Aarohi Singla’s Post

Last month, PhD researcher Dyah Adila told Snorkel researchers about her work on ROBOSHOT, a novel approach to get better performance out of foundation models without fine-tuning. This work shows promise and could meaningfully impact how enterprise AI teams approach FM applications. Watch the video here: #airesearch #foundationmodels

Digital Space: 2 Things Experts Must Be Doing Today with Dr. Joybert Javnyuy https://buff.ly/4bopFLl via Joybert Javnyuy, MBA, DBA of Cosdef Global Institute for Business and Technology on Thinkers360 #AI #BusinessStrategy #DigitalTransformation

For AI to truly be responsible, there must be clarity and transparency regarding the collection of data and its use. Tabbz Maina highlighted the government’s role in setting up legal frameworks for different scenarios, ensuring inclusive AI accessibility, and using it for positive impact.     Converge 4 updates are around the corner, so stay tuned!    #WeConvergeWeConverse 

NotebookLM is starting to look more like the note system I've always dreamed of 😁 and especially the Readwise highlights use case demonstrated in this video. I'm still using Obsidian as my primary knowledge repository, but giving NotebookLM a serious try now (maybe it's not a replacement anyway) after watching this excellent introduction by Dan Shipper and Steven B Johnson. #Readwise #AI

The next chapter is out: How Attention works https://lnkd.in/esqVS7fk

ANOTHER TYPICAL EXAMPLE OF YANN LECUN BEING COMPLETELY OUT OF HIS DEPTH - HE ADMITS HE HAS AGAIN NO IDEA, FOR HOW MUCH LONGER MARK ZUCKERBERG? "something we have no idea how to solve, at least I have no idea to solve is can we get machines to learn hierarchical representations of action plans." Yann LeCun THE 3 CHALLENGES "1. getting machines to learn to REPRESENT THE WORLD & proposing self-supervised learning 2. getting machines to REASON IN WAYS that are compatible with essentially gradient based learning 3. can we get machines to LEARN HIERARCHICAL REPRESENTATIONS OF ACTION PLANS" OPEN QUESTION Yann admits in this interview he has again no idea on "How to solve intelligence?", for how much longer Mark Zuckerberg? EPILOGUE Good gracious Yann LeCun , you are so embarrassing, please speak only for yourself. What are all your rewards & prizes worth? MY2CENTS What is learning? 1. ACCUMULATION IS NOT LEARNING, only accumulation, see education in school. 2. WE LEARN ONLY FROM OUR OWN EXPERIENCE, therefore we need: a. a BODY, with sensors, actuators & organs for "own needs". b. a SHORT-TERM MEMORY, in order to accumulate step by step new information c. a SELF-LEARNING MECHANISM, in order to link & integrate the accumulated information with the actual knowledge and store it into long-term memory d. the FEEDBACK is an absolute requirement for learning, as it proves via experience, IF our actions have success, or not e. a SOCIAL ENVIRONMENT is required, as we can act & react within a society & subsequently accumulate knowledge from others f. LEARNING HAS EVOLVED DURING EVOLUTION, therefore we should study the evolution of the brain, from insects via mammals/fishes/birds to primates. LINKS NEUROSCIENCE https://lnkd.in/d9H4SWxC NEUROROBOTICS https://lnkd.in/g_55qvFX NEURAL NETWORKS ARE IDEALLY SUITED TO IMPLEMENT KNOWLEDGE AND THINKING - TOO BAD OR THANK GOD MOST "AI" PEOPLE DON'T SEE THAT https://lnkd.in/eTVd9A2J A COGNITIVE FRAMEWORK ENABLES COMPLEX BEHAVIOR, REGARDLESS OF ITS LEVEL OF INTELLIGENCE - THAT IS A NERVOUS SYSTEM BASED ENTIRELY ON SPIKING NEURAL NETS https://lnkd.in/dJBsEzen SOME OF US QUESTION EVEN THE DEFINITION OF AGI - THE POINT IS WE HAVE TO CATEGORIZE INTELLIGENCE BY TAKING EVOLUTION INTO ACCOUNT https://lnkd.in/dhSJTby4 THE ROAD TO AGI ALSO IMPLIES THE DECISIONS HOW TO IMPLEMENT BIOLOGICAL FEATURES IN THE DESIGN OF INTELLIGENT SYSTEMS https://lnkd.in/diueBsGb AI IS NOT MATH - CONTRARY TO THE OPINION OF YANN LE CUN & CO https://lnkd.in/ec_XFZwp FALSE COMPARISON: MATH & INTELLIGENCE = APPLES & ORANGES https://lnkd.in/dMPEbFq5 YANN LECUN PROVES WITH EVERY NEW POST HIS INABILITY TO UNDERSTAND BIOLOGICAL INTELLIGENCE, LET ALONE DUPLICATE IT - HOW LONG WILL MARK ZUCKERBERG TOLERATE HIM? https://lnkd.in/dAUjYRfD PAYING HOMAGE TO YANN LECUN - IS THIS MEGALOMANIA OR SIMPLY FALSE MODESTY? https://lnkd.in/dB8EXUVf

🚀 Ever wondered if AI can truly "think" like humans? 🧠 Yann LeCun, a pioneer in AI, sheds light on this intriguing question! LeCun explains that current Large Language Models (LLMs) excel at tasks that require instinctive, automatic responses—what he calls System One thinking. Examples include simple pattern recognition and routine tasks. However, when it comes to System Two reasoning—tasks that require deliberate, planned thought—LLMs fall short. These include complex decision-making and strategic planning, where a deeper understanding and more thoughtful consideration are needed. The current models predict the next word in a sequence, which works well for generating text but lacks the depth needed for true reasoning. They are not capable of advanced planning or optimizing their answers in an abstract, meaningful way. LeCun envisions future AI systems overcoming these limitations through energy-based models. These models will measure how well an answer fits a prompt, optimizing responses efficiently and thoughtfully. This approach promises a significant leap forward in making AI more capable of reasoning like humans. 🌟 Curious about AI's capabilities and limitations? Watch the video and join the conversation!

For all my Strategy friends out there- I found this presentation by Zoe Scaman to be the most helpful and inspiring guide yet on how we can use AI to make our work and ourselves smarter, prettier, and more impactful. Do not miss it- and please let me know if there are any tools you're using that I might not know about yet, or that she doesn't reference herein. https://lnkd.in/gchk3XGn

Does the debate between open-source and closed-source AI still matter today? In this episode of AI Rising, hosts Leslie D'Monte and Jayanth N Kolla explore this with Sayandeb Banerjee, CEO of MathCo. They discuss resistance to open-source AI, the rise of Scalable Language Models (SLMs) like Gemma 2, and how its portability sets it apart. Get insights on AI's real-world applications and what it means for businesses looking to scale with AI-driven solutions! 🎧 Listen Now : https://shorturl.at/aQyPs #HTSmartcast #AIRising #OpenSourceAI #ClosedSourceAI #ScalableAI #DataDriven #AIModels #BusinessIntelligence #MachineLearning

AI for Strategists A ton has been written about how AI can be incorporated into the workflow of strategists. Of all the stuff I have seen, Zoe Scaman’s deck stands apart. What I like about Zoe’s guide is how it explains the role and use of different tools for each stage of a strategist's workflow. According to Zoe's analysis - Chat GPT plays a fundamental and foundational role- provided you use it right- think "conversations" more than "prompts", but she mentions many other tools that are useful for specific elements of strategy work- including Springboards.ai. Net- Comprehensive and useful Not to forget to mention that it's also a supreme act of generosity- there's a lot of work that's gone into this from the research and testing to the building of the presentation.