Andrea Rizzi

Postdoctoral researcher @fz_juelich 🇩🇪 and @IITalk 🇮🇹 | Machine learning and molecular simulations for drug discovery and (bio)chemistry

[Andrea Rizzi @andrrizzi]

Just arxived our latest work! We introduce multimap targeted free energy estimation and use it to converge free energy corrections from a FF to a DFTB potential with a few thousand single-point QM calculations. No sampling from QM needed. 1/n https://t.co/k8TjPPeRaz

[Parrinello Group @GroupParrinello]

📢JOB OPENING📢

We are seeking PostDocs with a strong background in biological simulations willing to incorporate enhanced sampling and machine learning in their research

Apply here to join us at @IITalk in the beautiful Genova ⛵️🍕
🔗: https://t.co/tUNL8TZ0Mk

#compchem #RT

[Parrinello Group @GroupParrinello]

🚨NEW PREPRINT🚨
@PeilinKang and @TrizioEnrico present a self-consistent method to:
-Compute the committor based on a variational principle🎢
-Extensively sample the transition state ensemble🎯
-Extract physical insights from the results🔍

📜 https://t.co/rGvsn0sRH8

@IITalk 🧵

[Andrea Rizzi @andrrizzi]

@GroupParrinello Tweetorial for the multimap targeted free energy perturbation work here

n/n
https://t.co/I3pJvDiY1P

[Andrea Rizzi @andrrizzi]

Just arxived our latest work! We introduce multimap targeted free energy estimation and use it to converge free energy corrections from a FF to a DFTB potential with a few thousand single-point QM calculations. No sampling from QM needed. 1/n https://t.co/k8TjPPeRaz

[Andrea Rizzi @andrrizzi]

This was a good week! Sent in my first ERC grant proposal, my multimap targeted free energy perturbation paper was published in PNAS (https://t.co/AZgLwIg0r3), and arXived a new preprint combining ML collective variables and enhanced path sampling!

1/n
https://t.co/WRa4BjWNks

[Andrea Rizzi @andrrizzi]

Fun work with Lukas Müllender, Davide Mandelli, Paolo Carloni, and @GroupParrinello ! Powered by mlcolvar

3/n
https://t.co/mDBLQ9mIBk

[Parrinello Group @GroupParrinello]

Wanna enhance sampling with #machinelearning-based collective variables?

Check out the library and the new paper on @JChemPhys by @LuigiBonati @TrizioEnrico @andrrizzi

📜 Paper: https://t.co/c7As75iZrP
📖 Repo: https://t.co/ZKWqaoHb9M
📁 Docs: https://t.co/ciKzpLVUN1

A short🧵

[Parrinello Group @GroupParrinello]

Wanna enhance sampling with #machinelearning-based collective variables?

Check out the library and the new paper on @JChemPhys by @LuigiBonati @TrizioEnrico @andrrizzi

📜 Paper: https://t.co/c7As75iZrP
📖 Repo: https://t.co/ZKWqaoHb9M
📁 Docs: https://t.co/ciKzpLVUN1

A short🧵

[Andrea Rizzi @andrrizzi]

@JCIM_JCTC @FraGalvani_96 @ScalviniLaura @alessio_lodola I wish! Great work! But wrong Andrea Rizzi (I don't think my namesake has a Twitter account). We do drive Google Scholar crazy.

[Simon Olsson @smnlssn]

Together with @AntonFKockum and @rahmlab we have an open position for a postdoc for AI-driven optimization of quantum computers for chemistry -- apply by July 1st here https://t.co/X2slG1ViqX

[Dhiman Ray @DhimanRay16]

Check out our new pre-print from @GroupParrinello @IITalk with @TrizioEnrico. We show how collective variables for enhanced sampling can be improved by including transition path data. Plus kinetics of the rare event is recovered during training: https://t.co/7VIkeAjm40 #compchem

[Andrea Rizzi @andrrizzi]

@MohsenSadeghi Thanks, Mohsen!

[Andrea Rizzi @andrrizzi]

Just arxived our latest work! We introduce multimap targeted free energy estimation and use it to converge free energy corrections from a FF to a DFTB potential with a few thousand single-point QM calculations. No sampling from QM needed. 1/n https://t.co/k8TjPPeRaz

[Andrea Rizzi @andrrizzi]

@__kraemer__ Thanks a lot, Andreas!

[Andrea Rizzi @andrrizzi]

And many many thanks to @fiona_chembot, @lcwarrensford, @hlwoodcock, Stefan Boresch, and Andreas Schöller for making the HiPen dataset available (https://t.co/eDK1b4rcz1, https://t.co/KoOwKrwokT)! Having input files and reference calculations was immensely useful. n/n

[Andrea Rizzi @andrrizzi]

The method is orders of magnitude faster than standard FEP and seems to compare favorably also to nonequilibrium calculations (about 8x faster) on the same set of molecules. 3/n

[Andrea Rizzi @andrrizzi]

@SoniaDhull8 Just click on "apply now" in the website I linked :)

[Andrea Rizzi @andrrizzi]

We have an open Marie-Curie PhD position to work on machine learning and molecular simulations! This is a joint PhD between KTH (🇸🇪) and Forschungszentrum Juelich/RWTH Aachen (🇩🇪). Deadline is Feb 28th. Retweets appreciated. https://t.co/IxC52ltxwP

[Theoretical & Computational Biophysics Department @CompBioPhys]

👉Open #phdpositions in the group of Aljaz Godec @mpi_nat in #mathbiophys in Göttingen 🇩🇪! Scroll through the website🔗https://t.co/8xVhVw0z3D for more details on job and research! #job #computation #theory RTs apprechiated! 🤩

[Andrea Rizzi @andrrizzi]

@zoecournia @GiEmmePi1 Thank you, Zoe!

[Andrea Rizzi @andrrizzi]

(1/7) We (Paolo Carloni, MP @GroupParrinello, and myself) have been experimenting with targeted free energy perturbation and normalizing flows to compute QM-accurate free energies from cheaper reference potentials. Here is what we learned.
https://t.co/3bKxjUMghv