UM-DeepOutbreak.yml

team_name: "University of Michigan, Computer Science and Engineering"
team_abbr: "UM"
model_name: "DeepOutbreak"
model_abbr: "DeepOutbreak"
model_version: "1.0"
model_contributors:
  [
    {
      "name": "Ruipu Li",
      "affiliation": "University of Michigan",
      "email": "liruipu@umich.edu",
    },
    {
      "name": "Sonika Potnis",
      "affiliation": "University of Michigan",
      "email": "potnissa@umich.edu",
    },
    {
      "name": "Alexander Rodríguez",
      "affiliation": "University of Michigan",
      "email": "alrodri@umich.edu",
      "orcid": "0000-0002-4313-9913",
    },
  ]
website_url: "https://alrodri.engin.umich.edu/"
license: "CC-BY-4.0"
designated_model: true
data_inputs: "Google's symptom search data, and CDC's ILI and WHO/NREVSS data."
methods: "Deep neural network model with conformal predictions."
methods_long: "Deep neural network model with conformal predictions. The neural network architecture is a sequence-to-sequence model based on recurrent units and self-attention modules. It is trained in a multi-task setting where each region is considered a task. The uncertainty quantification is conducted post hoc with conformal predictions that follows adaptive conformal inference to adapt to distribution shifts. Spatial correlation is not considered."
citation: "Rodriguez, A., Tabassum, A., Cui, J., Xie, J., Ho, J., Agarwal, P., Adhikari, B. and Prakash, B.A., 2021, May. Deepcovid: An operational deep learning-driven framework for explainable real-time covid-19 forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 35, No. 17, pp. 15393-15400). https://ojs.aaai.org/index.php/AAAI/article/view/17808"
ensemble_of_models: false
ensemble_of_hub_models: false
team_funding: "Start-up funds from the University of Michigan."