Shachi Dave
@shachi_dave
Followers
344
Following
32
Media
0
Statuses
85
Striving to be a better version of myself
Joined March 2012
Today, we have open-sourced the complete code for CUBE, including i) cube extraction code, ii) cultural diversity evaluations using Vendi score, and iii) CUBE dataset. Check it out! Github: https://t.co/V7oVzQX7sz Thanks for all the interest, patience and discussions during
github.com
Code for CUBE extraction and cultural diversity metric introduced in "Beyond Aesthetics: Cultural Competence in Text-to-Image Models" - google-deepmind/cube
🚨 New paper alert 🚨: T2I models are about to revolutionize industries...but do they work equally well for all cultures ? 🌏 What is the *cultural diversity* of images generated for under-specified prompts? 🤔 Our latest research delves into this! https://t.co/ZQOVWQHHOs
1
11
30
Application are now open for the Predoc 2025 program at Google DeepMind India. If you are pre-PhD (i.e., current undergrad or masters students or recent geaduates) and interested in research, then this should be a no-brainier. Please apply by Dec 18, 2024 https://t.co/Sm12S7CAPk
5
62
388
So happy to see @simi_97k get this well-deserved recognition! The seeds for this idea were sown during her predoc stint at Google Research India and she continued to nourish it with her passion. Our predocs always make us proud! :)
Thank you so much @emnlpmeeting for this wonderful recognition! I’m so honored and humbled 💕 Thanks @gneubig for your support throughout! We’ve been working on this for 1.5 years and everyone who has spoken with me in the recent past knows how passionately I feel about this
1
0
24
Excited to share our prompt tuning playbook! (Not an official product. Just authors tips & tricks for better prompting). I'm most excited about first half on mental models for post-training & prompting. Feedback/forks welcome! #LLM #PromptEngineering
https://t.co/TrVhKVJc64
13
133
610
We have an immediate opening for a Research Scientist in our team at @GoogleDeepMind India. If you have experience / are interested in the field of Culturally-aware Multimodal Models to make AI more globally inclusive, please apply.
job-boards.greenhouse.io
5
46
250
So happy that this work has been accepted at #NeurIPS24 (D&B). Nice culmination of the amazing journey with these awesome teammates! Very excited for what's next...
This is accepted to #NeurIPS2024 (D&B). Thanks to my amazing co-authors for the contribution and support throughout this project. @arif_ahmad_py @adjiboussodieng @utsav_prabhu. Special mention to @shachi_dave and @vinodkpg for all the guidance. See you in Vancouver!🌲
0
0
18
Continuing on our work on stereotypes, we introduce a multilingual dataset to detect generalizing language. To be presented at @COLM_conf by @aidaa and @gv_sagar.
NLP stereotype evals often focus on word co-occurrences, but ignore if those words appear in a generalizing context. See our @COLM_conf paper on new task, data & models for generalizing language: https://t.co/NcmpUi2OgO co-led @gv_sagar, w/ @sunipa17 @shachi_dave @vinodkpg
0
0
11
What a wonderful platform to interact directly with the authors! Thanks for featuring our paper, @askalphaxiv!
New from @GoogleDeepMind! Beyond Aesthetics: Cultural Competence in Text-to-Image (T2I) Models The first benchmark to evaluate T2I models across cultural awareness and diversity 🌍 Authors @NithishKannen @vinodkpg @adjiboussodieng @shachi_dave are here to answer your questions!
0
0
11
Exciting opportunity in the Languages team at @GoogleDeepMind India to advance LLM frontiers and bring their benefits to a lot more people! Get in touch for any queries. I shall also be at @aclmeeting next week, happy to chat in person. https://t.co/6Gj46TsaE7
job-boards.greenhouse.io
4
43
239
Our latest research on measuring cultural competence in T2I models is finally out! Kudos to our super talented predoctoral researcher @NithishKannen for his single-minded focus and hard work. So excited for what's next...
🚨 New paper alert 🚨: T2I models are about to revolutionize industries...but do they work equally well for all cultures ? 🌏 What is the *cultural diversity* of images generated for under-specified prompts? 🤔 Our latest research delves into this! https://t.co/ZQOVWQHHOs
0
0
17
Ted Gioia thinks our culture is stagnant. Algorithms have us addicted to distraction, Hollywood's out of creative ideas, and people don't read hard books like they used to. 16 lessons about writing and breaking free from dopamine culture: 1. Ted's golden rule for writing:
35
97
419
Another great contribution from our team to advance research in Indian languages. Work led by our very talented predoctoral researcher @Harman26Singh and senior researcher @nitish_gup.
IndicGenBench is accepted at ACL'24. Use it via huggingface: https://t.co/ycqCOzh9EF
@GoogleAI
@partha_p_t @nitish_gup @ShikharSSU @dinesh_tewari1
0
0
13
Excellent use of SeeGULL dataset to build visual stereotype dataset for evaluating Text-to-Image models. Shout out to our awesome student researcher @akshitajha and teammate @sunipa17 for all the blood and sweat that went into this! So happy to see this work land at ACL.
New #ACL2024 paper alert📢: Introducing ViSAGe: "Visual Stereotypes Around the Globe" - a dataset enabling evaluation of known nationality-based stereotypes in Text-to-Image models. Work w/ @sunipa17, @vinodkpg, @cephaloponderer, Sarah, @shachi_dave, @qadrida, @chandakreddy 1/n
0
3
15
Very important contribution in the multilingual space for stereotype detection. Kudos to @mukBhutani and @sunipa17 for their perseverance in seeing this through!
🎉 New ACL paper alert 🚨: SeeGULL Multilingual: a Dataset of Geo-Culturally Situated Stereotypes across 23 languages and countries, to detect and mitigate unsafe and unwanted generation by language models. Joint work with @sunipa17, @krob, @vinodkpg, @shachi_dave. @aclmeeting
0
1
12
Strongly recommended if you want to work with a group of highly motivated researchers on new research directions.
Postdoc opening in Languages group at Google Deepmind based out of Bangalore Topics: LLMs, multilinguality, multimodality, RAI, etc. Strong candidates may apply by sending cv to research-programs-in@google.com with [LLM-Postdoc] in subject by Apr 26 DM/email for any questions
0
0
11
Highly recommend attending this talk if you are around.
I will be at the @indoml_sym at @iitbombay this week. Looking forward to chatting about geo-cultural representation in AI. Come say hi if you are around!
0
0
9
Storytelling is intrinsic to India’s rich cultural heritage. We teamed up with GDI, USC and IAA to apply our AI to study how visual storytelling on Indian TV is representative of India's demography. Learn more → https://t.co/4Fta7Jr4wO
72
61
285
Excited to share our latest research on using AI to study demographic representation in Indian TV shows -
blog.google
Our machine learning technology & LLMs are providing first-of-its-kind multilingual analysis to inform equitable content in mainstream media
0
1
14
Storytelling is a superpower. Steve Jobs said: "The most powerful person is the storyteller." Unfortunately, most people don't know how to tell a good story. So they are at the mercy of those who do. Let's change that: Here's 5 principles for telling an amazing story: 1.
41
403
3K
The happiest people in the world are the ones who: • love what they're doing • accept things that they can't change • find pleasure in doing things • continue to grow and evolve • practice humility • expect less • enjoy and appreciate little things in life
148
2K
10K