APS8
Train deep nets powered
by the world’s smallest
inference engine
No AI expertise required

USE CASES

Machine health monitoring

Fall-predicting wearable for seniors

Running form indicator

In-body drug delivery device

Remote environmental monitoring

Neural net accelerator

Malware detection

CASE STUDIES

Machine health monitoring for the purpose of predicting the failure of air pressure systems in heavy trucks

Inputs
Outputs
Code
Data
RAM
Accuracy
Recall
Power

170
1
0.1 kB
131 kB
1.9 kB
0.99
0.68
93 µW

HAR7
a
i

Convert IoT data into insights & actions
EDG
XL

Ultra-low power edge inference
CLD
XL

High-throughput cloud inference

USE CASES

Machine health monitoring

Fall-predicting wearable for seniors

Running form indicator

In-body drug delivery device

Remote environmental monitoring

Neural net accelerator

Malware detection

CASE STUDIES

Human activity recognition
based on triaxial acceleration and triaxial angular velocity readings

Inputs
Outputs
Code
Data
RAM
Accuracy
Recall
Power

561
6
0.1 kB
2 kB
0.4 kB
0.95
0.95
1 µW

Turboprop8
Low-power, compact, fast
inexpensive deep nets for
the IoT edge & wearables
No AI expertise required

USE CASES

Machine health monitoring

Fall-predicting wearable for seniors

Running form indicator

In-body drug delivery device

Remote environmental monitoring

Neural net accelerator

Malware detection

CASE STUDIES

Predictive maintenance of aircraft engines based on temp, pressure, RPM, flow, and air-ratio sensor data

Inputs
Outputs
Code
Data
RAM
Accuracy
Recall
Power

35
3
0.1 kB
8.1 kB
0.3 kB
0.95
0.95
5 µW

PlayPause