Technologies Used
JavaScriptPandas
NumPy
scikit-learn
TensorFlow
bootstrap
Beginner Html/Css
keras
Flask-Restful
a bit of python
Project Team
About
Machine Learning practitioners are often are often faced with decision making challenges on arriving at a best-fit algorithm for the problem at hand. eMiLy, our virtual assistant for any ML enthusiast, beginner or seasoned expert will make your decisions easy by analyzing the dataset. This saves GPU, computation power and time that is usually exhausted while finding the best solution. eMiLy - the ML decision making as a service is available through a web interface. The algorithm is based on recognizing unique patterns from datasets and label them with performance metrics of pre-run machine learning fully tuned models. We have used CNN architecture along with the ML model labels as attention parameters to achieve this pattern recognition from datasets and predict the different algorithm's efficiency if this data were to run on them. This would then predict the performances of models without actually running them. The future application to this could be to apply the same techniques for checking efficiency on deep and very deep CNN architectures which usually require high powered GPUs and still consume a lot of time to check if they would be a really good fit for the data. So, this service helps the ML practitioner to save resources by letting the machine learn how it learns the best! :-)
Comments
Praveen svsrk - 1940d
Ritesh Modi ✔ - 1942d
Jim - 149d