On the occasion of International Women’s Day, the Women in Data Science Conference, a Stanford initiative, held its 2021 edition. Since it was fully virtual it brought together multiple regional events, including Pune region, in a successful effort of global collaboration.
The Women in Data Science Conference with its mission to involve and educate more people in this field addressed various aspects of Data Science — applications ranging from medicine to learning platforms, and ethics, inclusion and leadership roles for women in this industry. I attended several interesting sessions organised by Pune section as well as the ones being livestreamed from Stanford University. In this blog I note some of my key takeaways from these incredible sessions.
Keynote Address — Jean Liu, President, Didi Chuxing
Jean talked about resilience for an industry especially in the unprecedented times of today, and how resilience comes from user belief and kindness.
Jean mentioned Didi’s women network and their mission statement — be great, be you. While talking about being an inspiration for younger girls who approach her for guidance, she said that she’s normal just like us, adding “normal people can create greater good” which really stuck with me.
AI for Finance Automation — Deepti Chafekar, Director, Data Science, AppZen
AppZen is an AI platform for finance teams with nearly 1/3rd of the Fortune 500 companies as their customer base. Deepti talked about automation and the importance of feedback. In her words, “automation is only effective with human feedback”. She concluded by reasserting that AI is a powerful tool for humans but not something that can replace human resources.
Panel: Ethics and Responsible Data Science
Moderator: Shir Meir Lador, Data Science Group Manager, Intuit
Andrea Martin, Leader IBM Watson Center Munich & EMEA Client Centers, IBM Distinguished Engineer, IBM
Monica Scannapieco, Head of the Division “Information and Application Architecture”, Italian National Institute of Statistics
Nazareen Ebrahim, AI Ethics Officer, Socially Acceptable — South Africa
I attended this Panel Discussion on the theme of “Ethics and Responsible Data Science”. Shir summarised the themes of the discussion in three main categories — mindset, how to build solutions, and how to work as a community to advance ethical data science.
Some of the ideas and beliefs that really resonated with me were as follows:
Monica talked about the work being carried out at her organisation and how ethics are reinforced in the operations, from web scraping that doesn’t violate owner’s privileges to using privacy-preserving techniques for accessing personal data
Andrea elaborated on IBM’s efforts for AI & ethics, and how it is actually an end-to-end approach centred around processes, people and tools
Nazareen is involved with an AI ethics mentorship program for women all around the world paired with leaders. When asked what keeps her up at night, she answered “preserving dignity”
This was particularly interesting to me because I’m working on a project on differential privacy under IBM. While the subject falls under cryptography and privacy, its main application actually lies in data analysis.
Collaborate in an interdisciplinary way to achieve ethics in AI — Andrea
Towards the end, Nazareen stated that one should interrogate their question more than they interrogate the answer, food for thought for all of us.
What would you do if you weren’t afraid — Snehi Mehta, Head-Product Marketing, Facebook
Key points covered by Snehi:
Identifying microaggressions and blindspots
Diversity and inclusion is not just for HRs. If for most part, work culture affects the rest of us then shouldn’t all of us be working to inculcate it in everything we undertake?
Processes at Facebook — training, reporting & acting, community, accountability, measurability
I also had the chance to talk to her personally about leveraging the power of networks and she stressed upon the importance of communication, and reiterating to ensure you’ve correctly understood the other’s point of view.
Keynote Address — Joelle Pineau, Team Lead, Facebook Research, Associate Professor, McGill University
Themes — reproducibility, reusability and robustness
Joelle began by disclosing statistics about scientists expressing their concerns regarding reproducibility. Towards the end she introduced an ML reproducibility checklist, with elements like documenting dependencies, training code, results in a paper, etc.
Her research work has mainly been in the area of Reinforcement Learning of which she gave an introduction (there have been nearly 20,000 papers in this domain in the last few years, a number I find rather insane). She also talked about adaptive neurostimulation, an application of RL
Her concluding remark was — Science is a collective institution that aims to understand and explain. As someone interested in research, I’ve often been told about how competitive the field of academia is, and Joelle touched upon the same while urging us to not treat science as a competitive sport but as a collaborative platform.
