Andreea Danielescu

Last Thursday I gave a talk about our work at the Gaia Consortium, focusing on planet-scale model-based AI safety -- or more broadly, what we've been calling "everything safety": measurement, management and mitigation of catastrophic risk in an AI-enabled, interconnected planet. Check it out! https://lnkd.in/dP9x7bt3

19

1 Comment

Thinking Deeply about IDs, Names, and Renaming, https://lnkd.in/gnhcX45m

39

If you are involved in developing LLM applications, you may find this resource useful. It highlights key strategies for addressing the challenges in LLM application development. #LLMDevelopment #DataQuality #AIApplications 

3

“I think they [brain-computer interfaces] are a kinda neat idea.” – Geoffrey Hinton. Thought-provoking episode on neurotech and brain-computer interfaces by Bloomberg Originals. Alongside insights from neurotech pioneers like Jacob Robinson, Thomas Oxley, and Conor Russomanno, a highlight was hearing Geoffrey Hinton, Nobel laureate and “Godfather of AI,” share his perspective on implantable chips. Interviewer: “Would you put a chip in your brain if you knew it could enhance your neurological functions?” Geoffrey Hinton: “If I knew it could enhance my neurological function and it wouldn’t have any nasty side effects, and there was a very high probability it wouldn’t kill me… yes, I would!” The key challenge? Building capable, less invasive, and safe chips. That’s exactly why MintNeuro exists. – Dorian #neurotech #braincomputerinterfaces #bci #neuralimplants #brainimplants #chips #brainchips #semiconductors

58

3 Comments

A must-read for ML researchers and engineers. Robotics is a domain that keeps AI/ML research honest. If an AI/ML approach from your paper is applicable to robotics in principle but can't be deployed on a robot w/o "future work" 😉, keep on trying! This is also a reminder to self.

16

5th Annual Embodied AI Workshop -- what better way to spend your 2nd day at #CVPR2024? https://embodied-ai.org/ There is a lot going on in embodied AI these days, with the meaning of the term itself still a bit of a mystery to many. In this workshop, we will be looking at embodied AI from three perspectives -- embodied mobile manipulation, generative AI, and language model planning -- to get folks from different sub-communities to come together and share their progress on this common agenda. Stoked to be there! Come and say hi!

18

The AI in Health Conference is coming up on September 9-12. Themes for this year include #FoundationModels in health, applications of #LLMs in health, AI in #Neuroscience and #Neurotechnology, AI for the healthy industry, #DigitalTwins in healthcare, ad patient engagement and #Equity in #HealthAI.   Submit your abstract to disseminate the results of research! #AIHealthConference #AI #ArtificialIntelligence #DataScience #Health #KenKennedyInstitute

50

1 Comment

Completing the DARPA IMPAQT initiative this year stands out as one of the most rewarding milestones of my career. Leveraging PolarisQB's expertise in quantum annealing, we ventured into uncharted territory by applying those principles to #IBM's gate-based quantum computers. Despite encountering challenges at every turn, we successfully developed a hybrid classical-quantum solution. This approach enables us to efficiently explore extensive fragment libraries and uncover novel molecules that meet all of our stringent constraints. As the accompanying graphs demonstrate, this groundbreaking solution not only identifies more molecules but also delivers better characteristics at a pace that classical methods simply cannot match. These promising results underscore the transformative potential of quantum computing in accelerating discovery and innovation!

10

2 Comments

Ghost engineers I suspect many of you are fascinated by this next definition of disengagement in engineering - the ghost engineer 👻 There’s some usefulness in this, as there is in any discussion, but much of this speaks to management culture too. I’ve routinely seen good engineering leaders really *know* how their team is performing without any hard data. You will find those people can easily put their teams into Gaussian distributions on where they are with performance. Those who can’t really would struggle without any objective measure. That said, it’d be disingenuous of me to say it’s not important to measure developer productivity. I’ve routinely in my career used things like GitHub, Jira, Rituals, etc to help me define how much impact and productivity an engineer has in comparison to their peers. But these are only part of the picture. A person with no GitHub activity this month may have accelerated 4 features via pairing, they may have been heavily involved in architecture design that unlocks future value, etc etc etc. It is, of course, important to be able to understand individual productivity - even though success is a team sport, understanding where each of the individuals within that team are is crucial as a leader. Let’s see how these ghost engineer conversations play out. https://lnkd.in/e3yKtiBf

15

5 Comments

On #FederatedLearning for #AI models: “FL models often struggle with generalization across institutions due to non-uniform data distributions (non-IID), where variations in data size, labels, features, and quality across different sources complicate model generalization, and a lack of standardized data harmonization practices.… To overcome these barriers, the paper recommends improving model transparency, using adaptive frameworks, reducing communication burdens, and addressing statistical and system heterogeneity.”

