A Naive Plan for the Coming Experiment at AMNH

[linan7788626]

• [x] Construct a Specific Webpage for AMNH

  • [x] Create a web page as CLEAN as possible .
  • [x] Add tooltips to explain the buttons.
  • [x] Instruct students at AMNH to use Hoopla.
  • [x] Ask trained students to instrcut public visitors to use Hoopla.
  • [x] Collect models from users: locally or remotely?

• Agents: Three groups of users

  • Strong lensing Experts
  • Students on the first layer
  • Visitors on the second layer

• Properties of the agents in different groups:

  • Statistical properties of the accuracy of the models from agents in different groups: bias and scatter
  • Statistical properties of the efficiency of modeling procedures of the agents in various groups: mean, scatter

• Images would appear in the experiment:

  • [x] For instruction:
    • Some famous strong lensing systems, SDP.8.1, HorseShoe.......
  • [x] For testing:
    • bsz = 10" , 400x400
    • HST-like images using SIE: Re = [2, 3], ql = 0.73, Ori = 39 degree
    • HST-like images using SIE+ExtShear+ExtKappa: (0.03, 36), 0.02
  • [x] For prediction:
    • LSST-like images using SIE
    • LSST-like images using SIE+ExtShear+ExtKappa

[linan7788626]

Simulate properties of the agents

• JSON files with lens models and source models
  • LENS: Position, Einstein Radius, Ellipticity, Orientation
  • SOURCE: Position, Sigma, Ellipticity, Orientation
• Name each JSON file as __.JSON

Construct the first version of HazelTree

• CLASS human with layer_ID, user_ID, Bias, Scatter, Speed ......
• More coming soon ...