SciML

🔗 Where deep learning meets recommendation systems - Chris Rackauckas explores unexpected connections in open source development in a recent OSA community event. Watch how these insights could transform your next project: https://lnkd.in/dvF_9n9G #MachineLearning #OpenSource

𝗝𝘂𝗹𝗶𝗮 - 𝗣𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿 𝗳𝗶𝘁𝘁𝗶𝗻𝗴 𝘀𝘁𝗲𝗽 𝗯𝘆 𝘀𝘁𝗲𝗽 𝘈𝘯𝘥 𝘢 𝘤𝘩𝘢𝘭𝘭𝘦𝘯𝘨𝘦 𝘧𝘰𝘳 𝘺𝘰𝘶! Trying to understand Julia's SciML ecosystem, I wanted to build a complete chain of simulation that: 1. 𝙼̲𝚘̲𝚍̲𝚎̲𝚕̲𝚜̲ a suspended mass system and 𝚜̲𝚒̲𝚖̲𝚞̲𝚕̲𝚊̲𝚝̲𝚎̲𝚜̲ it 2. 𝙰̲𝚍̲𝚍̲𝚜̲ 𝚗̲𝚘̲𝚒̲𝚜̲𝚎̲ to the simulated data 3. Runs a 𝚙̲𝚊̲𝚛̲𝚊̲𝚖̲𝚎̲𝚝̲𝚎̲𝚛̲ 𝚏̲𝚒̲𝚝̲𝚝̲𝚒̲𝚗̲𝚐̲ 𝚘̲𝚙̲𝚝̲𝚒̲𝚖̲𝚒̲𝚣̲𝚊̲𝚝̲𝚒̲𝚘̲𝚗̲ to find back the correct parameters 4. 𝙲̲𝚘̲𝚖̲𝚙̲𝚊̲𝚛̲𝚎̲ results These images show the result and the code 𝘀𝘁𝗲𝗽 𝗯𝘆 𝘀𝘁𝗲𝗽. 𝗜𝘁'𝘀 𝗮 𝗽𝗲𝗿𝗳𝗲𝗰𝘁 𝗳𝗶𝘁𝘁𝗶𝗻𝗴! (The callback function is not a step listed above. It is just used to generate the images of the solution) As usual, I post the code on GitHub 👉 https://buff.ly/4gpdzTL 💪 Here is the 𝚌̲𝚑̲𝚊̲𝚕̲𝚕̲𝚎̲𝚗̲𝚐̲𝚎̲ 𝚏̲𝚘̲𝚛̲ 𝚢̲𝚘̲𝚞̲: 𝗪𝗵𝗮𝘁 𝗮𝗺 𝗜 𝗻𝗼𝘁 𝘁𝗲𝗹𝗹𝗶𝗻𝗴 𝗵𝗲𝗿𝗲? 𝗔𝗻𝗱 𝗵𝗼𝘄 𝘁𝗼 𝗲𝘅𝗽𝗹𝗮𝗶𝗻 𝗶𝘁? (𝘛𝘰𝘰 𝘤𝘳𝘺𝘱𝘵𝘪𝘤?) > Answer on Wednesday Feb 12th 🗓️, around the same time 🕝 𝘕𝘰𝘵𝘦: 𝘵𝘩𝘦𝘳𝘦 𝘢𝘳𝘦 𝘧𝘦𝘸 𝘴𝘩𝘰𝘳𝘵𝘤𝘰𝘮𝘪𝘯𝘨𝘴 𝘪𝘯 𝘵𝘩𝘪𝘴 𝘤𝘰𝘥𝘦 - 𝘴𝘶𝘤𝘩 𝘢𝘴 𝘢 𝘭𝘢𝘤𝘬 𝘰𝘧 𝘤𝘢𝘳𝘦 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘴𝘰𝘭𝘷𝘦𝘳 𝘴𝘦𝘭𝘦𝘤𝘵𝘪𝘰𝘯𝘴, 𝘳𝘦𝘥𝘶𝘤𝘪𝘯𝘨 𝘵𝘩𝘦 𝘯𝘶𝘮𝘣𝘦𝘳 𝘰𝘧 𝘪𝘵𝘦𝘳𝘢𝘵𝘪𝘰𝘯𝘴 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘰𝘱𝘵𝘪𝘮𝘪𝘻𝘢𝘵𝘪𝘰𝘯 𝘴𝘰𝘭𝘷𝘪𝘯𝘨 𝘢𝘯𝘥 𝘰𝘵𝘩𝘦𝘳 𝘤𝘰𝘥𝘦 𝘰𝘱𝘵𝘪𝘮𝘪𝘻𝘢𝘵𝘪𝘰𝘯 𝘵𝘩𝘢𝘵 𝘢𝘯 𝘦𝘹𝘱𝘦𝘳𝘪𝘦𝘯𝘤𝘦𝘥 𝘚𝘤𝘪𝘔𝘓 𝘶𝘴𝘦𝘳 𝘸𝘰𝘶𝘭𝘥 𝘥𝘰. 𝘛𝘩𝘪𝘴 𝘪𝘴 𝘯𝘰𝘵 𝘵𝘩𝘦 𝘢𝘯𝘴𝘸𝘦𝘳 𝘵𝘰 𝘵𝘩𝘦 𝘲𝘶𝘦𝘴𝘵𝘪𝘰𝘯/𝘤𝘩𝘢𝘭𝘭𝘦𝘯𝘨𝘦. ---- 𝘛𝘩𝘦 𝘣𝘦𝘴𝘵 𝘸𝘢𝘺 𝘵𝘰 𝘴𝘩𝘰𝘸 𝘵𝘩𝘢𝘵 𝘵𝘩𝘪𝘴 𝘵𝘺𝘱𝘦 𝘰𝘧 𝘱𝘰𝘴𝘵 𝘪𝘴 𝘷𝘢𝘭𝘶𝘢𝘣𝘭𝘦 𝘪𝘴 𝘵𝘰: • 𝚕̲𝚒̲𝚔̲𝚎̲ 👍 this post • 𝚏̲𝚘̲𝚕̲𝚕̲𝚘̲𝚠̲ me 🔔 on LinkedIn (https://buff.ly/4gzsb34) and • 𝚜̲𝚝̲𝚊̲𝚛̲ ⭐️ the GitHub repo (https://buff.ly/3Dmr20V) 𝗧𝗵𝗮𝗻𝗸 𝘆𝗼𝘂 𝗳𝗼𝗿 𝘆𝗼𝘂𝗿 𝘀𝘂𝗽𝗽𝗼𝗿𝘁!