Leveraging Data & Analytics to Solve Challenges in the Pandemic and Beyond — Gina Papush, Global Chief Data and Analytics Officer, Evernorth
Gina talked about a modeling approach for forecasting Covid-19 infections
She also introduced her company’s creation of an analytics suite of predictive solutions to better manage risks for employees and guide employer decisions
PS: I found it extremely endearing that she spoke majorly about her daughter’s achievements while introducing herself 🙂
Panel Discussion: Paths to Leadership in Data Science
Moderator: Martina Lauchengco, Operating Partner, Costanoa Ventures
Afua Bruce, Chief Program Officer, DataKind
Daniela Braga, Founder and CEO, DefinedCrowd, Inc.
Aishwarya Agrawal, Professor, University of Montreal; Research Scientist, DeepMind
Michelle Rodriguez, Dean, Engineering School Universidad del Pacífico
This was hands down my favorite discussion. I loved this congregation of powerful women from diverse cultures leading in diverse fields, from public sector to academia. I noted several points that I found insightful, they are as follows —
Michelle talked about the importance of communication and pursuit. In a society that sees women as different, it’s important to tell people about your ideas
Aishwarya talked about two kinds of skills she has learnt— some that are asked of her, and others that she observed in her environment and imbibed. An important aspect of doing research is to communicate your research — for instance, in presentations by giving more examples and demos of work instead of the math. Even while writing a paper, it’s important to communicate and design the flow of documents well. She said she learnt a great deal about these things while pursuing her Ph.D., including the skill of reasoning.
Afua raised the point of relevance of mentorship, and something that all of us worry about — finding a mentor and finding time with mentors. She said that one doesn’t need regular discussions with people to get inspired. Instead one can be motivated from just a few encounters, from public talks or even from social media. This approach has helped her serve the public better
Daniela is one of the few women holding leadership roles in a highly technical industry. When asked who motivates her, Daniela said her daughter is one of her biggest motivations. In her words, “she cannot live without me and what can be a greater motivation than that”.
As someone coming from a family of four sisters, Michelle said that it was natural for her to be a researcher, but walking into the world made her frustrated. She realised that not all people see the same, and understanding others’ perspectives has been a learning for her. She gave a quirky lil metaphor of how people have different glasses and during conversations she always wonders what the color of glasses of the people she’s talking to is. Martina added that it’s important to understand and acknowledge the vantage point of others, to learn to read a room in order to become a better listener.
I think I can definitely take that advice and be more considerate of conflicting opinions. Conflict is inevitable after all 🙂
Keynote Address — Shafi Goldwasser, Director, UC Berkeley and Professor, MIT and Weizmann Institute of Science
I was totally fangirling when I was informed of this session. I stayed up exclusively for this one and what a delight it was.
Shafi mainly spoke about the emerging role of cryptography in ethics of AI. She pointed out three relevant areas — privacy during training, robustness, and verifiability. She took us through these three key areas by describing her own research in all of them.
Once again, I was very happy to learn about the high applicability of data privacy for data science. Working on an open-source project centred around the same, it’s always nice to hear researchers from all around the world advocate the significance of it.
A statement that she made which I found very interesting was about the power of Machine Learning. She said — the power of ML comes from data but data belongs to individuals. Thus, the power lies in the hands of whoever owns the data.
Cryptography’s playground lies in the intersection of privacy and functionality and exploring the trade-offs between the two. My key takeaway from the sessions of the day has truly been about the rising significance of ethics in AI. I duly noted that there is a wide scope for research in the domain of cryptography and how it’s shaping up and reinforcing ethical practices.
All in all, my experience was amazing. I have previously volunteered with the Core Committee for WiDS Pune 2019 and even back then I was mesmerised by the sheer confidence, charm and experience of all the women speakers. Sitting at my desk, watching these conversations literally from a world apart, I felt the warmth, honesty and determination of every single woman who walked onto the stage (metaphorically). They continue to inspire me to be “free to be”.
March 8 — International Women’s day — is a day to respect and celebrate women and their talents and I hope someday the spirit of it becomes greater than the duration, that we consciously uplift and acknowledge women every single day.
Happy Women’s Day! Cheers!