The models are created so that these can be directly used to automate some common processes that can be used by many apps for various use cases.
What it does
The projects contain three models:
- Text summarization model can be used by various apps to automate the task of summarizing huge text into a few words. This is something that can be easily plugged into various apps for various use cases. Customer support can be one of the use cases of the apps that might need this.
- Spam filtering module can filter if the content is a spam or not. This too can be used as a module by various apps to automate things.
- The third model can be used to identify different verbal instructions. The data used for this is obtained from the Speech Recognition challenge on Kaggle. This model can be used by apps that want to use verbal cues in order to control navigations in the app. For example, instructions like stop, on, off, etc can all be intelligently used by apps to enable navigation in the applications as one of the use cases.
How we built it
We used TensorFlow, Python and NLTK.
Accomplishments that we're proud of
All the models have good accuracy and can enhance further if more diverse data is present.
What's next for Machine learning models
To improve the model accuracy by collecting customer satisfaction reviews and the data generated by the customers after integrating with PowerUp may even be used for further training the model using that data for better performance.