Shadi Hamdan retweeted
interested in 3D vision, motion, driving, or robotics? consider applying to our graduate programs. the spring deadline is tomorrow.
we do our best to provide a research environment where students can reach their potential and publish in top-tier AI venues.
we regularly collaborate with leading AI labs worldwide and offer EU-level scholarships (+dorm/housing support), solid GPU resources (a decent local cluster + access to supercomputers like Leonardo and MareNostrum), and a fun, supportive lab environment :)
apply.ku.edu.tr/
Important: one of my very dear friends, Mohammed Alsaqqa, is raising funds to help evacuate his family from Gaza. Mohammed is not only a great friend, but also an exemplary person and one of the brightest people I know. The past few months have been hell on earth for him and his family, as well as the millions of displaced Gazans. The situation has only gotten worse with time, and unfortunately now we see no other option but to help get his family out of Gaza.
The most critical situation at the moment is that of his father, Khalil (62), who suffers from critical health issues, including chronic high blood pressure and diabetes. The lack of essential medications and adequate nutrition in Gaza puts his life in constant danger. Please take a look at Mohammed thread, and consider supporting his fundraiser. Campaign link is within the thread, and if you can't donate, sharing is also caring.
Thank you.
I am reaching out to you to support a crucial cause: helping my family members who are in dire need to find safety and security away from the hardships they're facing in Gaza.
1/5🧵
#Gaza #GazaIsStarving #Palestine #Help #GazaStrip #fund #gofundmedonations #support #diabetes
Shadi Hamdan retweeted
Delighted that our paper "RbA: Segmenting Unknown Regions Rejected by All" has been accepted at #ICCV2023
paper: arxiv.org/abs/2211.14293
project: kuis-ai.github.io/RbA/
code: github.com/NazirNayal8/RbA
with @misrayavuz, João F. Henriques, @ftm_guney
we can now talk about “RbA: Segmenting Unknown Regions Rejected by All” accepted at #ICCV2023.
we found that when we find and classify regions with Mask2Former, each object query focuses on a known class. so we can find unknowns as pixels rejected by all, hence the funny name 😃
Shadi Hamdan retweeted
📚 Thrilled to have attended the MLSS^S 2023 in Kraków, Poland! 🇵🇱🤩 This summer school was an invaluable experience, introducing us to state-of-the-art topics in #MachineLearning and its applications across fundamental sciences.
This month, I had the pleasure of attending #ICVSS2023 and spending a week in Sicily, Italy! It was an intense week full of inspiring lectures by some of the best in the field! It was great to meet so many others that share similar interests in such a fun environment.
Shadi Hamdan retweeted
Modelling the underlying 3D scene structure greatly improves unsupervised multi-object segmentation. Check out our method for estimating segmentation mask and depth at the same time, accepted at #ICCV2023:
kuis-ai.github.io/multi-obje…
great work by Sadra @__Everwake__ @KuisAICenter
Shadi Hamdan retweeted
we can now talk about “RbA: Segmenting Unknown Regions Rejected by All” accepted at #ICCV2023.
we found that when we find and classify regions with Mask2Former, each object query focuses on a known class. so we can find unknowns as pixels rejected by all, hence the funny name 😃
Shadi Hamdan retweeted
Happy to see that @OpenAI used our Chart Question answering benchmark (arxiv.org/pdf/2203.10244.pdf) in evaluating their newly released GPT-4. Here's GPT-4's output for a ChartQA sample (cdn.openai.com/papers/gpt-4.…). Nice explanatory answer!
@JotyShafiq @Ahmed_Masry97 @dxlong2000
This quoted post is unavailable.
Shadi Hamdan retweeted
💫We're happy to share that our MS student @ozsoyserdar and faculty members @Alper_T_E and @denizyuret presented posters of the two papers at #NeurIPS2022 in USA, and our MS student @BozkurtBariscan made a successful virtual presentation in #NeurIPS2022 Oral Session.
Congrats!👏🏼
Very happy to announce our latest paper which has just been accepted to #NeurIPS2022! Very grateful to all my collaborators @ozsoyserdar @denizyuret @sercanarik and @Alper_T_E 😃
Immensely thankful to the Abdulla Al Ghurair Foundation for Education @AGFforE for their continuous support throughout my undergraduate education. Special thanks to all my fellow peers and mentors, without them none of this would have been possible :)
Five years ago, Shadi Hamdan set out on a journey at @kocuniversity in a different & new environment.
Today, he is ranked 1st at the university & is the 1st international valedictorian ever since its founding.
We are very proud of Shadi and his outstanding achievement. #AGFE
Shadi Hamdan retweeted
Üniversite mezuniyetlerindeyim; yeğenlerim, öğrencilerim için. Koç Üniversitesi’nde bir ilk...
İlk defa bir uluslararası öğrenci üniversitenin birincisi oluyor.
Bilgisayar Mühendisliği’nden mezun, geleceğin AI uzmanı Suriye - Filistin kökenli Shadi Hamdan’ı yürekten kutluyorum.
Shadi Hamdan retweeted
After 2 years of work by 442 contributors across 132 institutions, I am thrilled to announce that the github.com/google/BIG-bench paper is now live: arxiv.org/abs/2206.04615. BIG-bench consists of 204 diverse tasks to measure and extrapolate the capabilities of large language models.
Shadi Hamdan retweeted
For the past few months, @mmitchell_ai, @YJernite and I have been working on the exploration of popular NLP datasets.
Here are some fun things we discovered!
Shadi Hamdan retweeted
1- Letters of recommendation play a big role in the review process. Sure, there’s GPA, awards, SOP & maybe papers. But for me (& many colleagues), a single great letter >> everything else. Interestingly, 1 bad letter doesn't negatively impact a student's chances as much
Shadi Hamdan retweeted
You may have heard that algorithms don't just reflect bias, they amplify it
Here's a very visual example of that...
We generate portraits that are "closest" to a given text prompt🤳
Here, the prompt is "terrorist" and the output isn't a bearded guy, or a hijabi...
but BOTH😬
















































