RFQ 140G0120Q0090 Q-A_0001.pdf

RFQ 140G0120Q0090 – Q&A

Question: How many total images and also size (GB in TB) will be used for developing Artificial Intelligence for Streaming flow. So that we need to procure the required capacity of the Virtual Machines for developing this solution?

Answer: Approx 20 GB

Question: We have proprietary web solutions for storing the image data and processing the data.

Whether do you encourage this solution to be used or only open source technologies to be used for developing this solution?

Answer: Open source only

Question: Whether this project required us to publish or present research papers as part of these deliverables?

Answer: No

Question: What are the deep learning criteria that need to be used with the geographical data? What is the size of these datasets?

Answer: A regional database of images will be available (~20GB). The applicant will identify the most effective deep learning model structure.

Question: What are the deep learning criteria that need to be used with the local data?

Answer: The applicant will identify the most effective deep learning model structure.

Question: Most of the images are taken from the Conte Anadromous Fish Lab in Turner Falls, Massachusetts but how this data is extrapolated to individual local model versus a spatially weighted ensemble model across the entire study region?

Answer: Initial trials were run with existing data from the Conte Lab. This project will use the regional image database.

Question: Do we need to use the authentical model (UsedId protected) for the web application?

Answer: Yes

Question: Do we need to build the cloud environment for the demo of the application to the water resource managers and stakeholders.?

Answer: No, that will exist within Amazon CHS.