IUJ🍀
@demirbasayyuce
Followers
488
Following
18K
Media
126
Statuses
1K
Machine Learning Engineer & Published Researcher, BSc in Computer Engineering | Aspiring PhD, GSoC @TensorFlow, Google Dev Expert in ML
www.goodreads.com/ayyuce
Joined October 2023
I'll be presenting on Vision Language Models in Healthcare at Google's International Women’s Day 2025 event. 😇.
1
0
3
I love my job, keeping myself updated and constantly improving myself by learning new stuff.
4
8
227
Preserving Multi-Modal Capabilities of Pre-trained VLMs for Improving Vision-Linguistic Compositionality .This paper ( proposes a new method called Fine-grained Selective Calibrated CLIP (FSC-CLIP) to improve the compositional understanding of pre-trained
0
1
11
Enjoying some chocolate and salep while preparing lecture notes for my Vision-Language Models presentation on February 15th.
2
0
9
Hello, this is Ayyuce (IUJ). I'm looking for a PhD in machine learning. I hold a BS in Computer Engineering and have ML papers & projects. My current focus is on Vision Transformers and Vision-Language Models. Looking for a dedicated PhD student? Please DM me. #CVPR2024.
3
0
8
Disease-informed Adaptation of Vision-Language Models (. The paper proposes a novel method to adapt Vision-Language Models (VLMs) for medical image analysis, specifically targeting challenges related to diagnosing new diseases with limited data. This
1
0
7
I'm currently reading Hands on Large Language Models. Once I finish this, I will read Deep Learning for the Life Sciences.
0
2
7
If you want to visualize an ML model, I strongly recommend It shows all the details of your model.
0
1
7
How to built a Simple Retrieval-Augmented Generation (RAG) Pipeline
0
2
6
I'm on page 220 of #NexusBook. I can summarize what I've read so far as follows: Harari defines information as not just data but as something that connects people, creating networks essential for societies. He critiques the naive view that more information necessarily leads to
It's Friday! What are you reading this weekend?.
1
0
5
I have only managed to read the first 100 pages of the book so far. In these pages, Harari delves into how human societies have built, relied upon, and been influenced by information networks throughout history. Harari introduces the concept of information as a binding force in
It's Friday! What are you reading this weekend?.
0
0
5
Towards Robust Radiogenomic Models for Brain Tumor Characterization.
0
1
5
The MedSAM-2 paper ( proposes a novel segmentation model for medical images that extends the SAM 2 framework. This model addresses both 2D and 3D medical image segmentation challenges by treating medical images as videos. 1. Treating Medical Images as
0
1
5
MedCLIP - Contrastive Learning from Unpaired Medical Images and Text.
0
1
5
If you want to try LLaVA-Med on Google Colab, you need Colab Pro. The model has a high memory footprint.
1
0
5
Working on a brain tumor segmentation project (actually, it’s an object detection dataset) using #PaliGemma. I’m gonna share the jupyter notebook file. 😇
1
0
4
Here is a Rotary Positional Embeddings (RoPE) TensorFlow implementation.
1
0
4
This is a street cat who hangs out at Starbucks. 🐈💞 I love cats. I just submitted my application to be an emergency reviewer. @acl_srw @NickATomlin @enfleisig
This is me with Davis the cat at Oakland Cat Town! . Definitely consider being an emergency reviewer for the ACL Student Research Workshop 🥰
0
0
4
The Silk Roads
It's Friday! What are you reading this weekend?.
0
0
3
Humanity’s dreams of artificial intelligence go back much further than we might think. In Homer’s Iliad, you can find references to self-moving chairs called tripods.
0
0
3
There’s a cat on my yoga mat. Guess my workout will have to wait.
1
0
3
AI Godmother Fei-Fei Li Has a Vision for Computer Vision Her startup, World Labs, is giving machines 3D spatial intelligence.
0
0
3
Currently attending a fascinating online event, From The Library of Sigmund Freud: Krafft-Ebing's Psychopathia Sexualis, hosted by @FreudMusLondon
0
0
2
I've read 30 of 35 books for the Goodreads 2024 Reading Challenge! What should I read next? via @goodreads.
0
0
2
Eight years ago, when I was 18, I was awarded the Future Technology Star Candidate award by Microsoft Turkey. This video was recorded by my father at the award ceremony. I just found it in my gallery, and watching it took me right back to that day. It was one of the best days of
1
0
2
PaliGemma uses SigLIP image encoder, and here is a blog post about it
1
0
2
Making a cup of coffee at 2 AM to stay awake for late night studying. ☕️
0
0
2
My latest paper received its first citation. It's good to know someone read my research. I'm grateful for the recognition and looking forward to more research collaborations.
0
0
2
t-SNE visualization of patch embeddings helps to understand better how vision transformer models represent image data. Here is an implementation:.
0
0
2
The Language Instinct by Steven Pinker
It's Friday! What are you reading this weekend?.
0
0
2
I’m sharing my notes on Diffusion Models
1
1
2
Built a vision-language model from scratch for image captioning using TensorFlow. Check out the code here:
0
0
5
If you are curious about how sigLIP and CLIP works, you can check out this notebook .
