Subramanian Madhavarao’s Post

Over the recent Independent Activities Period (IAP), which is a four-week period in January between fall and spring semesters here at MIT, I took the "6.S087: Foundation Models & Generative AI" course by Rickard Brüel Gabrielsson, who is an Massachusetts Institute of Technology researcher specializing in foundation models and generative AI. I find this course quite entertaining and insightful at the same time. For the last two lectures Rickard has also invited Manolis Kellis and Artem Sorokin 🇺🇦 . Manolis is a professor of Computer Science at the Massachusetts Institute of Technology (MIT), where he heads the Computational Biology Group at MIT, and also a Principal Investigator in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). While Artem was my MIT Sloan School of Management classmate, who has recently dropped out of MIT to build his own AI start-up. He is the founder and CEO of CrackenAGI, which is focused on creating AI Agents for Cybersecurity and Digital Reliability. The course consisted of 6 lectures: - Introduction to Current AI Revolution - Exploring Foundation Models & Generative AI - In-depth Analysis of ChatGPT and Large Language Models - Insights into Image Generation and Stable-Diffusion - Emerging Foundation Models and Their Applications. Guest Speakers: Prof. Manolis Kellis & Artem Sorokin - AI Ethics & Regulation Discussion Panel with Prof. Manolis Kellis & Artem Sorokin I finally published my extensive summary of the course and key takeaways on my Substack (link in the comments). Hope you find it insightful. #AI #FoundationModels #GenerativeAI #AGI #MachineLearning #Innovation #Technology #Education #futureofai #future

Can't believe I missed the launch of this wonderful GenAI Essentials masterclass developed by ErasmusX and LifeVersity! Calling all Rotterdam School of Management, Erasmus University students and recent alumni as you can take this for free to learn more about generative AI, genAI tools, prompt engineering, and more. Next class is end of this month, so sign up now. https://lnkd.in/deyuy2H6

"We are launching LIVE Minor Program for Artificial Intelligence" Hear from our Co-founder Dr. Sreedath Panat MIT has the world’s best courses in artificial intelligence and machine learning. I developed a very strong understanding of the theory, fundamentals, and applications of various aspects of machine learning. This led to papers that I co-authored now and accepted to NeurIPS. I decided to come back to India in 2022 along with Rajat Dandekar and Raj Abhijit Dandekar, with a goal to make AI accessible to all. For the past 2 years, I have seen the way AI and ML is taught to students and engineers. I strongly believe that the way AI is taught currently needs to change radically. 1️⃣ Currently, there is a lot of focus on quickly solving Kaggle projects, copying Google Colab code given by ChatGPT and developing flashy GenAI applications. What about fundamentals? What about deep understanding? 2️⃣ How many of us can build a neural network from scratch? How many of us can build ChatGPT from scratch? How many of us can integrate scientific laws with machine learning? I really wanted to teach AI and ML the way it was taught to me at MIT. 3️⃣ For the past 1 year, we have prepared and developed the following courses: (1) Foundations of AI and ML (2) Machine Learning fundamentals (3) Building a neural network from scratch (4) Scientific Machine Learning (5) Building ChatGPT from scratch (6) Mastering Natural Language Processing (NLP) and Computer Vision (CV) (7) Generative AI fundamentals 4️⃣ The best part: all the courses will be conducted live by me, Rajat Dandekar and Raj Abhijit Dandekar. 5️⃣ You will directly learn from IITians and MIT, Purdue PhDs. We aim to make this a world class learning experience and have integrated lectures with ML, AI industry professionals too. 6️⃣ If you are looking to transition to ML, apply for jobs in ML, upskill with a deep understanding of ML, conduct ML research and stay relevant in the AI dominated world, these courses will be incredibly useful to you. 7️⃣ You can visit here and register: https://vizuara.ai/spit/ The early bird registration closes on 15 November. The batch size is limited, since we are conducting live lectures. Take a look at this video to understand more about our program: https://lnkd.in/dss33gVg

It's time to talk about AI... Competeracy Literaptitude Skilliteracy AI literacy or competency seem to be the go-to terms. I asked Gemini, and it suggested "proficiency." I like the term, but it reminds me a bit too much of my cycling proficiency 🚵♂️ I feel like the work we’re doing around AI almost needs to be split, much like Computer Science and ICT. What most teachers need is more like the ICT equivalent—focused on using AI effectively, with some technical understanding. What are your thoughts? Do you think there’s a better term, or does "AI proficiency" capture it well enough?

