Piccolo AI is a powerful AutoML modeling tool for creating predictive algorithms that transform raw sensor data into valuable application-specific insight. However, the model accuracy from the resulting firmware library generated is only as good as the data that was used to train it. Thus data cleansing is an important preparatory step in ensuring outcomes from the ML model building development phase. For this task SensiML offers its Data Studio application as a complementary companion tool to ensure high quality training data is achievable from the outset. Data Studio is a powerful ML DataOps client/server application that allows single users up to large development teams to:
- Collect and import raw sensor data from offline, connected and/or live-streaming sensor hardware
- Annotate datasets with user-defined metadata and label data
- Record and synchronize video to assist in after-the-fact and offline analysis and labeling workflows
- Manage and version control dataset labeling with multi-session labels with means for annotation auditing
- Augment empirically collected sensor data with synthetic data from Generative AI and DSP manipulated extensions to existing data
Data Studio is separate from Piccolo AI and offered as a paid subscription-based tool. For more info, see https://sensiml.com