Function
Ref. Model
I/O Data
SubAIMs
JSON MData
Profiles
Ref. Software
Conformance
Performance
1 Functions
The During-Training Watermark Insertion (NNW-DWI) AIM provides the Watermarked model in response to a given Training dataset and Payload: the Payload is inserted while the neural network is trained with the Training dataset. Where a Locator is given, the Payload is inserted where the Locator says, by the method it names; the AIM does not give the Locator back. Without a Locator, the method decides. Where the watermark cannot be placed there, the AIM gives no model and reports why.
2 Reference Model
3 Input/Output Data
Table 1 specifies the Input and Output Data of the During-Training Watermark Insertion (NNW-DWI) AIM.Table 1 – I/O Data of the During-Training Watermark Insertion (NNW-DWI)
| Input | Acronyms | Descriptions |
| Untrained model | AIP-MLO | The initial model of the neural network to be trained. Optional. |
| Training dataset | NNW-DST | The dataset with which the neural network is trained. |
| Payload | NNW-PLD | The information to be inserted in the neural network. |
| Locator | NNW-WLC | Where the watermark is to be placed in the model and the method that inserts it. Optional: without it, the method decides. The AIM does not give it back. |
| Output | Acronyms | Descriptions |
| Watermarked model | AIP-MLO | The trained neural network model with the Payload inserted. |
4 SubAIMs
No SubAIMs.
5 JSON Metadata
https://schemas.mpai.community/NNW1/V1.2/AIMs/DuringTrainingWatermarkInsertion.json
6 Profiles
No profiles.
