🎉🚀 We are super-excited to announce that our #tLaSDI: #thermodynamics-#informed #latent #space #dynamics #identification is published online in the #CMAME journal! 📚✨ 🔗 Check it out here: https://lnkd.in/g47q3GFv 🌟 tLaSDI introduces a thermodynamics partial differential equation, i.e., #GENERIC #formalism, to model latent space dynamics. This approach creates an intriguing link between full and reduced space dynamics, relating full order dynamics with thermodynamic quantities, such as #energy and #entropy, within the latent space. 🌐🔬 For instance, we observe that the rate of increase in entropy of the latent space dynamics corresponds to the shock-forming phenomenon in Burgers' equation. This #interpretable #relationship between latent space and full order model dynamics will be pivotal in advancing the data-driven approach, serving as an indicator to switch from one phase to another as the dynamic regime changes. 🔄💡 🔍 Besides these insights, many other fascinating aspects are discussed in the paper, such as the effect of #simultaneous #training, different #loss #terms, #greedy #sampling for #generalization, and #error #analysis. 📈🧩 🔥 Pretty cool stuff! The paper is available for #free for the next 50 days, so make sure to take advantage of it! 📖💥 👏 Kudos to Jun Sur Park, Siu Wun Cheung, and Yeonjong Shin for their outstanding effort on this development! 🌟🙌 #neuralnetwork #ML #simulation #science #dynamics #ROM
Day-after-day I am amused at the depth and breath of your work. Please, leave something for other researchers!!! 😄
Professor/catedrático, Universidad de Zaragoza
3moThank you for citing our work!