01/16/2020 16:13:50

IT tools for sustainable development: Chatbots aid library staff, blood donors, abuse survivors

Photo. Mikhail Burtsev, head of DeepPavlov project and Neural Networks and Deep Learning Lab at MIPT. Credit: Evgeniy Pelevin/MIPT Press Office

The sustainable development goals of the United Nations are a call to action. They are an appeal to all nations — whether rich, poor, or middle-income — to improve well-being and protect the planet. As a top-level global university and research center, MIPT is prepared to address the problems in technology and education, and supply the evolving international labor market with specialists whose expertise makes them indispensable for the emerging fields of the coming decade.

“The careers of the future are those in information technology. I mean artificial intelligence, machine learning, high-tech medicine. Robotics belongs there, too, naturally,” said Ekaterina Yakovleva, the head of the MIPT Career Center. “Research and development are precisely the tasks where machines cannot replace humans.”

The Moscow Institute of Physics and Technology adheres to the 2030 Agenda for Sustainable Development. As an academic and scientific center, MIPT is willing to openly share the tools for meeting the sustainable development goals: ending poverty, solving a range of health care issues, ensuring gender equality and social security, counteracting climate change, and protecting the environment.

One of such instruments is DeepPavlov, an open-source conversational AI framework for developing chatbots and virtual assistants. It has comprehensive and flexible tools that allows developers and researchers working on natural language processing to create production-ready conversational skills and complex multiskill conversational assistants.

Already in use in 92 countries, the DeepPavlov library offers ready-to-implement AI components for working with text in more than 90 languages. Over 30 Russian companies have implemented DeepPavlov-based solutions and use them.

Last year, the Russian State Art Library used solutions developed by the project to reduce the workload on the employees responding to reader inquiries in the library’s online chat. Implementing the technology made it possible to automate answering some of the standard questions, decreasing average waiting time and increasing reader satisfaction.

Klimova Blood Donation Center in Nizhny Novgorod, Russia, uses DeepPavlov-driven technology on its website to automate the responses to frequently asked questions. This resulted in a functional user support service on the website, faster response times, more transparency and better access to information, and a lower workload on the support team.

More recently, students at Maastricht University in the Netherlands used DeepPavlov technology as the core for building an AI solution that helps people who experienced sexual misconduct. The students developed a #MeTooMaastricht chatbot, which offers guidance and support to harassment victims. It aids in making a record of the incident and directs users to organizations that provide rehabilitation after traumatic events.

The prototype bot is powered by machine learning and operates on Telegram. It is trained to classify the type of harassment described by a survivor and determine whether an attack was verbal or physical. It is also able to ask for basic information, such as the time and date of the assault. After that, the bot recommends survivors to seek medical or psychological care, depending on the severity of the incident. Finally, the bot provides guidance on how to inform the police about the abuser.

“Meeting the sustainable development goals will take professions and technologies of the future. That is, IT and the related industries,” commented Mikhail Burtsev, who heads the DeepPavlov project and the Neural Networks and Deep Learning lab at MIPT. “DeepPavlov library components are already used by leading Russian and foreign companies, research and education centers. For example, our technologies underlie a chatbot developed to support people who suffered from abuse. We believe that open-sourcing our library is a contribution toward creating new technological solutions that will help us meet SDGs.”

*To ask a question about the DeepPavlov library or using the models, please use the forum.

If you have noticed a mistake on this page, select it and press Ctrl + Enter

Related posts

Contest at NeurIPS 2019: 2nd place for MIPT team’s walking skeleton simulation

MIPT lab invites you to have a chat with artificial intelligence
MIPT lab will create dialogue systems for telemedicine
MIPT partners with conference on machine learning in art