R-Ladies Mini Hackathon: Building a deep learning powered application in R

June 16, 2019 devadvin

Two months ago, we at R-Ladies San Francisco had this dream of bringing together people who do not have a deep learning background and helping them create a deep learning powered-application in a few hours.

Our dreams were fulfilled, and we had our first R-Ladies San Francisco mini-hackathon a few days ago. A hackathon can be the perfect place for inquisitive minds to meet and great ideas to be born.

Goal

The Model Asset eXchange is an Open Source Initiative from IBM that focuses on making deep learning models easy to consume. The idea at the hackathon was to use a model from the Model Asset Exchange and create an application around it.

The day

We kicked-started the hackathon with an introduction about R-Ladies, the Model Asset Exchange, a demo, and, of course, coffee!
The participants were given some time to understand the concepts of the Model Asset Exchange and how it works. Meanwhile, they formed teams and started brainstorming ideas.

Hackathon collage

The projects

  1. Deep learning in retail optimization

    Identify age groups of people entering a retail store at different times of the day. This allows retailers to better cater to their audiences at different times of the day, such as employing workers with different ages or credentials to work different shifts strategically. Or, possibly moving merchandise from their back-stock to their front shelves at different points in time.

  2. Classifying chromosomes of new organisms using Karyotyping images using Mask-RCNN

    Deep learning in the medical field is becoming more popular. Using Mask-RCNN to detect chromosomes and also performing pixel-wise instance segmentation to extract them can be a game changer as it eliminates a lot of manual work in classifying and extracting details from the images.

  3. Sentiment analysis using a text sentiment classifier

    Use Twitter data to perform sentiment analysis on trending tags and also create a Shiny app around it. This is all done using a state-of-the-art BERT model.

  4. Generate reviews using review generator

    How often you have skipped reviewing a product just because you don’t want to type it? Here, a review generator app generates reviews based on keywords.

  5. App for detecting objects using object detector

    A Shiny web app for detecting objects in an image trained using a COCO data set.

Presentations

The Model Asset Exchange is deep learning framework-agnostic, and we went one step beyond this to demonstrate it as a programming language-agnostic platform with these contributions.

If you are interested in learning more, make sure to look at our code patterns, tutorials, and our webpage.

Do you want to contribute?

Contact us through our slack channel, and follow us on Medium and Twitter for more information.

Previous Article
An introduction to the internals of the Model Asset eXchange
An introduction to the internals of the Model Asset eXchange

Take a look at how the Model Asset eXchange works.

Next Article
Ensuring equal human rights for vulnerable populations
Ensuring equal human rights for vulnerable populations

Konstantinos Sofikitis talks about his experience at the inaugural event, Call for Code Geneva.

×

Want our latest news? Subscribe to our blog!

Last Name
First Name
Thank you!
Error - something went wrong!