Corners of twitter are weird, while a lot of academic twitter (#NLProc/#ML) seems to have moved here, I did not see similar chatter on my book twitter account. Please suggest some good and fun non-#NLProc accounts to follow because I really do want to give this platform a fair shot, and I can't with ML alone. If it helps, I like books, coffee, dogs, ocean life, and dance.
Shaily Bhatt
- 95 Followers
- 117 Following
- 46 Posts
Pre-Doctoral Researcher at Google Research India, Chair at Widening NLP Workshop, Prev at MSR. Interested in reliable, responsible, and socially-aware multilingual #NLProc. Looking for PhD positions in 2023.
Just want to point out that one the coolest thing about our BINDI effort is that we are actively looking at biases encoded in the data used to train LMs.
For more details check this AACL paper (https://aclanthology.org/2022.aacl-main.55/), joint work with Shachi Dave, Vinodkumar Prabhakaran, Sunipa Dev, and Partha Talukdar
Re-contextualizing Fairness in NLP: The Case of India
Shaily Bhatt, Sunipa Dev, Partha Talukdar, Shachi Dave, Vinodkumar Prabhakaran. Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2022.
For those not wanting to head to the birdsite, here is the text from the tweet linked above that announced our project:
As India’s digital journey becomes powered by AI, it's crucial that we consciously invest in efforts to ensure that real-life societal biases don’t manifest in tech.
Introducing Project BINDI:
🔴Bias
🔵Interventions for
🟢NLP and
🟠Data in the
🟣Indian context
Also read about exciting research other members from my team are pursuing for developing better language technologies for India.
https://twitter.com/partha_p_t/status/1604908136267419649?t=4wIdGyyL5YmdjfNNzxg6CQ&s=19
Partha Talukdar on Twitter
“Super excited about 3 projects my group at Google Research has been working on getting announced at #GoogleForIndia #1 Vaani: collaboration with @iiscbangalore to capture India’s speech landscape by collecting and open sourcing speech from 773 districts https://t.co/fXx94xwgfV”
The project I work on got announced in the #GoogleForIndia event yesterday! I am so proud of being a part of this super important work on creating fair and inclusive language technologies with special attention to the socio-cultural context of India.
https://twitter.com/GoogleIndia/status/1604737600035880961?t=CJo9tKQTVLnDwNn1OnAcPA&s=19
Google India on Twitter
“As India’s digital journey becomes powered by AI, it's crucial that we consciously invest in efforts to ensure that real-life societal biases don’t manifest in tech. Introducing Project BINDI: 🔴Bias 🔵Interventions for 🟢NLP and 🟠Data in the 🟣Indian context #GoogleForIndia”
My feed here is uncannily full of posts on AI art. Is this from sigmoid.social? because I primarily follow NLP people and I don't recognize most accts whose posts I see.
Maria AntoniakI'm enjoying Mastodon more and more. My votes for what would make it even better:
- an algorithmic feed
- better conversation threading
- always viewing the original profile, not some mysterious percentage of the original profile (same for likes and boosts)
- seeing my overlapping follows when looking at someone else's follower list
- encrypted DMs
- DMs that don't visually operate like posts
- quote posts! I promise they can be used for good!
- an algorithmic feed
- better conversation threading
- always viewing the original profile, not some mysterious percentage of the original profile (same for likes and boosts)
- seeing my overlapping follows when looking at someone else's follower list
- encrypted DMs
- DMs that don't visually operate like posts
- quote posts! I promise they can be used for good!
aidoThere's no systemic racism but weirdly every time we train an AI on public data sets it becomes very racist.
How to make mastodon link from twitter open in mastodon (and not web browser)? There is no "open in mastodon" option.