SensiML web tool implements AI for IoT endpoints
April 21, 2020
Oregon-based SensiML, a developer of AI tools for building intelligent IoT endpoints, has released a web-based AutoML tool with a user interface that is said to make it easier for developers to implement AI for their endpoint IoT applications.
This tool is an evolution of the company's Analytics Studio, a core component of the SensiML Analytics Toolkit suite.
The Analytics Studio standalone web-based application is designed for users without data science expertise. Since it is a browser-based tool, there is no need for developers to download or install software locally. It features UI-driven AutoML capability with intuitive controls, allowing users to configure the software, obtain results, tune and iterate. New visualisation and reporting features provide clear and easy-to-understand feedback.
For data scientists and machine-earning experts, SensiML will continue to provide a separate version of the tool with a Python language interface. This version, called Analytics Studio Notebook, gives users programmatic access to the SensiML cloud and all its associated functionality for building compact ML sensor algorithms for consumer, industrial, automotive, commercial and agricultural IoT applications.
Both versions of Analytics Studio will continue to complement the Analytics Toolkit suite, which provides a complete design flow for developers to build intelligent endpoints from start to finish, including raw signal capture, data insight labelling, algorithm generation, firmware code generation, test, validation and support.
"The SensiML Analytics Toolkit was designed to simplify developing AI for IoT applications, making it possible for a single user or small development team to implement a complete AI-based solution for IoT endpoints," said Chris Rogers, SensiML CEO. "The new browser-based Analytics Studio improves the user experience through a sophisticated yet simple to use interface, so building intelligent sensor algorithms is easier than ever."
Analytics Studio filters and optimises labelled sensor data through machine-learning algorithms. It generates a device optimised SensiML Knowledge Pack (event detection algorithm), which is then ready to be installed in the local microcontroller's flash memory. The Knowledge Pack can detect events without reliance on cloud processing, creating an independent smart node.
Analytics Studio includes the following features and capabilities: AutoML code-generation engine; classic ML classifiers through xNN algorithms; automatic or directed feature engineering; web application interface; and Python client option for data science experts.
SensiML, a subsidiary of QuickLogic, offers software that enables low-power IoT endpoints that implement AI to transform raw sensor data into meaningful insight at the device itself. The SensiML Toolkit supports Arm Cortex-M class and higher microcontroller cores, Intel x86 instruction set processors, and heterogeneous core QuickLogic SoCs and QuickAI platforms with FPGA optimisations.