Women data in data science can reduce algorithmic bias; Praxis launches a scholarship program to bridge the gender divide
Delve into history of computing and you’d find that women played integral roles in the World War II war efforts in the United States. As men were drafted in to fight, many ‘computers’ (a role performed by humans who computed manually) during WWII were women – frequently with degrees in mathematics. Women were involved in calculating ballistics tables during the Wars. In the sixties too, women employees engaged in computing work were called “Computers”; but since then there has been a steady decline in women in technology.
Studies have shown that only a third of the IT workforce is women, and it is dwindling specially in India when women employees drop out in their third year of employment to raise a family. The gender gap in data science is even sharper. In a class of 35 students hardly six of them would be women.
Some international studies put women in as little as 10% of the industry’s executive leadership positions, compared to 23% across all industries. This is even though diversity, including gender diversity, is just plain good for business. In 2019, the top quarter of companies for gender diverse executive teams were 25% more likely to have above-average profitability than their counterparts in the bottom quartile.
Women hold just 18% of data science jobs in the US, and the problem is worse in most lower-income countries, where women are less likely to have access to the science, technology, engineering, and mathematics (STEM) education that provides a entry-point to a career in data science. In addition to increasing the risk of bias, gender imbalances in STEM and data science training make it harder for women to succeed in high-paying professions linked to the digital economy, further widening gender pay gaps.
Every study indicates that there is a compelling business case to have women in leadership roles in companies and it is even more important to include women data scientists. The reason is simple; bias in artificial intelligence is a major challenge and bringing in more women into the teams is a sure way of eliminating that unfairness in data models to a large extent.
As we navigate the lasting effects of the pandemic and social unrest, mitigating AI bias will continue to become more important. Here are several ways to get your own organization to focus on creating fairer AI:
- Ensure that training samples include diversity to avoid racial, gender, ethnic, and age discrimination.
- Whether labelling audio samples or generic data, it is critical to ensure that there are multiple and different human annotations per sample and that those annotators come from diverse backgrounds.
- Measure accuracy levels separately for different demographic categories to see whether any group is being treated unfairly.
- Consider collecting more training data from sensitive groups that you are concerned may be at risk of bias – such as different gender variants, racial or ethnic groups, age categories, etc.–and apply de-biasing techniques to penalize errors.
- Regularly audit (using both automatic and manual techniques) production models for accuracy and fairness, and regularly retrain/refresh those models using newly available data.
It is being noticed that gender bias is creeping into algorithmic models as those are being designed by men. In an attempt by Amazon to design a computer program to guide its hiring decisions, the company used submitted resumes from the previous decade as training data. Because most of these resumes came from men, the program taught itself that male candidates were preferable to women. While Amazon realized this tendency early on and never used the program to evaluate candidates, the example highlights how relying on biased data can reinforce inequality.
We at Praxis have decided to do something about getting more diversity, especially women, in data science. We have launched Praxis Women in Tech (WiT) Scholarships, an initiative to encourage and support women participation in tech, data and management careers. This is in line with our belief that gender diversity in the workforce, brings immense value to the organization, the economy, and the society. As a part of the Praxis WiT Scholarship program, women candidates who get selected to the prestigious PG programs* at Praxis and fulfil the WiT eligibility* criterion, will win a Scholarship of Rs 1 Lakh.
Click here for the details of the Praxis Women in Technology Scholarship program