Rob Ferguson’s Post

🌍 Excited to be part of the solution! 🤝 Join me at DataKind's #DataDive event, where we're diving deep into the challenges of our time: climate change and global access to financial services. At this critical juncture, harnessing the incredible power of science and AI is more vital than ever. Together, we're working towards a brighter future for all. 💡 Let's create a world where technology serves humanity's greatest needs. 🌟 #DataKind #AIforGood #ClimateAction #FinancialInclusion #TechForGood. https://lnkd.in/di5yZBfz

At #CivicSabha, I had the honour of moderating this incredible panel on 'Thriving Together - Scaling Open Data Collaboratives', here are some of my key takeaways: 1) Open Data Collaboratives can help prioritise resources and re-use better, helping us shine light on root causes and catalysing systems change. Claudia Juech from Bloomberg Philanthropies gave examples of how various Latin American countries like Buenos Aries, Brazil & Mexico City are sharing best practices as part of their City AI Connect program. 2) Open Data Collaboratives bring cross-sector collaborations and enhance sustainability, equity and efficiency by better sharing of existing knowledge and resources. Deepali Khanna from The Rockefeller Foundation shared examples of Lacuna Fund and ARTPARK on how they are co-creating reusable training datasets for improving public health & climate action. 3) We are witnessing Data Cooperatives, Data Trusts, Data Exchanges and other forms of data-led collaborations focusing more on use-case driven approach to unlock value from data. Sushant Kumar from Omidyar Network India mentioned that such collaborations have potential to improve Air Quality & Coastal Community Resilience. 4) Multi-stakeholder collaborations are essential to make #OpenData help grassroot communities, otherwise data platforms are at the risk of becoming data graveyards. Subrat Das from Ford Foundation shared his experience of conceptualising and co-creating Open Budgets India along with multidisciplinary teams from Centre for Budget and Governance Accountability (CBGA) and CivicDataLab. Thanks to the support of The Rockefeller Foundation for helping us make #CivicSabha a diverse & engaging celebration of data for public good ecosystem. Stay tuned, we will keep sharing our learnings from the event, will release videos of panel discussions, workshops and other documentation shortly.

In the darkest of nights, the dawn of possibilities awaits.🇵🇸 I am pleased to share that our team, LeakGuard, won the second place at a recent AI hackathon held online from December 8th to 10th. It was intense three days of attending insightful sessions, brainstorming ideas, and tirelessly developing our conducted idea. What emerged from this collaborative effort is a promising startup focused on early leak detection using AI, a game-changer in minimizing water damage for sustainable practices. Our mission goes beyond just detecting leaks; we're dedicated to ensuring a fair distribution of water to all subscribers. In a world where water conservation is crucial, we aspire to make a meaningful impact.💧 #LeakGuard #AIInnovation #SustainableWaterFuture

If you are heading to today's Applied AI Fall 2023 Conference in St. Paul, grab a seat at 10:35. Our own Chief Experience Officer Jason Tell walks through a Modern Climate case study of creating AI-generated imagery for a client. He shares his favorite tools and prompts, and passes along lessons he's learned. If you are there, we would love to hear your thoughts! #appliedai #MNtech #TwinCitiesTech #aiconference #artificialintelligence #generativeai #clarity

Great conversation between DeepMind's Hassibis and Dwarkesh Patel on AGI, safety, scaling, etc. Of the many areas covered, two things stand out for me: 1. Eval frameworks for large models aren't doing well to capture and measure some key underlying risks. Yes we have leaderboards and benchmarks and so on, but these only tell us how well models do on a set of well worn paths. Hassibis points to deception as a risk that lurks in the shadows, but beyond that there are clear gaps between real-world risks models face / create, on the one hand, and the narrow things measured in current evals, on the other hand. This is silimilar to the kind of gap that exists between the SAT and true intelligence. What's lacking is stronger attention to the gap between risk area and metric/test. 2. Self play for alignment and safety. Humans have a role to play in supervision and feedback, but scaling is a huge challenge, and as models grow in capability and as model tasks grow in size and complexity, big questions arise around the quality and reliability of (sufficiently fast) human feedback. Self play and other forms of model supervision have to be a big part of the path forward. Lots of interesting work happening in this area already (e.g., Deepmind's double efficient debate https://lnkd.in/dHr7yJQt, Meta's multi-round automated red-teaming https://lnkd.in/e3XuT5g7, etc). I can't imagine a near/medium term future where this doesn't need to become a key tool in our alignment and safety belt. #artificialintelligence #aisafety #trustworthyai #llm

It's important to be an optimist when it comes to the future of how technology will help us figure out big data problems. Thanks for the shoutout For ClimateTech! With AI automating pattern analysis/correlation and quantum computing opening the flood gates of access to high performance computing, big data problems will be no more. We will know how to analyse the data and construct more accurate next steps to solve big data problems affecting the climate. FLUIX Inc. is part the Venture For ClimateTech Cohort 3, where we engage with industry experts and mentors through programming, strategy sessions, and access to non-dilutive funding. If your interested in learning more about the program, visit: https://lnkd.in/eP325t8C #fluix #ai #quantum #data #datacenter #cleantech #climate #cooling #energy

📢 As we celebrate crossing 9,000 subscribers, I want to thank all of our community members at the The GenAI Collective for your collaborative engagement, valuable insights, and continued support! 📖 🙏 #GenAI #CommunityEngagement ✏ If you're interested in the convergence of Generative AI and humanoid robotics, check out the 17th edition of The GenAI Collective newsletter! I dive into how innovation in humanoid robotics is poised to revolutionize industrial productivity and daily life. Join the conversation and see where this technology is heading next! 🤖 #HumanoidRobots #AIRevolution

Cheers to the #AirForce and #DARPA for this impressive achievement, pitting AI against humans in an aerial dogfight. Besides being incredibly interesting, this video is also a good example of DevOps in the AI world. Very cool. AI will have a different impact on society, on a sector by sector market basis, depending on the density and fidelity of data available in the sector. More data, more better. There’s a lot of talk about disrupting the world with AI, but it will be disrupted, vertical slices at a bite. Byte. ;)

I am really excited about our ICLR workshop paper on novel applications of energy-based models to optimal experimental design for Bayesian inverse problems! View the paper here: https://lnkd.in/eTTY-mmN 😊 I hope some of you can meet Paula Cordero Encinar at the AI4DiffEq workshop today to learn all about it! #ICLR24 #AI4Science

The #DeepTechAward slowly approaches and we cannot wait to honor the Deep Tech Stars 2024 on the 11th of July! 🚀 Time to spotlight our visionary finalists in a 15 Day countdown. ⏳ Welcome to our Deep Tech Talk with SAFIA Technologies GmbH. The finalist in the category #Sustainable and #SocialImpact offers a rapid testing solution for toxins in groceries. Learn more about their potential for change in their #DeepTechTalk. Senatsverwaltung für Wirtschaft, Energie und Betriebe #DeepTechAward #DeepTechBerlin #BerlinTech #DTA24 #Berlin