Date: April 7th, 2020
Earlier this week, Crayon’s Sr. Data Scientist Dr U Kanimozhi spoke at Women in Data Science (WiDS) Ashoka 2021, a conference organized by Ashoka University in collaboration with WiDs Stanford University.
The virtual event kicked off with a panel discussion featuring Kanimozhi, along with Bindu Narayan, an AI leader; Gargi Dasgupta, IBM Research India; Kalika Bali, Microsoft Research Labs; and Lipika Dey, TCS Innovation Labs. The topic of discussion was “Data Science for Social Good” and what it meant to each of them. The panelists spoke about how aspiring data scientists can learn from the data available to them and build computational models. This in turn can enable services and tech for marginalized communities. They shared personal anecdotes as women in the tech industry, and examples of how data science can be applied to tackle real-world problems.
Panel Discussion: Data Science for Social Good
Here are some key highlights from the session:
1. The paradigm of centralized computing is slowly changing to distributed computing and federated learning techniques. This will help scale-up solutions for social good.
2. With the help of education, open-source tools, and innovative solutions, we can create inter-disciplinary programs and content to train the next generation of data scientists. As responsible and collaborative individuals, they will be equipped to tackle social problems that matter to the community.
3. Data science for social good is about fostering the use of data science for positive social impact through research, engineering, and initiatives to build an AI ecosystem by addressing societal challenges.
- Education: Creating interdisciplinary programs and content to train the next generation of responsible data scientists to tackle social problems that matter.
- Tools and solutions: Incubating ideas, tools, and solutions to scale the impact of data science for social good. All codes being available as open source.
- Community: Growing a global community to collaborate on problems with a social impact.
4. While tech researchers and developers seemed to have solved the “last mile” problem, their next challenge is to crack the “access-point” problem. There is currently a large gap between people who have access to technology and those who do not. Until this problem is solved, the gap will only grow, creating classes in society based on accessible technology.
5. Data scientists need to be very aware of the socio-cultural context in which their technology is being used.
6. Human judgement and discretion while collecting, cleaning and structuring data is important. Especially while creating technology that is trustworthy.
7. AI needs to be defined as augmented intelligence, instead of artificial intelligence. Create models that aid, rather than simply automate.
8. For aspiring data scientists: build your foundation and sharpen your analytical thinking skills. Then identify a problem you want to solve and find applications to it, rooted in the community around you.
Fireside chat with Dr Kanimozhi
In her Fireside Chat following the panel discussion, Kanimozhi answered questions on NLP (natural language processing) and its applications in Healthcare, Agriculture, Law, Cybercrime and Education and Research. She elaborated on the future of machine learning and how it could impact the labor force, the ethics of AI-models and how it varies from nation to nation. She also shared her journey as a data scientist: from a researcher looking for the right datasets to building a personalization engine with Crayon Data.
Watch the full Fireside Chat:
Fireside chat with Dr Kanimozhi at WiDS Ashoka 2021
There was a series of Lightning Talks in regional languages by industry leaders, following the Fireside Chats. Speaking in her mother tongue, Tamil, Kanimozhi shared her journey as a researcher, data scientist and teacher, and recounted how she picked up coding by learning online. She also provided a road map for aspiring data scientists and students looking to enter the field of data science.
Watch the full Lightning Talk: