XACS team wishes you Happy Holidays and all success in New Year 2024!

2023 was a productive year as we improved a range of XACS services and program packages. Here we want to share with you the overview of the main news and upgrades. Improvements are often based on your valuable feedback – thank you for your continued support!

New cloud computing interface

We rolled out the major upgrade of our web interface for the XACS cloud computing with impressive list of capabilities:

• improved job submitter where you can write or upload your input files and submit the job.
• online terminal allowing the experienced users to use all the power of the Linux terminal, write scripts and submit their jobs through the command line.
• file manager allowing to easier manage your data, i.e., view, edit, upload and download the files and directly submit the jobs.
• job manager where you can navigate through the computing jobs, manage them and directly jump to the directory with the input and output files.

All of the above features are seamlessly integrated with each other.

More powerful cloud computing hardware

We have also substantially upgraded our hardware and the users can now enjoy more and faster CPUs.

Upgraded XACS software

We are continuously improving all our core software packages, XMVB, XEDA, and MLatom. All of them were substantially upgraded in 2023:

• XMVB 3.1 and 3.2 released with
  • hybrid MPI/OpenMP parallelization
  • implementation of symmetry-constrained valence bond SCF calculations
  • up to 60 times speedup of valence bond orbital optimization.

XEDA (standalone) launched on XACS cloud with over 10 times speedup owing to

• an independent new integral module.
• independent SCF module.
• the density matrix batch processing algorithm.

• MLatom 3 was released with
  

  • Python API.
  • a wider range of quantum mechanical and machine learning methods.
  • more simulation types ranging from geometry optimizations to dynamics and spectra calculations.

  [Figure below from MLatom paper, see preprint]

  

Upgraded documentation and tutorials

In 2023 we provide much more user-friendly online documentation and tutorials for the users of the XACS cloud computing and all core packages, XMVB, XEDA, and MLatom. While most of them are in English, we have also prepared Chinese documentation of MLatom. 

XACS workshops

In addition to the improved documentation and tutorials, we also provide on-site trainings. Special tutorials were prepared for the workshop participants and they are also available online.

In May, we had the second XACS workshop which more than 70 participants.

In December, a smaller MLatom@XACS workshop was held in Torun University, Poland.

Highlights of research done with XACS platform

We are happy to see that the XACS platform is used by independent experts to perform cutting-edge research and hence we started Research highlights section on our website and social media platforms.

One of such examples is the contribution by Rui Liu et al. on the influence of hydrogen bonds on the roaming reaction. If you want to share with us your research, please contact us and we will be happy to highlight it!

Below is the an overview of the main publications from the XACS team documenting new implementations and showcasing the usage of the packages:

• Articles on valence bond theory with XMVB:
  • Using block-localized wavefunction implemented in XMVB for studying planar four-membered diboron actinide compound with doble Möbius aromaticity (J. Am. Chem. Soc.)
  • A novel strategy for the construction of many-electron symmetry-adapted valence bond wave function (J. Chin. Chem. Soc.)
• Articles on energy decomposition analysis with XEDA:
  • Energy decomposition analysis of intermolecular interactions in excited states using XEDA (Phys. Chem. Chem. Phys.)
  • Assessment of three DFT-based Energy Decomposition Analysis Methods for Intermolecular Interactions (J. Chem. Phys.)
• Articles on AI-enhanced computational chemistry done with MLatom:
  • showcasing the power of AIQM1 in out-of-the-box applications for elucidating real-world ground- and excited-state experimental observations of large organic compounds (Chem. Eur. J.)
  • efficient and interpretable ML-two photon absorption method (Adv. Sci.)
  • MLatom 3 (submitted, preprint available), MLQD (Comput. Phys. Commun.), Sparrow (J. Chem. Phys.) packages.
  • MD and IR spectra simulations (Phys. Chem. Chem. Phys.)
  • Four-dimensional-spacetime atomistic artificial intelligence models - a novel concept for more efficient MD and beyond (J. Phys. Chem. Lett.)
  • implementation of a very accurate machine learning interatomic potentials (p)KREG (J. Chem. Theory Comput.)
  • benchmark of AIQM1 on barrier heights (it works better than semiempirical methods, J. Chem. Phys.).

Stronger team

The success of the XACS platform is owed to the big team of enthusiastic and excellent developers, partners, and support staff.

We are very glad that our team have become significantly stronger this year after new members joined us in 2023. Welcome and thank you for your hard work!

As we work in the university, many of the contributors to the XACS platform are our students, who have to graduate and leave our team. We thank them cordially for all their contributions and wish them lots of success in their next career stage!

What is next in 2024?

We are eagerly looking forward to the growth of XACS platform in 2024 and you are going to see next year:

• better tutorials and documentation.
• bilingual support of the website and documentations, which should be in available in both English and Chinese languages.
• improved XACS web GUI
• introduction of XACS Jupyter notebooks
• more and better AI models, extension of general-purpose models to more elements, widening applicability domains to, e.g., chemical reactions.
• more user-friendly input and more informative output of MLatom.
• MLatom tools for active learning.
• more applications of AI to solving real-world problems.