We misinterpreted the research paper and constructed the wrong models. Solution: Conduct regular tests and monitor the training progress in order to see if the model converges to an optimal solution
The training environment fails because of a power outage or some other unforeseen problem. Solution: make frequent checkpoints of the training model. That way we can safeguard from crashes
The model is stuck in a local minimum. Solution: Restart the model with randomly initialized parameters.
One or two of the team members is unable to work on the project because of an accident or a medical issue(e.g. COVID-19). Solution: The best solution is to get most of the project finished as soon as possible. If one of the team members is indisposed, then the other should probably take a bit more of the work
Possible risks:
We misinterpreted the research paper and constructed the wrong models. Solution: Conduct regular tests and monitor the training progress in order to see if the model converges to an optimal solution
The training environment fails because of a power outage or some other unforeseen problem. Solution: make frequent checkpoints of the training model. That way we can safeguard from crashes
The model is stuck in a local minimum. Solution: Restart the model with randomly initialized parameters.
One or two of the team members is unable to work on the project because of an accident or a medical issue(e.g. COVID-19). Solution: The best solution is to get most of the project finished as soon as possible. If one of the team members is indisposed, then the other should probably take a bit more of the work