Stijn De Baerdemacker
@cortogantese.bsky.social
The reason why the sky is blue is #quantum. The reason why it is sometimes pink is sunsets. #theochem #xai #qubits #quovadis
Most sunsets observed in #Fredericton 🍁
Most sunsets observed in #Fredericton 🍁
As always, lively discussions and a whole-lotta-learning during QuNB seminars! (2/2)
November 15, 2025 at 3:28 AM
As always, lively discussions and a whole-lotta-learning during QuNB seminars! (2/2)
Recommend checking out these references:
arxiv.org/abs/2412.13088
doi.org/10.1039/D4DD...
as well as the impressive work on MACE-Osaka24 models:
github.com/qiqb-osaka/m...
Foundation ML work FTW! (2/2)
arxiv.org/abs/2412.13088
doi.org/10.1039/D4DD...
as well as the impressive work on MACE-Osaka24 models:
github.com/qiqb-osaka/m...
Foundation ML work FTW! (2/2)
Taming Multi-Domain, -Fidelity Data: Towards Foundation Models for Atomistic Scale Simulations
Machine learning interatomic potentials (MLIPs) are changing atomistic simulations in the field of chemistry and materials science. However, constructing a single universal MLIP that can accurately mo...
arxiv.org
November 12, 2025 at 8:32 PM
Recommend checking out these references:
arxiv.org/abs/2412.13088
doi.org/10.1039/D4DD...
as well as the impressive work on MACE-Osaka24 models:
github.com/qiqb-osaka/m...
Foundation ML work FTW! (2/2)
arxiv.org/abs/2412.13088
doi.org/10.1039/D4DD...
as well as the impressive work on MACE-Osaka24 models:
github.com/qiqb-osaka/m...
Foundation ML work FTW! (2/2)
Many among us still hungover from the blue bird’s toxic potion …
November 1, 2025 at 3:04 AM
Many among us still hungover from the blue bird’s toxic potion …
Te weinig, en eerder in een droge co-author list. “Joint work” is mooi expliciet.
October 17, 2025 at 1:53 PM
Te weinig, en eerder in een droge co-author list. “Joint work” is mooi expliciet.
“It was all trivial”
October 12, 2025 at 6:48 PM
“It was all trivial”
Excited to see this one out an learn more👏. Enjoy the choice for a grid-based approach, and wonder about the internal representation and transferability of the model. 🧪
October 11, 2025 at 3:10 AM
Excited to see this one out an learn more👏. Enjoy the choice for a grid-based approach, and wonder about the internal representation and transferability of the model. 🧪
And on a darker note: it is a failure of society that we need the investment of $B private companies to deliver cornerstones of research and progress #AIsouvereignty #alfafold #omol25
October 2, 2025 at 10:39 PM
And on a darker note: it is a failure of society that we need the investment of $B private companies to deliver cornerstones of research and progress #AIsouvereignty #alfafold #omol25
Already reposted it on LinkedIn today, (@justinoberman.bsky.social), so might as well throw it in here as well:
“AI is not a replacement of understanding, it’s an amplifier of understanding”
Case in point: you’ll need to understand DFT if you want ML to have an impact in #compchem/#quantchem
“AI is not a replacement of understanding, it’s an amplifier of understanding”
Case in point: you’ll need to understand DFT if you want ML to have an impact in #compchem/#quantchem
October 2, 2025 at 10:30 PM
Already reposted it on LinkedIn today, (@justinoberman.bsky.social), so might as well throw it in here as well:
“AI is not a replacement of understanding, it’s an amplifier of understanding”
Case in point: you’ll need to understand DFT if you want ML to have an impact in #compchem/#quantchem
“AI is not a replacement of understanding, it’s an amplifier of understanding”
Case in point: you’ll need to understand DFT if you want ML to have an impact in #compchem/#quantchem
Can’t really pinpoint it
October 2, 2025 at 4:13 AM
Can’t really pinpoint it
Why does it look like this fellow is AI- generated?
October 1, 2025 at 6:31 PM
Why does it look like this fellow is AI- generated?
September 25, 2025 at 2:24 AM
Nevertheless, the question got the team engaged in some fun philosophical meanderings, involving undergrad Incé Husain (currently PhD candidate in computational neuroscience @westernu.ca) and QuNB grad Amer El-Samman, among others. Preprint is here: doi.org/10.21203/rs.... (5/5)
Occupancies vs Embeddings: Internal representations of Ab Initio and Graph Probabilistic Models*
Chemical modeling has traditionally been separated into two fairly non-overlapping frameworks: the exact framework of \textit{ab initio} models and the data-driven framework of statistical models. How...
doi.org
September 25, 2025 at 2:24 AM
Nevertheless, the question got the team engaged in some fun philosophical meanderings, involving undergrad Incé Husain (currently PhD candidate in computational neuroscience @westernu.ca) and QuNB grad Amer El-Samman, among others. Preprint is here: doi.org/10.21203/rs.... (5/5)
and straightforward investigation are not conclusive. Many "hyperparameters" need to be further explored before we can phrase hard claims, such as the rigour of the population analysis (Hirshfeld, etc), the choice of rung on Jacob's ladder, or in/co/equivariance of the ML models. (4/5)
September 25, 2025 at 2:24 AM
and straightforward investigation are not conclusive. Many "hyperparameters" need to be further explored before we can phrase hard claims, such as the rigour of the population analysis (Hirshfeld, etc), the choice of rung on Jacob's ladder, or in/co/equivariance of the ML models. (4/5)