3

1 Comment

One of the major challenges in developing anything related to hardware is the threat that ultimately one of the really big players is going to come along with something that fundamentally disrupts your entire business case. Looking at Apples recent patent applications for biosensors in earbuds, this might be such a moment coming up for a lot of MedTech startups working with sensor-hardware to solve big health challenges. It is definitely something to be aware of and monitor closely. Nobody knows how long Apple are going to be before the biosensors start shipping embedded in their popular AirPods, and no one knows yet whether they will be truly medical grade in a way that HCPs will find them 'good enough'. But it would probably be a good idea to prepare for it to happen. Having said all this the jury is still out on what patients feel about this? Historically, patients have not been too fond of getting to closely attached to big pharma. Are they going to feel the same about entrusting big tech with their healthdata, when it really matters? Maybe. Interesting times. #healthtech #bigtech #competition #biosensors https://lnkd.in/dnEcVPGU

2

Exciting Opportunity for AI Researchers: Agora Compute Grants We are thrilled to announce the launch of Agora Compute Grants, an initiative designed to empower AI researchers with the computational resources they need to push the boundaries of multi-modal research. https://buff.ly/4clNtQF What is Agora Compute Grants? Agora Compute Grants is an open-source AI research lab initiative aimed at providing AI researchers with access to powerful GPU clusters. These resources are essential for conducting advanced multi-modal research, which integrates diverse data types such as text, images, and audio to create more sophisticated and capable AI models. Why Focus on Multi-Modal Research? Multi-modal research represents the future of AI. By combining various data types, researchers can develop more robust and versatile models, leading to innovations that can revolutionize industries and improve lives. At Agora, we are committed to supporting this cutting-edge research and helping you achieve breakthroughs. How to Get Started: -> Join our Discord community: https://buff.ly/4clNtQF -> Sign up and introduce yourself to the community. -> Share your project in the Creator Showcase to be considered for the compute grants. We invite all passionate AI researchers to take advantage of this incredible opportunity. By joining Agora Compute Grants, you'll gain access to the computational power needed to drive your research forward and make significant contributions to the field of AI. Let's shape the future of AI together. Join our Discord today and start your journey with Agora Compute Grants. For more details and to sign up, visit our Discord community: https://buff.ly/4clNtQF We look forward to seeing the innovative projects you bring to life!

8

1 Comment

Very interesting, insightful study. Flare Capital Partners AI in healthcare is hype. Most AI stories for healthcare are tech advancements. Not real clinical advancements or fitting into their workflows, bearing few. MedTech Innovators face these paradox. 1.Adoption requires Evidence. But evidence comes only from adoption. Clinical data can not be synthetic data for good AI. 2.Regulatory, validation requires money. But investors require regulatory, validation. 3.MedTech need faster adoption, but healthcare is slower among every profession, a cut-throat industry. They won't use it, if it is more than 2-3 mouse clicks. https://lnkd.in/gUh8sDHe

1

Loving the vision presented in this article from Obvious Ventures of how generative science (AI trained on the knowledge of the physical sciences) will enable breakthroughs in everything from physics to chemistry, biology, and beyond. Great read. https://lnkd.in/eWfhuRdH

30

We don't have to be heading towards a collision course between meeting our climate goals and managing AI's intense energy consumption. As federal policies shift, New York has a unique opportunity to set the standard for sustainable AI, creating a blueprint for other states to follow. Governor Kathy Hochul has set an ambitious goal for New York: to become a national and global leader in AI research and development. Perhaps New York’s true competitive advantage lies in attracting funding and investment precisely because it can lead in sustainable AI. Imagine a future where sustainability isn’t just a commitment—it’s New York’s key selling point. Read more below: https://lnkd.in/g_fih3WR

7

Mike Morrison, PhD has just launched his #scienceux journal using Curvenote's publishing infrastructure -- if you are curious about how this works or want to be able to publish in the same way, you can! Whether that's: 📗 open access copies of your lab group's research 📗 a preprint service for your project or community 📗or your own journal -- you can use this too. It's modern, github integrated (if you want), gives you submission and editorial controls, DOIs, interactive graphs, attached compute... If you'd like a demo, message me or comment. #publishing #science

14