search

pagand / model_optimze_vessel

Queen of Oak Bay - visualization/classification/prediction
5 stars 4 forks source link

Sprint #16 (August 31) #36

Open pagand opened 10 months ago

pagand commented 10 months ago
  1. Model: if need to train again, consider removing the random selection for start point (always start from beginning)
  2. Model in environment: do not need the loops over horizon as the factual data is not available. Only use the first prediction, compare with the actual label in a selected trip.
  3. Conclude the project by one week extension: 3.1. Successful, so we can send the student proposal paper (deadline sep 15) to AAAI [need to finish writing the draft by sep 10th]. Then we can ask for extension of this work for next term for a full version paper. 3.2. Unsuccessful, wrap up the reports and code done throughout the summer term. Then change your focus to implementation side of the project and work closely with Jack.
pagand commented 10 months ago

in close area to the docking point, we need to consider only mode 2.