0
0
1
@DeryaTR_ Also, I would be delighted to offer you a cup of one of my unique coffee blends if you let me know when you come to Turkey.
0
0
1
LipsFormer: Introducing Lipschitz Continuity to Vision Transformers. Paper: Code:
0
0
1
Spread the love for open source with #Hacktoberfest, a month-long celebration of open-source projects, their maintainers, and the entire community of contributors.
0
0
1
ColPali: Efficient Document Retrieval with Vision Language Models.
0
1
1
“Based on what you know about me, draw a picture of what you think my current life looks like.”
0
0
1
The rate at which information becomes outdated, especially in the field of machine learning, is incredibly fast, almost daily. Given this situation, I find it absurd that academic journals in this field still require peer review and make authors wait almost a year for.
0
0
1
@anxious599 Here is the image captioning VLM that I'll be using for my presentation. I'll share the slides in the coming days.
Built a vision-language model from scratch for image captioning using TensorFlow. Check out the code here:
1
0
1
@minsuk_chang Hi, I would be happy to help with reviewing. I recently reviewed two LLM papers for ACL-SRW 2024, and I’m available over the next couple of days. Here is my Google Scholar page: Let me know if I can help. Thanks.
1
0
1
@DeryaTR_ This is really interesting. But I have to say, back in my first year at university, I once ended up in the hospital after drinking several cups of coffee to stay awake and study all night before a final exam. Since that day, I limit myself to just one cup a day. By the way, I’m.
2
0
1
Preparing a latte macchiato but there is no milk option between 270ml and 340ml. I can't get the right amount of coffee for my cup.
0
0
1
@xtmmwrr Absolutely. Gossip is a great example of how information in Harari's sense goes beyond mere facts to build social bonds. As Harari discusses, shared stories, whether myths, news, or even gossip create networks by linking people through common narratives. It is less about factual.
1
0
1
MoE-LLaVA: Mixture of Experts for Large Vision-Language Models. This paper ( proposes a novel approach to enhance the efficiency and performance of Large Vision-Language Models (LVLMs). The key contributions of the paper include:. · MoE-Tuning Strategy:
0
1
1
@ASophiaKoepke @TU_Muenchen @MunichCenterML Dear Professor Koepke, I am really interested in this opportunity. I hold a Bachelor's degree in computer engineering and have experience in computer vision and multi-modal learning with published papers and projects. Although I don’t have a Master’s degree, would it be possible.
0
0
1
This one (The Turkish edition of.Electronic Devices and Circuit Theory)
Which book scared you the most?.
1
0
1
First PayPal, then Instagram and Roblox. What's next? They even banned GitHub for a while, if my memory serves me right it was 2016. We will be using abacuses soon. #Turkey #FreedomOfExpression
1
0
0
There is nothing better than drinking linden tea brewed from the linden flowers I picked with my own hands while working.
0
0
1
@microth Hi! I'm really interested in this opportunity. I hold a Bachelor's in computer engineering and have experience in machine learning, with published papers and projects. Would it be possible to apply for a PhD position with this background? Thanks!.
2
0
1
Here is my TensorFlow implementation of Rotary Position Embedding for Vision Transformers. Check it out on GitHub:.
0
1
1
A Path Where No Man Thought:.Nuclear Winter and the End of the Arms Race
It's Friday! What are you reading this weekend?.
0
0
1
I have a lot of work to do, but since there's no milk at home, I can't make coffee, and since I can't drink coffee, I don't want to work. I might as well sleep.
0
0
1
I’m going to see #Macbeth at the theater in a few days, so I decided to read the play beforehand. While it tells the story of a man's ambition for power and the series of murders he commits to achieve it, I would like to share some details that caught my interest. When the.
0
0
1
Conversations with Neil's Brain. It explores how the brain works, focusing on thought, language, and consciousness. Using the story of Neil, a composite epilepsy surgery patient, the book explains how different brain regions control functions like speaking, understanding, and
Is there a book that’s not that well known/popular that you wish more people would read?.
0
0
1
The Personal MBA: A World-Class Business Education in a Single Volume
It's Friday! What are you reading this weekend?.
0
1
1
I’m looking for academic journals or conferences to review papers in the fields of deep learning, vision transformers, and vision language models. I’m able to review papers in English and Turkish. Any recommendations?.
0
1
1
I love this: Vision-Language Models (VLM) at the Edge by MJRoBot (Marcelo Rovai) in Raspberry Pi and Machine Learning & AI
0
0
1
@TradeTexasBig Yes, you can find it here.
Here is my TensorFlow implementation of Mixture of Experts.
0
0
1
Oh god, I’m in love with computer architecture and organization. 😍
1
0
0
After reading that thought can be a separate process from language, contrary to common belief, I came across this paper while researching various ways to enable reasoning in ML models other than multimodal models. I'll read it as soon as possible. Also,.
0
0
1
I’m joining @Google's AI Sprint. I’ll be working on fine-tuning PaliGemma on a brain tumor VLM dataset. The first thing to do is to create a dataset with images and radiological reports. @GoogleDevExpert
0
0
1