Negar Rostamzadeh
@negar_rz
Followers
10K
Following
22K
Media
254
Statuses
5K
Staff RS @Google Research, Adjunct Professor @McGillu, Associate Industry Member @MIlAMontreal
Joined August 2016
Today @CVPR I had the honour of (1) giving a keynote on *From System-Level Assessment to the Geometry of Manifold: A Framework for Auditing via Controllable Synthesis* at @emacs workshop https://t.co/yKKrlxqAqb (2) giving a keynote on *Towards Meaningful Cultural Inclusion in
sites.google.com
This workshop focuses on generative techniques for experimental auditing of machine learning models, to analyze performance, bias, and failure modes
0
0
12
I was impacted by the Meta PAR/FAIR layoffs today. I’m a research scientist passionate about computer vision, generative models, multimodal understanding, and AI safety. Over the past decade, I’ve contributed to advancing a broad range of applied machine learning problems — from
50
70
1K
It’s both spooky and cute to see all the Halloween decorations at my daughter’s daycare! 🎃
0
0
6
For those who claim there's no genocide against Christians in Nigeria this is my response! The Dead Christian Bodies in Plateu Don't Lie! The Dead Christian Bodies in Benue Don't Lie! The Dead Christian Bodies in Kaduna Don't Lie! The Dead Christian Bodies in Taraba Don't
10
9
161
I am quite excited to share that our efforts in organizing and running "The VQA series of challenges" have been recognized with the 2025 Mark Everingham Prize -- https://t.co/oZrGEBBU5B for "stimulating a new strand of vision and language research". Thank you to the PAMI TC
9
15
117
Congratulations to Aishwarya Agrawal (@aagrawalAA), Core Academic Member at Mila, who has received today the Everingham Prize 2025, one of the most prestigious awards in the computer vision community, in recognition of her foundational contributions to the field.
4
8
49
Congratulations to Mila members @ShalalehRismani, Leah Davis, Negar Rostamzadeh (@negar_rz) and AJung Moon (@RoboEthics) for their Best Paper Award at AIES 2025!
I am incredibly excited and honoured that our #AIES2025 paper on *Measuring What Matters: Connecting AI Ethics Evaluations to System Attributes, Hazards, and Harms* got the *best paper award* This paper is led by @ShalalehRismani and coauthored by @renee_m_shelby Lea Davis and
0
1
5
🚨 NEW: Introducing Deep Research Your on-demand research analyst, powered by MessariAI. Generate bespoke, analyst-grade research reports with just one click. Exclusive to Enterprise users.
11
28
89
Félicitations aux membres de Mila @ShalalehRismani, Leah Davis, Negar Rostamzadeh (@negar_rz) et AJung Moon (@RoboEthics) pour leur Best Paper Award à AIES 2025!
I am incredibly excited and honoured that our #AIES2025 paper on *Measuring What Matters: Connecting AI Ethics Evaluations to System Attributes, Hazards, and Harms* got the *best paper award* This paper is led by @ShalalehRismani and coauthored by @renee_m_shelby Lea Davis and
0
3
16
If you are attending #ICCV check out our paper *Erasing More Than Intended? How Concept Erasure Degrades the Generation of Non-Target Concepts* presented and led by @amara_ibtihel 🥳 https://t.co/XGhnxzwcRz
https://t.co/I0ChUTi8xT
lnkd.in
This link will take you to a page that’s not on LinkedIn
0
3
7
I am incredibly excited and honoured that our #AIES2025 paper on *Measuring What Matters: Connecting AI Ethics Evaluations to System Attributes, Hazards, and Harms* got the *best paper award* This paper is led by @ShalalehRismani and coauthored by @renee_m_shelby Lea Davis and
4
1
27
Riot Platforms Reports Third Quarter 2025 Financial Results and Strategic Highlights. - Record quarterly revenue of $180.2 million - Net income of $104.5 million, or $0.26 diluted EPS - Adjusted EBITDA of $197.2 million - Announces initiation of 112 MW of core and shell for data
17
34
223
2 papers each accepted by 2 reviewers out of 3. No strong arguments against the paper, AC accepting the paper and the final result reject with no additional explanation? 😳 Any insights about @NeurIPSConf #NeurIPS2025 decision process this year?
1
1
8
Heureux d’annoncer aujourd'hui mon nouveau rôle de directeur scientifique au @Mila_Quebec! Grand honneur d'avoir cette opportunité de servir cette communauté de leaders et d'innovateurs en IA, que j'ai toujours chérie et dont j'ai moi-même bénéficié. https://t.co/Jm8lQQlIvg
mila.quebec
Montréal (Québec), le 2 septembre 2025 – Mila, l’institut québécois d’intelligence artificielle (IA) annonce la nomination de Hugo Larochelle, professeur associé à l’Université de Montréal et ancien...
11
6
145
Excited to share that I begin today as Scientific Director at @Mila_Quebec! Truly honored by this opportunity to serve this community of AI leaders and innovators, that I've always cherished and have benefited from myself. https://t.co/l0d77Afbry
mila.quebec
Montreal (Quebec), September 2, 2025 – Mila, the Quebec Artificial Intelligence Institute, announces the appointment of Hugo Larochelle—Adjunct Professor at the Université de Montréal and former head...
86
36
621
So proud of @_AKassem's amazing work! My first #EMNLP2025 paper 🥳
🎉 Thrilled that our work has been accepted at #EMNLP2025 (Main Conference)! TL;DR: We propose a framework to predict & explain unintended side effects in models (e.g., emergent toxicity, forgotten knowledge) using OOD data. Huge thanks to @gfarnadi, @negar_rz, and Zhuan Shi 🚀
0
0
6
Well, @OpenAI just banned @ChatGPT from giving medical advice. @Doctronic is built different: peer-reviewed clinical validation, a network of licensed physicians across all 50 states, and partnerships that let us keep expanding. We're not pulling back. We're moving forward.
3
27
190
🎉 Thrilled that our work has been accepted at #EMNLP2025 (Main Conference)! TL;DR: We propose a framework to predict & explain unintended side effects in models (e.g., emergent toxicity, forgotten knowledge) using OOD data. Huge thanks to @gfarnadi, @negar_rz, and Zhuan Shi 🚀
🚨 What happens under the hood when you fine-tune or unlearn LLMs? We introduce MNEME—a sparse model diffing method that predicts & explains unintended side effects (like emergent toxicity or forgotten knowledge) without needing fine-tuning data. @ActInterp #ICML July 19 Poster
0
2
18
Come see our poster about how to predict side effects of unlearning and Fine-Tuning at @ActInterp
1
4
25
صاحبخانههای سودجو، پول رهنخانهها را نمیدهند و حتی مهاجران را تهدید میکنند که گزارشتان را بخاطر همکاری با موساد به پلیس میدهیم! چه شد آن تاریخ و فرهنگ چندین هزار ساله؟خجالت نمیکشید؟ به هر چیزی که باور دارید حق غریب خوردن ندارد! #نه_به_نژادپرستی
51
77
538
Anyone from #iran looking for a phd/postdoc/research internship in statistical learning theory, deep learning theory etc, contact me. Please retweet.
5
50
103
Crypto 101: seed phrases are broken. We don’t use them. Each wallet is independent, backups encrypted, PIN-protected. No photo, glance, or leak can steal your funds. Yet the old, vulnerable story is still everywhere.
2
12
65