"We are launching LIVE Minor Program in Artificial Intelligence" MIT has the world’s best courses in artificial intelligence and machine learning. I developed a very strong understanding of the theory, fundamentals, and applications of various aspects of machine learning. This led to papers that I co-authored now and accepted to NeurIPS. I decided to come back to India in 2022 along with Rajat Dandekar and Raj Abhijit Dandekar, with a goal to make AI accessible to all. For the past 2 years, I have seen the way AI and ML is taught to students and engineers. I strongly believe that the way AI is taught currently needs to change radically. 1️⃣ Currently, there is a lot of focus on quickly solving Kaggle projects, copying Google Colab code given by ChatGPT and developing flashy GenAI applications. What about fundamentals? What about deep understanding? 2️⃣ How many of us can build a neural network from scratch? How many of us can build ChatGPT from scratch? How many of us can integrate scientific laws with machine learning? I really wanted to teach AI and ML the way it was taught to me at MIT. 3️⃣ For the past 1 year, we have prepared and developed the following courses: (1) Foundations of AI and ML (2) Machine Learning fundamentals (3) Building a neural network from scratch (4) Scientific Machine Learning (5) Building ChatGPT from scratch (6) Mastering Natural Language Processing (NLP) and Computer Vision (CV) (7) Generative AI fundamentals 4️⃣ The best part: all the courses will be conducted live by me, Rajat Dandekar and Raj Abhijit Dandekar. 5️⃣ You will directly learn from IITians and MIT, Purdue PhDs. We aim to make this a world-class learning experience and have integrated lectures with ML, AI industry professionals too. 6️⃣ If you are looking to transition to ML, apply for jobs in ML, upskill with a deep understanding of ML, conduct ML research, and stay relevant in the AI-dominated world, these courses will be incredibly useful to you. 7️⃣ You can visit here and register: https://vizuara.ai/spit/ The early bird registration closes on 15 November. The batch size is limited, since we are conducting live lectures. Take a look at this video to understand more about our program: https://lnkd.in/dJStWzDq

This past semester, Xavier entrepreneurship students created what they believe is the first language-model designed to help students and faculty use generative AI in an ethical and academically honest way. The bot's name? D'Artagnan. ⚔️ #AIAppreciationDay

To truly reap the benefits of AI, executives must embrace the need to develop a deep understanding of the technology, its use cases, ethics and more. Sertac Karaman, MIT professor of aeronautics and astronautics and director of the MIT Laboratory for Information and Decision Systems (LIDS), talks about AI Executive Academy, a new executive education course jointly offered by MIT Sloan and MIT Schwarzman College of Computing, that looks at the technical and business aspects of AI and its impact across industries. https://lnkd.in/g7kZBTtU

Is AI really "intelligent"? It's a question that's been debated since the term was coined in the 1950s. We've traditionally measured AI intelligence through benchmarks like the Turing test, chess and Go championships, and more recently, language understanding and reasoning tasks. But we’ve hit a roadblock: today's state-of-the-art AI systems have maxed out many of these benchmarks. GPT-4 achieves above human-average performance on most language tasks. AlphaGo and its successors dominate board games. Even the Turing Test needs updating. We've hit a ceiling. So what's next? We need new, more sophisticated ways to evaluate AI capabilities. Benchmarks that go beyond pattern matching and test for qualities like: - Adaptability to novel situations - Causal reasoning - Ethical decision making - Creative problem solving As educators, we should be part of this conversation. What does "intelligence" mean in our classrooms? How can we design assessments that truly capture human cognitive abilities? #artificialintelligence #aieducation

Another important impact of AI that has to be considered. AI tools require huge amount of energy and water. Do the owner-companies of these AIs, that have taken the world by storm, take any responsibility of their societal and environmental impacts? Are they being hold accountable of their negative impacts? So many questions to discuss and teach our students. (null)

What is an AI anyway? This is the title of Mustafa Suleyman's recent TEDTalk... I'm not always a fan of TED talks for some obvious reasons, but this one by Mustafa Suleyman has insightful ideas about AI that are worth discussing. Mustafa Suleyman proposes viewing AI as a "new digital species" or "digital companions" to help conceptualize the profound implications of this rapidly advancing technology. He argues that AI will become our knowledgeable, empathetic partners with an "actions quotient" (AQ) - the ability to get things done in both digital and physical worlds. As educators, there are some key takeaways we can reflect on: 1. We need to prepare for the transformative impact AI will have on education. Embracing and integrating AI tools/technologies in teaching and learning is inevitable. 2. Using the "digital species" metaphor can foster critical discussions among students about the societal implications of AI. 3. AI's increasing knowledge, empathy, and AQ present opportunities to enhance educational experiences and enable personalized learning that we should explore. 4. Promoting responsible AI development/use and educating on potential risks/ethics will be crucial roles for educators. 5. We should encourage students to envision how AI can be imbued with human qualities like empathy and creativity for the benefit of society. I found Suleyman's perspective thought-provoking. Please share your views in the comments on preparing for the AI revolution in education. #AI #EducationTechnology #DigitalTransformation #EdTech #LearningInnovation #FutureOfEducation #ArtificialIntelligence #EthicalAI #CriticalThinking #PersonalizedLearning Link to his talk: https://lnkd.in/d4VHWYvk