While the world celebrates women’s day in different ways – both big and small – in the AI and analytics space, the need for collective voices on equality, the gender pay gap, and gender biases have become more important than ever.
Enough has been said about how women and girls in science and technology are adding value and how it accelerates progress towards scientific advancements . But there are much bigger things at stake in the AI and analytics landscape . Ever wondered why most voice assistants – Siri, Alexa, Google Assistant – all have female voices? That’s gender bias and stereotypes embedded in AI right there, notes UNESCO.
Women are given smaller research grants than their male counterparts, and while they represent 33.3 percent of all researchers , and only 12 percent of national science academies are women.
Only one in five professionals is a woman in the AI field – i.e. about 22 percent.
Female researchers tend to have less well-paid careers. They are underrepresented in high-profile journals and are often passed over for promotion.
Check out the gender pay scale gap in data science domain below:
Meet Gender Pay Gap Bot . Earlier this week, one of the fact-checker accounts, called @PayGapApp , has been calling out names of companies on Twitter and revealing the percentage gap of how much women are getting paid as compared to men.
Thanks for reading The Belamy | Weekly dose of best AI stories! Subscribe for free to receive new posts and support my work.
In a bid to bring together some of India’s smartest and brightest women in AI and data science, Analytics India Magazine recently announced the launch of the fourth edition of the coveted ‘The Rising’ event (powered by Fractal), held in person on April 8 in Bengaluru, Karnataka. You can register for the event here .
Check out the best Indian conferences for women in tech in 2022 here .
Check out top data engineering conferences in 2022 here .
Women in Tech
Read Google head of product – responsible AI and ML fairness Tulsee Doshi’s interview on responsible AI and ML fairness here . In this exclusive interview, Doshi talks about the concerns surrounding AI and ML fairness and breaks the myths that often come with controversial issues.
Asha Vishwanathan leads the machine learning division at Verloop, a Conversational AI platform. Her expertise lies in NLP and computer vision, and her career spans over 14 years. She got introduced to data science in 2014 and made a pivot from analytics. Read her complete journey in data science here .
Want to know what is happening in the world of NFTs? Check out the video below.
Startups, Apps & Tools:
Nagarro is a leading global software development industry with its technology and software consulting services with over 15,000 employees in over 27 countries. The company is at the forefront of the AI disruption with its AI accelerators, data-driven business insights, big data, chatbots and more.
Goopt is an interesting search engine tool for a ‘procedural simulation’ of the web with GPT-3. This new web will use procedural content generation to create different content, completely synthetic.
Article in Focus:
At the Peak Performance Event, 2022, Apple announced the new M1 Ultra chip. Built on M1 Max’s transformational architecture, M1 Ultra is claimed to be the world’s most powerful chip for personal computers. Check out more details on Apple’s M1 Ultra chip here .
Watch the key highlight of Apple’s Peak Performance Event below:
Bengio & LeCun Debate on How to Crack Human-level AI
Lex Fridman had an interesting discussion recently with Yann LeCun and Yoshua Bengio at Meta’s Inside the Lab event. The trio discussed the latest advancements in AI and machine learning and possible paths to human-level intelligence. Read the complete story here .
Research & Papers
Meta’s chief AI scientist Yann LeCun, alongside scientists from MIT and NYU, recently introduced two variants named Direct (projUNN-D) and Tangent (projUNN-T) projected Unitary Neural Networks – an alternative method based on rank-k updates that maintain performance at a nearly optimal training runtime.
In collaboration with the University of Washington Columbia and Tel Aviv University, researchers from Google and Meta proposed a more accurate and robust alternative to the second step of the conventional recipe in the context of fine-turning a large pre-trained model. The second step of the conventional recipe includes picking the individual model that performs best on a held-out validation set, besides training multiple models with various hyperparameters.
Social media dialogues
Top News of the Week