Systems, methods, and apparatuses for determining data relevance and labeling, model updating, and deployment
Techniques for training a machine learning model based on captured images are described. A method described include filtering a first set of collected images using one or more machine learning models; labeling the first set of filtered, collected images using a data labeling service using a service of the provider network; training a machine learning model from a machine learning algorithm using the first set of filtered, collected images using a service of the provider network; and causing deployment of the trained machine learning model onto a device.
1. A computer-implemented method comprising:
collecting images using a camera to generate collected images;
filtering the collected images using one or more machine learning models to generate filtered, collected images;
transmitting the filtered, collected images to a provider network;
labeling the filtered, collected images using a data labeling service of the provider network;
training a machine learning model using the filtered, collected images using a model development service of the provider network to generate a trained model;
optimizing the trained model for a target device using the model development service of the provider network to generate a trained, optimized model; and
causing deployment of the trained, optimized model onto the target device.
2. The computer-implemented method of claim 1 , further comprising:
suggesting a machine learning algorithm to train into the machine learning model based upon a desired outcome and the filtered, collected images.
3. The computer-implemented method of claim 1 , wherein the trained, optimized model is deployed outside of the provider network.
4. A computer-implemented method comprising:
filtering a first set of collected images using one or more machine learning models to generate a first set of filtered, collected images;
labeling the first set of filtered, collected images using a data labeling service of a provider network to generate a first set of labeled, filtered, collected images;
training a machine learning model from a machine learning algorithm using the first set of labeled, filtered, collected images using a model development service of the provider network to generate a trained machine learning model; and
causing deployment of the trained machine learning model onto a device.
5. The computer-implemented method of claim 4 , wherein the filtering of collected images is performed using a service of a provider network.
6. The computer-implemented method of claim 4 , wherein the filtering of collected images is performed outside of the provider network under control of a service of the provider network.
7. The computer-implemented method of claim 4 , wherein the trained machine learning model is deployed external to the provider network.
8. The computer-implemented method of claim 4 , wherein the trained machine learning machine learning model is deployed in the provider network.
9. The computer-implemented method of claim 4 , wherein the machine learning algorithm is an image classification algorithm.
10. The computer-implemented method of claim 4 , wherein the machine learning algorithm is an object detection algorithm.
11. The computer-implemented method of claim 4 , further comprising:
optimizing the trained machine learning model for a particular device configuration prior to deployment.
12. The computer-implemented method of claim 4 , further comprising:
receiving one or more parameters to configure one or more services to perform the labeling the first set of filtered, collected images using the data labeling service, the training a machine learning model from a machine learning algorithm using the first set of filtered, collected images, and the causing deployment of the trained model onto the target device; and
configuring the one or more provider network services based on the received parameters.
13. The computer-implemented method of claim 4 , further comprising:
suggesting the machine learning algorithm to be trained based on a desired outcome and the filtered, collected images.
14. The computer-implemented method of claim 4 , further comprising:
filtering a second, different set of collected images using the one or more machine learning models;
labeling the second, different set of filtered, collected images using the data labeling service; and
re-training the machine learning model using the second, different set of filtered, collected images.
15. A system comprising:
an image capture device to capture a plurality of images; and
a model development service implemented by one or more electronic devices, the model development service including instructions that upon execution cause the model development service to invoke one or more other services to:
filter, using one or more machine learning models, a first set of collected images provided by the image capture device to generate a first set of filtered, collected images,
label the first set of filtered, collected images to generate a first set of labeled, filtered, collected images,
train a machine learning model from a machine learning algorithm using the first set of labeled, filtered, collected images to generate a trained machine learning model, and
cause deployment of the trained machine learning model onto a device.
16. The system of claim 15 , wherein the trained machine learning model is deployed external to the provider network.
17. The system of claim 15 , wherein the machine learning algorithm is one of an image classification algorithm and an object detection algorithm.
18. The system of claim 15 , wherein the model development service is further to invoke a service to optimize the trained machine learning model for the device.
19. The system of claim 15 , wherein the model development service is to be configured based upon received parameters for the services it is to invoke.
20. The system of claim 15 , wherein the model development service is to re-train the machine learning model when the trained machine learning model does not have a desired accuracy.