diegocasmo / iis-project Goto Github PK

Use outline as a starting point, feel free to modify it as you see fit.

Please add a few words about the pipeline folder in the main repo's README.md file, similar to the "Results" section

• Project's notebook is full of TODOs which explain the analysis we are doing, these must be explained

• Record a video showing a demo of the system developed during the project

• Parse Computer Vision input (see format)
• Export/import best model (either MLP or SVM)
• Use best model predict Computer Vision input

Notes:

• Kaylan mentioned in the meeting that we need to be able to tell what features are the most important ones and communicate that to the Computer Vision group. It would be a good idea to send an email to Kaylan to clarify this requirement.

Use outline as a starting point, feel free to modify it as you see fit.

• Create .tex folder structure
• Must include all sections that are mandatory (i.e. Elaborate on potential ethical concerns we have identified with the emotional responsive agent)

In addition to the tasks mentioned in the report's skeleton, the following must be covered:

• The report should include, at the beginning:
  • a summary of the work and main results
  • what has worked and what has not
  • role played by each individual group member
  • one paragraph explaining the contribution of each individual student should be included
• Others (mandatory as well):
  • Your original project specifications as an Appendix (the Appendix does not count towards the 5-page limit)

• Create format for Emotion Synthesis group (must include accuracy and confidence)
  • Update final-solution.ipynb notebook to include confusion matrix/accuracy
• Send email to Emotion Synthesis group with format and a documentation of it
• Classify a couple of examples and send them to the Emotion Synthesis (in the same email from above)

Use outline as a starting point, feel free to modify it as you see fit.

Others (mandatory):

• Reflection on how well the integration in the final pipeline worked (what worked well, what were challenges)

Use outline as a starting point, feel free to modify it as you see fit.

Use outline as a starting point, feel free to modify it as you see fit.

Others (mandatory):

• Discussion on possible ethical issues with the full pipeline or your specific part of the pipeline

• Update README.md with up-to-date instructions
• Verify .yml file is up-to-date
• Include a README.md file in the /code directory explaining the structure of it
• Clean /code directory (too many files, difficult to follow for a non-collaborator of the repo)
  • Place parsers in a /parsers dir
• Remove .csv from /data dir (update .gitignore)

Use outline as a starting point, feel free to modify it as you see fit.

• Implement MLP (Tristan)
• Implement Random Forest (Ankur)
• Implement K-fold validation (Ankur)
• Abstract methods into a single notebook for each of them (Tristan)