We are surrounded by smart machines! The machines that speak our language, perform mundane tasks for us and solve our queries and problems. Conversational AI is a field that has opened doors for human-machine interaction and soon this technology will enable machines to talk to zillions in their native languages enabling their technology inclusion. But as we move forward to that path, it is of utmost importance to carry out frequent checks to ensure that the chatbots can interpret our query and language correctly and respond accordingly. Ananya B. Sai, a research scholar at the RBCDSAI centre at IIT Madras aims to be a bridge that improves the way chatbots interact with humans to ensure that they don’t misinterpret each other.
Born and brought up in the city of palaces-Mysuru, her inspiration to take the long laden path of research came from her mother who is a research scientist at the CSIR- Central Food Technological Research Institute (CSIR-CFTRI). Ananya was always interested in logic and computers owing to which she enrolled in an engineering degree in computer science. Although her heart was fixed on pursuing higher education after the B.Tech degree, she joined CISCO for a brief stint while she was preparing for the GATE Exam. Her good performance in GATE Exam enabled her to get enrolled in the M.Tech Computer Science course at the prestigious IIT Madras from where her exciting research journey started. She enjoyed various courses like cloud computing and modern compilers but what struck her most was a course on “deep learning” which had endless applications and huge potential in future. Intending to learn more about the subject, she joined Dr Mitesh Khapra for her master’s project who worked in the same area. Her master’s research project was a hit as it got her a research paper in a prestigious journal and boosted her confidence to take up a PhD in the area. The wings to her dream of pursuing a doctorate were given by two prestigious fellowships- Google PhD fellowship and Prime Minister’s Doctorate Fellowship.
Her research at RBCDSAI involves studying various dialogue systems such as chatbots, language learning assistants, that are rapidly being deployed in the real world. Ananya aims to understand the limitations of these models mainly by scrutinizing and improving the scientific methods used to evaluate such models. Her research shows that the current models can learn to “game” or “fool” the existing evaluation metric by learning certain unexpected shortcuts that are not desired. She thinks it is important to improve the metric system used in this domain of text generation so as to account for various adversarial scenarios and also provide better actionable feedback while developing these systems.
One of the outcomes of her research would be to help corporates and organizations better handle customer queries through chatbots and only the most complex queries will be transferred to the human agents. This automation will spare humans from performing mundane and boring tasks.
When her mind isn’t wandering around her research, she loves to paint and play the mandolin for relaxation. When she finds spare time, she finds solace in the readings of Richard Feynman and books like Logicomix, Freakonomics. She believes that climate change is a bigger problem and awareness and action is required on that front. Apart from that, she believes that misinformation and rumors are the epidemics that we should worry about and perhaps is one of the reasons she took up research in Natural Language Processing which has the potential to check the spread of misinformation on social media.
To know more about her research visit her webpage.