aidaa Profile Banner
Aida Davani Profile
Aida Davani

@aidaa

Followers
1K
Following
16K
Media
90
Statuses
9K

Catsplaining science | CS PhD @USC | Research Scientist @Google | NLP/AI Ethics | She/her

Joined September 2008
Don't wanna be here? Send us removal request.
@aidaa
Aida Davani
4 months
RT @MohammadAtari90: Thanks to my mentors, collaborators, and students. @JoHenrich @MortezDehghani @DOyserman @JF_Schulz @JonHaidt @cam_l….
0
9
0
@aidaa
Aida Davani
4 months
RT @HannaHajishirzi: Excited to drive innovation and push the boundaries of open, scientific AI research & development! 🚀 Join us at @allen….
0
15
0
@aidaa
Aida Davani
4 months
RT @jlambert_: I am hiring a Student Researcher at Google DeepMind for 2025!. 👩🔬 Interested in improving multi-turn optimization and reason….
0
90
0
@aidaa
Aida Davani
5 months
RT @MohammadAtari90: 🧿New Preprint🧿. We argue for diversifying conceptual diversity, i.e., studying topics that are not well represented in….
0
21
0
@aidaa
Aida Davani
5 months
@vinodkpg @blahtino . wonderful collaborators and inspiring teammates: @dylnbkr @justaturbo @laroyo @AliciaVParrish @cawelty @charvvvv_.
0
0
2
@aidaa
Aida Davani
5 months
Thrilled to share our research vision on the Google Research Blog. This work reflects 4 years of passion & dedication with a brilliant team. Deepest thanks to Vinod, my inspiring manager, and Mark, my invaluable teammate, #GoogleAI #ResearchPublication.
@GoogleAI
Google AI
5 months
Modern ML models are used by people around the world who have a variety of perspectives that often align along socio-cultural and demographic lines. Today we describe effective ways to integrate these considerations into the ML modeling pipeline.
Tweet media one
2
9
31
@aidaa
Aida Davani
5 months
RT @lujainmibrahim: 📣New preprint!📣We’ve long known humans tend to anthropomorphize computers. But with the rise of social AI applications,….
0
44
0
@aidaa
Aida Davani
8 months
RT @WOAHWorkshop: 🥳We are excited to share that WOAH 2025, our 9th edition, will take place at #ACL2025 in Vienna! @aclmeeting . Our speci….
0
9
0
@aidaa
Aida Davani
9 months
RT @IasonGabriel: Are you interested in exploring questions at the ethical frontier of AI research?. If so, then take a look at this new op….
0
68
0
@aidaa
Aida Davani
9 months
RT @canfer_akbulut: How do we anticipate and prepare for the impacts of anthropomorphic AI on users and society? We explore this question i….
0
74
0
@aidaa
Aida Davani
9 months
RT @payampiray: 📣 I'll be recruiting PhD students for Fall 2025. Take a look at the lab website and the FAQ for prospective students:.http….
0
85
0
@aidaa
Aida Davani
10 months
RT @MohammadAtari90: The Historical Psychology Pre-conference @SPSPnews is up! We have a fantastic lineup! . Themes include the role of his….
0
17
0
@aidaa
Aida Davani
10 months
RT @MohammadAtari90: We’ll be presenting our paper on historical text analysis in classical Chinese at #EMNLP2024 (the Main Conference).….
0
7
0
@aidaa
Aida Davani
10 months
(6/).By presenting GeniL task, dataset, and experiments, we call for more meaningful LLM evaluation methods that capture contextual nuances by looking beyond word-level associations. See you @COLM_conf, we will be presenting the work during Session 3 on Tuesday, Oct 8th.
0
0
4
@aidaa
Aida Davani
10 months
(5/).Why go multilingual?. We tested classifier with (1) English-only, (2) English + machine translations, & (3) annotated target language data; . (3) outperformed across all languages 🏆 as it captures locale-specific intricacies!. So, Multilingual data matters!
Tweet media one
1
1
5
@aidaa
Aida Davani
10 months
(4/).Using GeniL we asked: can we filter stereotypes with word pairs?.Not really. Simple word pairs overpredict stereotypes. Why?.(a) stereotype pairs often appear in non-generalizing context.(b) offensive generalizations are rare, and have different rates across languages
Tweet media one
1
0
3
@aidaa
Aida Davani
10 months
(3/).GeniL ( consists of 50k sentences in English, Arabic, Bengali, Spanish, French, Hindi, Indonesian, Malay, & Portuguese, mentioning identities & attributes from SeeGULL (, annotated by native speakers (no LLMs here 🙄).
Tweet media one
1
0
3
@aidaa
Aida Davani
10 months
(2/) GeniL focuses on the linguistic context and asks if a sentence mentioning a pair of identity and attribute is an instance of generalization. We distinguish two types of generalizations : (1) mere *mention* of a generalization, (2) *promotion* of a generalization.
Tweet media one
1
0
4
@aidaa
Aida Davani
10 months
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: co-led @gv_sagar, w/ @sunipa17 @shachi_dave @vinodkpg
Tweet media one
1
7
59
@aidaa
Aida Davani
10 months
RT @MohammadAtari90: The Historical Psychology preconference at @SPSPnews is back! . See you in Denver! . (and send us your abstracts) . St….
0
10
0