During the last few days, I had the privilege of mentoring an international competition in computer vision organized by Prof. Paul Feigin and the Israel Data Science Initiative (IDSI) at the Technion - Israel Institute of Technology, collaborating with Helmholtz Information & Data Science Academy. It was satisfying and rewarding to help teams of computer vision experts solve a semantic segmentation task and overcome barriers in utilizing deep learning for a real-life application. As a mentor of all the 7 groups, I addressed questions and problems that limited their progression—ranging from providing insight into research dilemmas and algorithm designing to guiding how to debug some technical, DevOps-related issues. The wide variety of approaches, tools and models was fascinating! I want to thank all our outstanding competitors for fertilizing fruitful discussions from which I have learned a lot. Congratulations to the winning team: Yi Wang and Chenying Liu, from the German Aerospace Center (DLR) of Data Science!🏆🥇 Check out their open-sourced solution on GitHub: https://lnkd.in/dbtBAhMX
Nadav T.’s Post
More Relevant Posts
-
AI researchers/entrepreneurs in Boston – join an intimate mixer event with VPs of ML and top executives from leading companies like Wayfair, State Street, TD Bank, HubSpot, DataRobot, Liberty Mutual Insurance, and Vertex Pharmaceuticals. On June 7th, 6-8pm. For more details, DM me or Glenn Ko. Limited spots available.
-
Happy to be part of the 💫 #StarCoder team and contribute to open-access AI. https://lnkd.in/eXqUCnmg
So this week we've finally released StarCoder (https://lnkd.in/eSU9Tfst) StarCoder is the first large model (15B) which is both *high performance* (beating the like of PaLM, LLaMa, CodeGen or OpenAI code-crushman-001 on code generation) and also trained *only* on carefully vetted data (only permissive code licences, comprehensive PII removing). If you need a clean and high performance code generation model, check it out. Even though it can act as a chat model with a dialog prompt, it's not a real chat model though as we haven't done dialog finetuning yet so don't use it like ChatGPT. You should rather use it as the basis for internal finetuning or for code completion e.g. with the vscode extension https://lnkd.in/eAbRAyQ3 Happy coding!
-
-
The #LLMs for Code Seminar is launching this week! 🧠 Join us online: https://lnkd.in/daazWcYi #AI4Code #LLMs4Code
-
-
How can a programming problem be solved without writing a single line of code? Last week, as part of lablab.ai's hackathon about OpenAI's #Codex, I had the opportunity to talk about code completion tools like GitHub Copilot and to present a live demo of typing natural language comments to create a Python solution to the daily LeetCode challenge. #AI4Code #LLMs4Code
We are all using AI to revolutionize the world of intelligent applications, and happy that lablab.ai community has real success there 🚀 In just 7 days, 259 teams used OpenAI tools to create innovative solutions. It’s been an incredible journey with a total of 2653 people building amazing applications, and 54 submitted projects we have got at the end 🏄🏼♀️ (you can check all of them on events page) This is just the beginning of the limitless potential of AI innovation. We encourage everyone to continue building and innovating with AI. You already know the finalist, congratulations guys once again 🏆 Thank you all for participating and having such a productive week together 🦾 Mathias Asberg Arjun Patel Pawel Czech Simon Olson Olesia Zinchenko Elizabeth Marchuk Anastasiia S Adam Raźniewski Nemo Shi, Ed.M. Wilbert Osmond Ervin Moore Omar Atef Sesa Skander K. Nadav T. Sherry Horowitz - Brand Identity Designer Mohammed Arsalan Fabian Stehle Simon Olson Stay updated and join us for the next Hackathon in 2023 🎅🏼 #innovation #ai #openai #machinelearning
-
The #LLMs for Code Seminar is about to launch! https://lnkd.in/daazWcYi Our first speaker is Jacob Andreas from MIT (abstract below). Register for his talk at the link in the comments. #AI4Code #AI4SE
-
-
The group I led in the OpenAI Whisper Hackathon ("The Future of #Speech_Recognition") won 3rd place!🥳 We have developed a state-of-the-art #speech_to_text engine for people who stutter. While creating our proof-of-concept over a weekend, we collaborated with the Israeli Stuttering Association NGO and learned about their needs. We found that adaptations in the AI algorithm (like customed sampling) can make it more inclusive. A Telegram bot we implemented allowed anyone to transcribe voice messages using our facilitating deep learning model. It was a small reminder for ourselves (and hopefully the AI community) that inclusion is reachable.
We did it! Another fantastic and inspiring hackathon at LabLab.ai 🚀 Thank you to all 1700 participants🧑💻 for creating innovations with AI. Last weekend we had more than 20 incredible submissions, and it was tough to choose just three winners 🤩 🏆 In the first place, we have Team Blue with their project, "InvestogAId" 💻 InvestogAId uses OpenAI Whisper and GPT-3 to create an automated transcription tool to watch your favorite stock trading videos and implement your strategies. Nikhil Sehgal Martyn Ben Ami Lito Saro Antonello Lovito Sebastian Neri Breeze-ops Kudos, Team Blue! 🥈 In the second place, we have Team Biscoff with their project, "Moriarty" 💻 The solution to this problem is to utilize each user's microphone to assess whether his speech is obscene, toxic, threatening, insulting, etc., using cutting-edge Machine Learning tools. Spiros Chatzigeorgiou Christina Athanasiadou Way to go, Team Biscoff! 🥉 And in third place, we have Team Boss with their project, "Butter." 💻 Butter is a chatbot that helps people who stutter. It uses AI-based speech-to-text conversion to output messages from live voice recordings accurately. It also answers personal questions regarding stuttering. Nadav T. Rimjhim Singh Neal Sharma Bensis Congratulations, Team Boss! Thank you again to everyone who participated. We can't wait to see what you all come up with next! 🚀 #Hachathon #ArtificialIntelligence #Programming #machinelearningengineer #computerengineering #AIHackathon
-
Watch my talk with Peter Henstock, ML Lead at Pfizer and Harvard University lecturer, about current trends in deep learning. We covered some bottlenecks, breakthroughs, and opportunities for businesses. Thank you, Ramesh Dontha from Digital Transformation Pro, for hosting us in your great webinar. https://lnkd.in/e_KyZhac
-
As a child, although I participated in programs for gifted students, I was bored most of the time at school. That's why I was happy when I had the opportunity to volunteer today to professionally guide, for the second year in a row, the most talented boys/girls in Israel at the Israeli Mathematics Olympiad held at the Weizmann Institute of Science. I wish I had this kind of a challenge as a pupil. Think you can beat them? Test yourself! 🤓
-
AI Researcher in LLMs | PhD student, Weizmann Institute | Expert Advisor, MIT Sandbox
2yTomer Sidi, Tomer Fried, Gerardo Diaz