Parinaz Sobhani (@PariAIML) header
Parinaz Sobhani (@PariAIML) profile picture

Head of Applied Research @georgian_io de-risk innovation and research for growth-stage startups, feminist, advocate for women in AI and fairness in ML

Toronto, Ontario
Joined March 2017
Thanks @jgibbs333 for highlighting my journey in STEM. I was born in IRAN, nine years after the Islamic revolution, which was pretty similar to @MargaretAtwood handmaid's tale. The new regime took away five decades of progress in women's rights over a few months. (1/5)
I’m excited to highlight amazing women in STEM who are Leading The Way like the incredible @PariAIML! Be inspired by her story here: instagram.com/p/CFiQfezH3ob/…
3
3
23
Parinaz Sobhani retweeted
Want to learn more about the latest AI research and commercialization? Catch these and other brilliant leaders in AI on February 22 at MaRS Impact AI. @PariAIML (@georgian_io), @shelbyzaustin (Arteria AI) and Chris Eliasmith (@UWaterloo) .
1
1
5
947
Experimenting with open-source LLMs, amazed at how easy & affordable it's become to fine-tune them. Sometimes, cheaper than a ☕️! #TechInsights rb.gy/bc05g
4
505
Parinaz Sobhani retweeted
How many of the signatories have used their expertise, their influence, their resources to rigourously analyze and expose the risks of current AI systems, and to build solutions that address real harms? That is what we need. Not letters, headlines and inflammatory statements.
We’ve released a statement on the risk of extinction from AI. Signatories include: - Three Turing Award winners - Authors of the standard textbooks on AI/DL/RL - CEOs and Execs from OpenAI, Microsoft, Google, Google DeepMind, Anthropic - Many more safe.ai/statement-on-ai-risk
23
65
11
461
156,341
A comprehensive study but I really hate using the word "AGI" for a language model!
At @MSFTResearch we had early access to the marvelous #GPT4 from @OpenAI for our work on @bing. We took this opportunity to document our experience. We're so excited to share our findings. In short: time to face it, the sparks of #AGI have been ignited. arxiv.org/abs/2303.12712
2
431
Parinaz Sobhani retweeted
The bravery of these women and the filthy brutality of the supposed guardians of morality. #MahsaAmini
Replying to @Shayan86
In Mashhad, a plainclothes officer beats women protesters on the side of a road with a baton as one woman tries to protect another by standing in front of her, on day five of protests in Iran over the death of #MahsaAmini in morality police custody. nitter.cf/1500tasvir/status/1572…
1,329
8,269
526
31,993
Parinaz Sobhani retweeted
Meet @PariAIML— our Head of ML and Applied Research! 🤩She leads our applied research team and continues to make an impact by helping our Georgian companies solve real-world problems. Want to work with Pari? Check out the opportunities at Georgian👇 gptrs.vc/3odS8NI
2
2
State of AI report is out and the most fascinated part for me is using large language models to generate new functional proteins and going beyond what evolution has designed! stateof.ai/
1
1
4
Parinaz Sobhani retweeted
On the Trustworthiness of Tree Ensemble Explainability Methods deepai.org/publication/on-th… by Angeline Yasodhara et al. including @PariAIML #MachineLearning #Hyperparameter
4
11
Parinaz Sobhani retweeted
We have an upcoming Paper Club Session on Accelerating AI with @databricks and how @tealium and Arteria AI are accelerating data and machine learning initiatives with Databricks. Register here 👇 gptrs.vc/3nBUwhh
2
1
Wednesday’s ☕ break just got a lot more exciting. Mark your calendar for our #FFA x @georgian_io Virtual Forum 🗓 #ArtificialIntelligence roadmaps, Q&A, and more! RSVP here: https://nitter.cf/t.co/h8nUypsjDQ
3
7
Parinaz Sobhani retweeted
Mark your calendars! 🗓The Female Founders Alliance @2point19 is hosting a fireside chat with @PariAIML and @maddie_elkhazin on "Creating an AI Roadmap" 🤩 Register now to secure your spot 👇 gptrs.vc/3zX7dGO
4
5
Parinaz Sobhani retweeted
Fantastic summary 🙌🏾 and very insightful contributions today!! Great views from academia, tech, industry and policy! We peaked at ~100 people but more engagement is needed on #FairnessInAI!! Thanks @nsrg_shah, @PariAIML @CukierWendy @VectorInst & @OCInnovation – at Toronto, Ontario
It was a pleasure to participate in a panel discussion on fairness in AI, together with very accomplished @PariAIML and @CukierWendy, brilliantly moderated by @Tjido, and organized by @VectorInst and @OCInnovation. 1/13
1
1
4
Parinaz Sobhani retweeted
We've all heard examples of unfair AI. Parinaz Sobhani @PariAIML, Head of Machine Learning and Applied Research at Georgian, shares seven principles of building fair machine learning systems 👇#machinelearning #AI gptrs.vc/3oVWQhJ
7
3
GPT-3 Hype vs. Reality: At @Georgian_io, we’re super excited about @huggingface new open-source GPT-Neo and how it performs against @OpenAI's GPT-3.
Replying to @BudhkarAkshay
TLDR is #opensource GPT-NEO model performance is very comparable to the #gpt3 on both the generative and the classification task. But seriously, check out the blog post for more details: medium.com/georgian-impact-b…
1
3
11
Can't agree more! We both equally committed and took 6 months' leave, and guess what it was the best decision to develop a strong bond with our daughter!
This Post is from an account that no longer exists. Learn more
3
Parinaz Sobhani retweeted
At @Georgian_io , we compared these two models to see if how they stack up against each other, check it out here: medium.com/georgian-impact-b…
2
13
32
Parinaz Sobhani retweeted
(1/3) Machine learning model sizes are getting out of hand, and we did our part by compressing @facebookai 's wav2vec model for automatic speech transcription in collaboration with @chorus_ai ! Check out our preprint on shrinking this bigfoot here: arxiv.org/abs/2103.15760
1
2
7