Tech questions ITB-23-1317-JC.pdf
PDF 41 KB Posted
- Attached to
- Technical Questions and Answers
- Solicitation number
- IN-02552
- Issued by
- Polk County, Florida
About this file
This document contains technical questions and answers related to the FDOT District One City of Sarasota Connected Vehicle Project ITB-23-1317-JC. The project is primarily intended to be a functional system upon installation, with a research element. While there are no defined accuracy targets for the AI classification elements, the expectation is that the AI will learn from a "tabula rasa" or blank slate starting upon installation. The project allows for the application of learning from previous data collection efforts to improve the accuracy and speed of the AI detection.
The document indicates that the project is acceptable to apply learning from previous data collection efforts that improve the accuracy of the AI detection and the speed with which it achieves good accuracy. The project is focused on the deployment of a connected vehicle system and does not provide specific details on pricing, set asides, incumbents, funding requirements, or other contractual terms.
View the file
On GovTribe
Work with this file
- Download the original file
- Contacts named in this file
- Similar government files
- Ask GovTribe AI about this file
Text version
DOT-ITB-23-1317-JC
FDOT District One City of Sarasota Connected Vehicle Project
Technical Questions
1) Is the system described expected to be turn-key and operational on day one of install, or is this project intended as an R&D project to be evaluated throughout development and at the end of the 3 year R&D cycle?
A) While there is a research element of this project, it’s primarily intended to be a functional system upon installation.
2) System requirements #3 details the AI requirements of this project. If known, what is the expected AI’s accuracy for correctly classifying events (such as jay walking etc.). If known, what are the acceptable limits on the AI incorrectly classifying events?
A) Presently, there are no accuracy targets for classification elements.
3) As this is an AI project, it is expected that classification reporting accuracy would improve the longer the AI system is permitted to collect data and learn from the environment. Does this project expect the vendor to “bring our own brain” from a previous data collection engagement, or is the department expecting “tabula rasa?
A) Our expectation is that the AI will learn from a “tabula rasa” / blank slate starting upon installation. It is acceptable to apply learning from previous data collection efforts that improve the accuracy of the AI detection and the speed with which it achieves good accuracy.
File details come from the government source that posted it. Updated .