The IIT Patna CSE team works to combat hate speech online. First stop: Twitter

IIT Patna: IT engineers working on an AI and ML tool to combat hate speech online.

IIT Patna researchers are developing an AI-based model to tackle hate speech issues in the country.

NEW DELHI: It works through a browser extension. Once installed on a web browser, it can scan a page to detect “hate speech” and prevent it from spreading. The software module called Hate, Hyperpartisan, and Hyperpluralism Elicitation and Observer System – or HELIOS – was built by the AI-NLP-ML group at the Department of Computer Science and Engineering (CSE) of the Indian Institute of Technology (IIT ) Patna, in collaboration with the Indraprastha Institute of Information Technology (IIIT) Delhi and the University of Texas Austin.

Recommended : Get important details about IIT Patna. Download brochure

Still being refined, the tool uses artificial intelligence (AI), natural language processing (NLP) and machine learning (ML) – the terms from which the research group IIT Patna derives its name – to identify hate on text page including social media. stations, and disinfect it.

“With hate speech problems increasing in India, how do you detect offensive content and ensure it is stopped when it happens? If you allow hate speech to spread, there may be serious situations in the maintenance of law and order,” said Asif Ekbal, who leads the project at IIT Patna. “Lately, the number of hateful or offensive comments targeting certain people has increased dramatically. is how to detect this, because 50 million tweets or messages are generated daily in the world. Of this number, two to three percent are hateful content. How do you filter it so that you can control offensive content on the platforms? ?”

Read also |IIT Delhi and ITC to conduct collaborative research in STEM fields

IIT Patna: HELIOS and hate

The IIT Patna team started planning the project in 2019 and was looking for collaborations. In 2020, he found a partner in major IT company Wipro to create a commercially viable product to address the problem of hate speech. “Hate speech is always targeted at someone or something and we wanted to identify this content using an AI-NLP-ML based technique developed at IIT P,” Ekbal said. Other project collaborators include Pushpak Bhattacharya, Professor, CSE Department and also Head of the Research Group; Sriparna Saha, Associate Professor, CSE; university researchers; and his colleagues from IIIT Delhi and the University of Texas.

The definition of hate varies geographically, Ekbal explained. “It depends on the country, language and other factors,” he said. “Currently, we are developing research for Hindi and English. Our main objective is to develop solutions for Hindi and English capable of automatically detecting hate speech on social networks. A large amount of data – “offensive content” – is first entered into this system so that it learns to detect hate speech or offensive content in the browser. “This is done through an algorithm based on AI and NLP-ML technology,” Eqbal said. “Once detected, it automatically converts offensive content to non-offensive content. This program is applicable to websites or any social media platform, blog, etc. Currently, we are only working on tweets. It takes text as input, then detects offensive content and converts it automatically. It will work as an extension on the browser.

The primary goal of HELIOS is to stop this content and prevent its further spread. It is intended to prevent mail from being forwarded. It also prevents the spread of fake news.

Read also |Vedic mathematics, astrology, anti-vaccine: this engineer denounces pseudoscience on YouTube

AI, machine learning

“In AI and machine learning models, machines learn to fork between hateful and non-hateful content. It cannot automatically detect this data and this data creation and further development of this data requires a lot of money,” Ekbal said. Wipro funded the project, putting Rs 42 lakh into the consortium’s work for three years, according to Ekbal’s website.

The project consists of three parts: data creation, research and product engineering construction. Ekbal noted, “Wipro is giving us funding in terms of data creation, labor, and building the product. Once we have completed our research and prototype, Wipro will refine the research and create an app or product from it.

Moreover, research and products that have a greater impact and solve a socially relevant problem are relevant to business and industry because they want such solutions. “Their ultimate goal is commercialization and through this project they aspire not only to prevent hate speech but also to ensure commercialization. Through our research, Wipro will build a product for their client,” Ekbal said.

Read also |“To affirm the constitutional policy”: what are the messages of Professor DU prof Ratan Lal on YouTube?

Software engineering students

Overseen by Bhattacharya, Ekbal and Saha, the AI-NLP-ML is a thriving research group and this project has trained and exposed a large number of students.

“This project includes customer interaction that exposes students to marketing and interaction. Additionally, students are empowered to solve a problem that is relevant to society and industry, as these technologies have multiple applications,” Ekbal said.

Technology development also involves ideation and the exchange of ideas. “Students can exchange their ideas with other researchers and international organizations. Also, as the students solve the problem by being part of the team, it makes them more professional,” Ekbal said.

Ekbal said the team plans to extend the project beyond December 23, when the project’s three years are over, and develop the same solution for other languages.

Read also |35% of male dropouts ‘not interested in school’: What the NFHS 5 says about school education

“We started in 2020 and developed a prototype and research papers. We used Wipro’s basic software and they’re building some sort of product using our research submissions,” he said.


Follow us for the latest education news on colleges and universities, admission, courses, exams, schools, research, NEP and education policies and more.

To contact us, email us at news@careers360.com.


Learn more about


  • NIRF classification
    21

  • Evaluation
    4.4/5

  • Questions and answers
    95

Comments are closed.