Rubin Observatory’s Post

Speaking of ethics in the context of AI reminds me of Ambedkar, who wrote our (India) constitution, stating that at the end of the day (after all debating is done and penned), it's the morality of the people / society, and not that of the written text, that has the last word, so to speak. With being good, difficult as it is, becoming popular, penal system will rest in peace, or so I seem to think. Speaking of morality, here's Bilgrami on its romance: https://lnkd.in/gDE-u4vi

"The telescope made us discover that we were not the center of the solar system. Scary thought at first, but we became accustomed to it, and it made us realize the beauty of the cosmos and our modest place in it. Similarly, AI made us discover that intelligence extends beyond humanity, and that we don't have the monopoly on intelligence. Scary thought at first but we will soon realize our role, as just one part of a broader intellectual tapestry. Maybe then, we will be worth our planet." super interesting and eyes-opening lesson on Generative AI today with Hoss - Zawali from Stanford University Graduate School of Business. thanks The European House-Ambrosetti for organizing and inviting me to attend!

What if the architects of the social media revolution had engaged the world’s leading sociologists, political scientists, and educators to evaluate the potential impact of the technology before it was deployed? With so many conversations focusing on the future of AI, getting ahead of the game now is imperative so we have a pathway to responsible AI. And that's what we're hoping to address at #SXSW this coming March. Please support us by casting your vote so you can join the discussion with Joel Fischer, Professor of Human Computer Interaction at the University of Nottingham and Research Director of Responsible AI UK, Lisa De Bonis, Global CEO Huge, and yours truly representing Genie. Link to vote is in the comments below 👇

Sharing this invitation with my network in #London. Join this C-level conversation with distinguished Professor Aaron "Ronnie" Chatterji on the critical topic of the implications of #DeepTechnology, #AI and #Innovation on #business and #society.

Happy Lunar New Year! Recently, netizens have been both shocked and appalled by the flood of X-rated, AI-generated images of Taylor Swift on X (Twitter). However, many may not realize that with the advancements in personalized generative AI, it's now possible to create such images of not only celebrities but also any individual (e.g., yourself) with just a few photos. In my latest blog post, I outline various strategies to combat this issue, as well as ways these measures can be circumvented. So, why not take a break from Lunar New Year celebrations and weekend activities and read the post and learn more about addressing this critical concern :D Link to the post: https://lnkd.in/gCFC22qg

Ready for a spectacle almost as breathtaking as the #SolarEclipse? 🕶️ Highlighted in green are the states buzzing with solar eclipse excitement, precisely where our AI technology, IntentKey, detects intense relevance to the celestial event. The concept's interest surged by a staggering 73% in just the past week and a whopping 98% in the last month! 📈 And guess what? It aligns seamlessly with the path of totality, making perfect cosmic sense! Peek closer, and you'll notice California and Florida glowing with high indexing too, indicating bustling activity around the eclipse. Could this mean residents are gearing up for an unforgettable journey to the path of totality? This is the power of real-time insights at work. When we say IntentKey AI uncovers your prime audiences precisely when they matter most, we're not kidding! Explore more on this model or create your own using our FREE tool, the Audience Discovery Portal: https://lnkd.in/gFfDPgZz #pathoftotality

Always nice to be asked to share my thoughts on AI. I'm sometimes asked, how did a theoretical chemist get into AI. This used to puzzle me, as modern neural networks were invented, in part, by theoretical chemists and related applied mathematicians and physicists. Some people who inspired me to move into this field: Jack Cowan : Applied mathematician (University of Chicago). Co-Invented the Wilson-Cowan mathematical model of spiking neurons (and brought the sigmoid function into computational modeling of neurons). And personal friend of my PhD advisor. Christopher Longuet Higgins: Quantum chemist (University of Chicago). Doctoral advisor to Geoffrey Hinton. Started the school of Machine Intelligence. Very famous guy from my department. John Hopfield: Professor of Physics, Chemistry, & Biology at different times. Invented the associative neural network in 1982, now known as the Hopfield Network or Hopfield Associative memory. He spearheaded 100s physics papers on the theory of NNs. Today, there is so much capital flowing into AI research that's impossible to keep up, but I try to bring to bear my background in theoretical chemistry and physics to make a small dent in a huge field. I hope the work and thew weightwatcher tool (which now has 125K downloads!) has been useful to you.

✨ I'm thrilled to have attended the 26th International Conference on Human-Computer Interaction this week. 📄 🔍 Our paper, "DATA SINS: Speculative Design Unveiling Data Colonialism Through AI Imagery," aims to spark crucial conversations within the design community about the power dynamics embedded in the infrastructures supporting emerging technologies, particularly generative AI. Additionally, we explore strategies to creatively envision alternative futures without reinforcing historical cycles that jeopardize both the planet and human autonomy. #hcii24 #hci #speculativedesign #criticaldesign

🚀 Excited to unveil the fusion of the Shadow Clown Dare model with the BioMistral Dare model: The Shadow-Clown-BioMistral-7B-DARE. Engineered to supercharge AVG with vital medical insights, this model not only raises the bar but also outperforms benchmarks, including the esteemed Mistral 7B Instruct V2, with evaluations conducted using the Eleuther AI Language Model Evaluation Harness: - Average: 71.21 - ARC: 67.41 - Hellaswag: 86.78 - MMLU: 64.07 - TruthfulQA: 67.68 - Winogrande: 81.61 - GSM8K: 59.74 And the innovation doesn't stop there. This model is finely tuned for efficiency, boasting quantization and nested quantization capabilities that slash inference consumption from 14.6 VRAM to a mere 4.8 VRAM. For seamless integration and further exploration, check out the model's Hugging Face repository: https://lnkd.in/dSajXXFr Ready to witness its prowess firsthand? Dive into testing on Colab with a T4 GPU through this notebook link: https://lnkd.in/dj8mYy2S Join me as I redefine possibilities in the realm of AI and medical applications. #AI #MedicalAI #Innovation #MachineLearning #DeepLearning #LLM

When it comes to understanding the underlying nature of the world and observing emergence, generation from videos might not be the key players. What truly matters is making predictions in an abstract representation space, steering away from the constraints of the pixel space. #generatieveai #foundationmodels #emergence https://lnkd.in/gjeTPFqj