Molecular Quantum Solutions

This summer we have initiated and guided several student projects at DTU - Technical University of Denmark and Københavns Universitet - University of Copenhagen with the master thesis and project titles: "Modular Robotic Lab Automation System for Sample Handling and Processing" by Alan Mansour and Nima Jalili "Cloud Resource Cost Optimisation for Machine Learning Model Training using Kubeflow in Kubernetes Clusters" by Mathias Kristensen "Quantum Computing on Periodic Plane-wave Electronic Structure Calculations" by YIZHI LIU Congratulations to all of you and we wish you all the best! The door is always open at MQS for kickstarting new projects to push the needle in quantum chemistry, quantum computing, lab automation and high performance computing for quantum chemistry / AI / LLM / machine learning models. At MQS we focus on computational intensive models within pharma, life science and chemistry; highly optimized on hybrid and energy-efficient CPU-GPU-QPU infrastructure.

To understand quantum computing papers such as this very recent one: "Solving an Industrially Relevant Quantum Chemistry Problem on Quantum Hardware" by Nützel et al.; https://lnkd.in/d9AqwMWJ our "Quantum Computing for Chemistry Workshop" can help you with learning quantum chemistry theory and mapping problems to quantum computers. Sign up for our workshop via [email protected] Theory and code exercises include: - Define and visualize a quantum state on the Bloch sphere - Cartesian to polar form conversion; Polar to cartesian form conversion - Braket notation and linear algebra - Creating quantum circuits - Measuring qubits - Understanding 1-,2, 3-qubit quantum gates; Toffoli gates; Universal Quantum Gates - Pure and mixed states - Adapting to quantum hardware topology - Basic quantum computing methods/algorithms - Quantum and digital annealing - Variational quantum algorithms (VQA) - Parameterized quantum circuits for quantum machine learning (QML) and with tensor networks - Classifying via QML - Quantum convolutional neural network - Quantum graph neural network - General adversial network (GAN) and Quantum GAN More information on our webpage: https://mqs.dk

New MQS blog post about MLOps, Quantum Chemistry and Machine Learning: https://lnkd.in/gEFMDrXh We are harmonizing our TaskTypes based Cebule API with Kubeflow so that different TaskTypes, be it quantum chemistry models, machine learning training&prediction tasks, quantum computing and data analysis algorithms, can be chained together in a pipeline. Our highly optimized Kubernetes and HPC infrastructure also includes now the newest release of Kubeflow where we achieved in a master thesis project in collaboration with DTU up to 50% reduction of GPU energy/usage costs. As always, sharing is caring: Please reach out if you are interested in these kind of topics: contact at mqs dot dk.

Next week Wednesday, the 28th of August, the Danish Quantum Community is organizing the first online webinar on the topic: "Quantum Technology for Chemistry". Mark Nicholas Jones will present all the exciting stuff going on at MQS ranging from quantum chemistry to multi-scale modelling, machine learning, laboratory automation, and the specific use-case applications we have been working on within formulation development, process simulations and materials design. Register here: https://lnkd.in/gnG_HJeV

We have updated our article how to use our graph neural network (GNN) model to easily build prediction pipelines with the Cebule API. In this way you are able to train and fine-tune GNN models with quantum chemistry descriptors and experimental data to obtain highly accurate prediction models: https://lnkd.in/gGdq7EDZ See in the below picture which use-case applications can be tackled with this holistic pipeline setup and send us a message if you want to know more about the specific use-case you are interested in (contact at mqs dot dk).

Quantum chemistry models should be easy to run for e.g. quantum chemists, chemists¸ formulation experts, chemical engineers and physicists. Cebule (polish: onions) is a cloud based solution incorporating quantum chemistry packages, thermodynamic property calculation methods in combination with a high performance infrastructure setup. This allows you to perform multi-layered modelling and simulation tasks. The initial release supports plane wave DFT and Car-Parrinello molecular dynamics calculations via the Quantum Espresso PWscf (pw.x) and CP (cp.x) packages. Further, an extensive library of calculation routines has been implemented based on the COSMO solvation model. COSMO surface charges and sigma profiles can be easily generated with Cebule in a high-throughput manner. We will add step by step our already implemented methods to the REST-API access in the next upcoming summer and autumn releases: - COSMO-SAC: https://lnkd.in/gMbvBckn - COSMO-SAC for Ionic Liquids, Electrolytes and Deep Eutectic Solvents: https://lnkd.in/g2XjY99F - COSMO-SAC for Polymer-Solvent-Mixtures: https://lnkd.in/gBTAp8Km - COSMO-SAC for Cocrystal Formation of Pharmaceutical Compounds: https://lnkd.in/g8rSVnWs The MQS Python SDK holds examples how to easily send tasks to Cebule. More will be explained in a blog article (https://blog.mqs.dk) and dedicated webinars will be given the upcoming weeks and months. The high performance infrastructure runs currently on a combination of highly energy efficient Intel and Arm processors. GPU access and QPU interfacing can be requested after a support meeting so that your use-case can be solved in the most efficient way. Please reach out via our contact and demo forms on our webpage if you have questions or want to start scoping a project with us: https://mqs.dk

The Quanticom H2Q project meeting at the German Aerospace Center (DLR) in Cologne this week did provide very good sparring between d-fine, planqc, DLR and MQS. For more information about the DLR Quantencomputing-Initiative follow this link: https://lnkd.in/g7V3ieAJ Alexandra Krause and Mark Nicholas Jones from MQS participated in the meeting and we look forward to further push our contributions of hybrid classical-quantum computing for atomistic simulations. This initiative is a large scale investment into the European quantum computing landscape: https://lnkd.in/gPaP_V5v

MQS is featured in the amazing Future of Computing blog run by Benjamin Wolba. If you would like to get an update on the current developments at MQS, then this is the article to read. If you are an investor, feel free to reach out. We are currently fundraising a seed round. contact(at)mqs(dot)dk

This should not to be missed by the serious AI/ML/Digitalisation enthusiasts within the pharma and biotech space. A whole day about how AI is transforming the pharmaceutical industry! The event is organized by Stine Rønholt and Mie Kristensen from the Department of Pharmacy, University of Copenhagen and the Danish Pharmaceutical Society. Contributing presenters will be: Rasmus Knappe (Microsoft) Haiping Lu (The University of Sheffield) Jens Christian Nielsen (Gubra) Mark Nicholas Jones (Molecular Quantum Solutions) Troels Ravn Bærentzen (LEO Pharma) Paula Antunes (Novo Nordisk) Mads Nielsen (Københavns Universitet - University of Copenhagen, Pioneer Centre for AI) Sign up via this link: https://lnkd.in/gNM8_BTP

MQS will participate with two poster contributions and a digital demo at the Chemical Compound Space Conference in Heidelberg. The titles of the contributions are: "Solvation Modelling and Bayesian Design of Experiments for Pharmaceutical Formulation Development"; https://lnkd.in/gasXMWjv Mark Nicholas Jones "HOMO-LUMO Gap and Toxicity Predictions with Graph Neural Networks and Quantum Chemistry Data"; https://lnkd.in/gy-kSy-C Mark Nicholas Jones and Milind Upadhyay "The MQS Platform for High-throughput Screening of Molecules with Quantum Chemistry Models and Machine Learning"; https://lnkd.in/gdxfM_JJ Mark Nicholas Jones, Botond Horvath, Patrik Andó