Artificial Intelligence model and system development is still much more of an art than an exact science, requiring developers to interact with model code, training data, and other parameters over multiple iterations. Training datasets may be acquired from unknown, untrusted sources. Model weights and other training parameters can be susceptible to malicious tampering.
Format: |
|
Topics: | |
Website: | Visit Publisher Website |
Publisher: | National Institute of Standards and Technology (NIST) |
Published: | July 1, 2024 |
License: | Public Domain |