The BELGRADE MACHINE LEARNING HACKATHON will be co-hosted by Startit and Avast, and will take place on June 1st & 2nd in the Belgrade Startit Center (Savska 5).
It will last for 24 hours, with the top three teams receiving prizes of 2500€, 1500€ and 1000€.
The hackathon will be a Kaggle-style machine learning competition. Participants will receive a training dataset provided by Avast and will be asked to analyze the data and solve a classification problem using machine learning techniques. A test dataset will be used for evaluating the performance of the solutions and determining the winners.
The contest is open to both industry and academia.
In this competition, you have a chance to work with the data from an exciting domain – cybersecurity.
Connected and smart devices are becoming increasingly popular in modern homes: smart TVs, security cameras, home assistants, smart door locks, connected baby monitors… It is important to see what devices you have connected to your network: is that device my smart bulb or somebody hacked into my network? And in order to better protect connected devices, it is also extremely beneficial to know device types. Smart TVs and smartphones, security cameras and network storages – they all might have different vulnerabilities and require different protection.
Avast antivirus products include Wi-Fi inspector – a component that helps users secure IoT devices and other computers on their home networks. Wi-Fi inspector scans the user’s network and provides an inventory of connected devices. It identifies the class of a device whether it is a smartphone, TV, a camera or something else. This information helps Wi-Fi inspector to ensure that the device is protected.
In this competition, you will be analyzing raw data collected by Wi-Fi inspector for thousands of different IoT and non-IoT devices. The goal is to design your own classifier for device class prediction from network data. Use benefits of data exploration, feature engineering and machine learning techniques of your choice.
By playing with the dataset, you will learn how devices communicate with each other in a network, how different are security cameras from smartphones and what makes a smart lighting bulb so special.
A small subset will be released shortly after the registration is open; the full dataset will be available on the day of the competition.
*It is beneficial to have some background on computer networks, but we believe you will be able to quickly learn all necessary concepts during the hackathon.
Since all participants will be using sensitive data during the hackathon, they will be asked to sign a Non-disclosure agreement before the event.
Od 11:30 — Registration
12:00 — Presentation of the dataset and the task
15:00 — Lunch and consultations with mentors
20:00 — Dinner and consultations with mentors
00:00 — Snacks
10:00 — Breakfast
11:00 — Presentations of the solutions
12:30 — Belgrade AI & Security Meetup
13:30 — Announcement of winners, celebration and networking
Applications will be open until May 23.
The selection process will begin on May 24. Participants who qualified for the hackathon will be notified on May 27.