Autonomous vehicles are one of the fastest-evolving fields in AI research today. Representing the future of driving in an AI world, a lot of investment is happening in this area.
Many data scientists have also cut their teeth on developing deep learning models for use in AVs, as it is a great starting point for creating models. Moreover, deep learning and machine learning technologies continue to increase in complexity and applicability as the field as a whole continues its growth.
Owing to this, Intel® has a community program known as the Software Innovator program that supports independent developers who demonstrate the ability to create forward-thinking projects. Intel®’s Student Ambassador program aims to find students with a developer affinity all over the world. As a part of this program, participants will gain access to new frameworks, technologies and hands-on resources for the creation of new models.
Intel® has partnered with MachineHack to launch a hackathon, open only for Intel® Student Ambassadors and those under the Software Innovator Program. The hackathon will be called “Making Autonomous Vehicles Safer For Humans”. By harnessing the power of the OpenVINO™ toolkit for deep learning models, participants will engage in creating a model for the effective recognition of objects on the road.
It will be open from April 22nd to June 15th.
OpenVINO™, or the Open Visual Inference and Neural network Optimization toolkit provides developers with improved convolutional neural network performance. Intel®’s distribution of the toolkit is based on their hardware specifications, thus bringing better performance across their hardware and maximising performance.
Intel®’s distribution also functions in conjunction with their accelerators due to the toolkit’s support for heterogeneous execution. This means that the CNN can utilise Intel® CPUs, GPUs, FPGAs, VPUs, IPUs and various compute sticks using a common API. The toolkit can be used to deploy computer vision, neural network inference, and deep learning capabilities.
About The Dataset
The dataset for this hackathon includes images from the Indian Driving Dataset, which was created as a part of a collaboration between IIIT Hyderabad and Intel®. The dataset aims to provide data that is specific to driving on Indian roads and conditions.
The training dataset will consist of 400 images, with the test set containing 100 images. The problem statement will be to predict the presence of a large automobile or a person or both in the test samples.
How To Participate In The Hackathon
- Open your favourite browser and go to https://www.machinehack.com/course/making-autonomous-vehicles-safer-for-humans-hackathon-by-Intel®/
- If you are a new user click on sign up and register on MachineHack with your email. Click the login button if you have already registered to MachineHack.
- After logging in, click on “Hackathons” which will take you to the page with a list of all the MachineHack hackathons. Select the latest hackathon from the list titled “Making Autonomous Vehicles Safer For Humans – Hackathon by Intel®”. Read through each of the highlighted sections of the Hackathon to understand the problem clearly.
- Click on “Start Hackathon” to enter the challenge, read through the rules and click on “Start Course” which will take you to the submission page.
- In the submission page, the training set, test set, and sample submission can be downloaded from the “Attachment” section and your submissions can be made by following the submit link in the “Assignment” section. Please note that after clicking on “Finish Hackathon” you cannot submit any more entries. Finish the hackathon only when you have no more submissions to make.
- The scores of your submission will be updated within five minutes after submission. To check the leaderboard after solution, click on the hackathon’s leaderboard.
MachineHack comes up with a number of interesting hackathons to give many challenging problems to data science enthusiasts. Here are all our recent hackathons. MachineHack recently concluded its 10th hackathon titled “Predict Restaurant Food Cost Hackathon”.