Here's an Integrated Virtual Cell Simulation with Julia combining differential equations, neural networks, and stochastic modeling to simulate complex biological processes. This Julia notebook contains: - ODE-based virtual cell – models molecular reactions, ATP synthesis, glycolysis, cellular signaling, and the cell cycle. - AI-driven predictions – a neural network (Flux.jl) learns from ODE simulations to forecast cell behavior. For instance, at t=15, the predicted cell state is: Float32[-0.43540806; 0.60547364; -0.5930182;;] - Stochastic modeling – adds biological randomness using JumpProcesses.jl. This approach paves the way for virtual experiments, optimizing research in biotechnology, pharma, and systems biology while reducing costs. The Julia notebook: https://lnkd.in/dxKg8JND Would love to hear your thoughts! Could this be useful in your field? #JuliaLang #AI #Biotech #MolecularBiology #ComputationalBiology

Julia is on a remarkable #growth trajectory! 🚀 Over the past 5 years: - Discourse views increased by +494% (136M+) - GitHub stars grew by +412% (511K+) - Published citations rose by +391% (8K+) - Registered packages expanded by +322% (11K+) These numbers reflect Julia's impact and rapid adoption in the #tech and #scientific communities. Explore how Julia is shaping the future! https://lnkd.in/gS9WS2T2 #JuliaLang #OpenSource #programming

The 25th Geilo Winter School is well underway here at Dr. Holms Hotel! ❄️ So far, Chris Rackauckas has given us an introduction to inverse modeling with Julia's SciML ecosystem, and Adrian Kirkeby has started showing us how we can do inverse modeling of water waves. This morning, we even had a session by the pool to see how well the linear wave theory holds in practice—most likely the first time doing practical experiments in the history of the Winter School! Yesterday evening, we had a very interesting poster session, with 14 contributions ranging from computational biology to reinforcement learning for decision making. Looking forward to the rest of the week, where we will also have lectures from Sergey Alyaev and Giovanni S. Alberti Øystein Klemetsdal Torkel Andreas Haufmann SINTEF Digital Norges forskningsråd

💡 Just Released: Watch our latest OSA Community session where Chris Rackauckas reveals how JuliaHub turned JuliaSim into a successful product while strengthening its open source roots. Full recording: https://lnkd.in/dvF_9n9G #OpenSourceArchitect