IndiaAI, an Independent Business Division (IBD) under the Digital India Corporation (DIC) of the Ministry of Electronics and IT (MeitY), is spearheading the IndiaAI Mission. This mission seeks to democratize AI's benefits across all sectors, bolster India’s global AI leadership, promote technological self-reliance, and ensure AI's ethical and responsible use. As part of this effort, the IndiaAI Application Development Initiative (IADI) is driving innovation by promoting AI applications with the potential to catalyze socio-economic transformation in key sectors.

In line with these goals, IndiaAI is collaborating with the Indian Cybercrime Coordination Centre (I4C), a Ministry of Home Affairs initiative that supports law enforcement agencies in tackling cybercrime. The I4C manages the National Cybercrime Report Portal (NCRP), which receives approximately 6,000 cybercrime cases daily.

IndiaAI and I4C are launching the IndiaAI CyberGuard AI Hackathon to address the increasing threat of cybercrime. The hackathon is a strategic effort to leverage cutting-edge AI technologies to improve cybersecurity. It allows innovators, researchers, and entrepreneurs to design advanced AI models to enhance cyber fraud detection, investigation, and mitigation.

The Hackathon will challenge participants to develop sophisticated AI solutions, focusing on Natural Language Processing (NLP) models. These models will analyze, classify, and improve the accuracy of citizen-reported cybercrime incidents on the NCRP platform.

Event Overview:

• Stage 1: A virtual competition where teams work on an NLP model to classify fraud incidents, running over 20 days.
• Stage 2: The top 20 teams will participate in an in-person hackathon in New Delhi.
• Stage 3: The top 3 teams will pilot and deploy AI-driven cybersecurity solutions.

Key Objectives:

• Develop NLP models that can classify fraud based on victim descriptions.
• Enhance the accuracy of reporting and analysis on the NCRP.
• Contribute to national cybersecurity by addressing the rising complexity of cybercrime.

Prizes:

First Prize: INR 25 lakhs and the opportunity for national solution deployment.

Special prize of INR 5 lakhs for all-women teams.

Guide to the Hackathon Participation

I) Stage 1:

a) Participation requirement:

• Participants may participate individually or in groups per the below eligibility criteria.

b) Event Process and Schedule

• All participants are required to Sign up here before filling the application form to access the application form.

• A team leader must register and apply for the challenge individually on the IndiaAI portal by clicking the submit link.

• After the initial sign-up, the Team Leader should list all the Team Members under the Management Team and complete all organization details.

• Team Leaders will also create a GitHub/Open Forge account to upload Source Code. Next, the Team Lead shall paste the link for their

GitHub/Open Forge source code in the application form.

• Participants may access the Datasets and the Data Dictionary upon registration.

• Additionally, the Team Leader must answer all additional questions, including uploading documents and clicking ‘Submit’.

• The Hackathon under Stage 1 would take place virtually over 20 days from the start of registration and the final date for submission of developed prototypes. Any edits to the Source Code after the final submission date will lead to immediate disqualification of the application.

c) Problem Statement:

• Problem Statement: Development of an NLP Model for Text Analytics and Classification

• Objective:

To develop an NLP model that categorizes complaints based on victim, type of fraud and other relevant parameters used for text classification and preparing the final model.

• Outputs:

• Text Preprocessing: Tokenization, stop word removal, stemming, and text cleaning to prepare data.
• Model Development: Selection of a suitable NLP model for text classification
• Accuracy Measurement: Evaluate the model based on accuracy, precision, recall, and F1-score metrics.

II. Stage 2:

a) Participation requirement:

• Top 20 Solutions will be selected for an in-person round.

• Three members per team will be provided with travel and accommodation support to participate.

b) Event Process and Schedule:

• The second round will be organised as an in-person 3-day event in New Delhi.

Details on the location and logistics will be emailed to the shortlisted participants.

• Participants would receive a dataset from the NCRP database.

• Participants need to work on the given problem statement detailed below.

c) Problem Statement:

Problem Statement: Development of an NLP Model to guide citizens in filing cybercrime reports on the National Cyber Crime Reporting Portal (NCRP) correctly through a real-time analysis of the description and incident-supporting media files uploaded by the citizen.

• Objective:

To develop an NLP model that categorizes fraud based on victim descriptions, focusing on improving model accuracy.

• Outputs:

• Text Preprocessing: Tokenization, stop word removal, stemming, and text cleaning to prepare data.
• Model Development: Selection of a suitable NLP model (e.g., Logistic Regression, LSTM, or BERT) for multi-class fraud categorization.
• Accuracy Measurement: Evaluate the model based on accuracy, precision, recall, and F1 score metrics.

• Result:

Participants will deliver a working NLP model emphasising achieving the highest possible accuracy within the 3-day timeframe.

III. Stage 3:

a) Participation requirement:

• The Top 3 teams will be selected after evaluating the results from Stage 2.

b) Event Process and Schedule:

• The teams will pilot and deploy their prototype based on the datasets and development environment provided in collaboration with IndiaAI, I4C, and a MeitY empanelled CSP.

• The teams will be free to choose any of the empanelled CSPs as per their choice. The solution will then be presented to the Jury along with a presentation made by the team.

• The winning team will be awarded INR 25 lakhs based on their working prototype. The winner will also get a contract from the government of India and its associated entities to deploy their solution for four years.

Expected Deliverables from Participants

• Model Code and Documentation (OpenForge/GitHub)

• Clear and well-documented code used to build, train, and test the submitted model.
• Explanation of the key methodology and steps taken in model development.

• Project Report (PDF)

Discuss the significant findings from your NLP analysis (Summarize in 500 words)

For instance:

- Sentiment trends over time

- Commonly recurring themes/topics

- Present visualization as needed

- Text classification accuracy and key drivers behind correct/incorrect predictions

IndiaAI’s CyberGuard Hackathon provides an unparalleled platform for AI enthusiasts to contribute to India’s cybersecurity framework while fostering innovation and collaboration in AI.

For more information and to register, visit the IndiaAI portal.