Fields of Application
- Internet of Things / Intelligent Sensors
- Medical Devices / Wearables
- Smart Home / Smart City
- Smart Factory / Condition Monitorig
Since AIfES is very versatile, we do not limit ourselves to special applications. We are currently working on the following research topics:
Machine learning algorithms and hardware accelerators:
RISC-V with neuroaccelerator
The Fraunhofer IMS is currently developing a RISC-V microprocessor, which will contain a hardware accelerator especially for neural networks. AIfES will be optimized in a special version for this hardware, to assure that the hardware resource can be used optimally.
AIfES especially suitable for sensors
In future, AIfES will contain functions for signal pre-processing in addition to the machine learning algorithms. These functions are particularly suitable for feature extraction.
Further network types
Since AIfES is constantly growing and new algorithms are being implemented, the next step is the integration of further neural network topologies. These include, for example, recurrent networks or a convolutional neural network (ConvNet).
Compatibility with other software applications
In the future, AIfES will also be able to import trained models from e.g. Tensorflow. Requirement for this is that the selected network structure is also available in AIfES.