Back2Future: Leveraging Backfill Dynamics for Improving Real-time Predictions in Future

Authors: Harshavardhan Kamarthi, Alexander Rodríguez, B. Aditya Prakash

ICLR 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our extensive experiments demonstrate that our method refines the performance of diverse set of top models for COVID-19 forecasting and GDP growth forecasting.
Researcher Affiliation Academia Harshavardhan Kamarthi, Alexander Rodríguez, B. Aditya Prakash College of Computing Georgia Institute of Technology {harsha.pk,arodriguezc,badityap}@gatech.edu
Pseudocode No The paper describes the model components and their interactions but does not include any formal pseudocode or algorithm blocks.
Open Source Code Yes We also release the code and datasets at www.github.com/Aditya Lab/Back2Future.
Open Datasets Yes We collected and pre-processed important publicly available signals from a variety of trusted sources that are relevant to COVID-19 forecasting to form the COVID-19 Surveillance Dataset (Co VDS)... The code for B2F and the Co VDS dataset is publicly available at https://github.com/Aditya Lab/Back2Future.
Dataset Splits Yes We tuned the model hyperparameters using data from June 2020 to Aug. 2020 and tested it on the rest of dataset including completely unseen data from Jan. 2021 to June 2021.
Hardware Specification Yes All experiments were run in an Intel i7 4.8 GHz CPU with Nvidia Tesla A4 GPU.
Software Dependencies No The paper does not provide specific software names with version numbers for libraries or frameworks used in the experiments.
Experiment Setup Yes We tuned the model hyperparameters using data from June 2020 to Aug. 2020... We observed that setting hyperparameter τ = c|F| where c {2, 3, 4, 5} provided best results... We also found setting l = 5 provided the best performance.