately presented this to Ferðamálastofa/Icelandic Tourist Board (this is connected to the Ministry)
• INFO (CP): Presentation about to be sent to Snohetta about the value of the simulation for talking to stakeholders.
• INFO (CP): Talked to Bilal (candiate for faculty position)
• IDEA (CP): Use a simple app, with one button to tap, for getting data on people's perception of crowdedness (they press when they feel it's crowded, more presses, more crowdy). Bilal supported this idea, and offered to provide an app to start with.
• IDEA (CP): From discussion with Bilal, instead of focusing on how correct the simulation is, with regards to real world data, one should focus on to what extent it diverges. Scatter plots with differences between pictures.
• DISCUSSION: How do we interpret the data, could be many levels of interpretation. Is it possible to associate sense of crowdedness with the space, considering time as well. Start with gathering information on densities which we can then use to plan further research/information gathering. Will people arrive in groups (e.g. in a tour group from a bus), always feel crowded? Maybe not. Maybe crowdedness is experienced to a greater degree when the crowd is coming from somewhere else, e.g. another group crashes with yours.
• DISCUSSION: Useful to refer to existing theories of Crowding, e.g. the Crowding Perception Model to understand e.g. what we need to know about the people reporting on crowding (e.g. former travel experience).
• DECISION: The summer project should be done by two students from two different departments, psychology and computer science.

8th of April 2019¶

• READ: Work from Chris Peter's Embodied Social Agents Lab at KTH (In particular on learning and groups).
• READ: Interesting work related to the evaluation of a crowd simulation, based on official guidelines. Question: Where do the guidelines get their model of human behavior from?
• READ: Good source of information on how to evaluate crowd/group behavior is work from INRIA Rennes
• QUESTION: How can we use Machine Learning (ML) to validate simulation from real world data from a specific location?
• INFO: We have a Student Innovation Fund grant for the summer, 2 students for 3 months. Project is focused on the validation of the simulation. The Application . We need to hire the students (they can be just graduated, both BS and MS level).
• GOAL (April/May): Set up the software and methodology for the students to execute during the summer.
• SUGGESTION (CP): We should keep an eye on the trends in AI in Europe, e.g. through reports like the WIPO Technology Trends.
• SUGGESTION (MD): May need to formalize the environment to a greater degree to have a better chance at applying AI technologies, e.g. game environments are very well defined.
• SUGGESTION (CP): For validation the AI (ML) may really just be focused on adjusting parameters of a model that has already been constructed. I.e. tweaking the simulation/knobs to produce results closer to the real data. Not so much to alter the underlying AI of the agents.
• SUGGESTION (HV): After we have some evaluation of how close we can get with the baseline model, we continue to improve the model, with territoriality and social models getting added (CP: Realistically, bit by bit).
• SUGGESTION: We should meet with the ML people in the department, such as Yngvi Björnsson and Stephan Schiffel
• IN PROGRESS (CP): Presentation for Snohetta (top architect in the world) regarding experience simulation for an exhibit.
• IN PROGRESS (CP): Contacted Paloma, but waiting for reply (perhaps follow up).
• IN PROGRESS (CP): Talking to the Icelandic Technology Transfer office about the paths to transfer.