sqIRL lab

We are sqIRL (squirrel), the #Interpretable #Representation Learning Lab based IDLab - @UAntwerpen & #imec
#ML #AI #Explainability #RepresentationLearning #XAI

[sqIRL lab @sqIRLlab]

Another sqIRL graduated!
Congratulation Saja (@sajatawalbeh91) on the successful defense of your PhD.
Lots of success

#AI #ML #CapsNets #XAI #interpretability #explainability #sqIRL #IDLab #UAntwerp

[sqIRL lab @sqIRLlab]

TriLite: Efficient Weakly Supervised Object Localization with Universal Visual Features and Tri-Region Disentanglement
Arian Sabaghi, Jose Oramas
CVPR, 2026.

Code + additional resources:
https://t.co/xFxLM4fD6O

Paper URL:
https://t.co/ajpxSSVDPC

[sqIRL lab @sqIRLlab]

Accurate object localization at reduced annotation/computational cost?
#TriLite achieves this by pairing a pre-trained backbone, with our #TriHead component which separates ambiguous, foreground and background features.

More at #CVPR2026

#WSOL #CV
#sqIRL #IDLab #UAntwerp

[sqIRL lab @sqIRLlab]

The main highlights of #TriLite include:
• State-of-the-art performance
• < 1M trainable parameters
• single-stage training

We will be glad to get in touch at #CVPR2026 to discuss further.

We thank #DEFRA for supporting this research

#CV #WSOL #AI
#IDLab #UAntwerp #imec

[sqIRL lab @sqIRLlab]

The review process of scientific articles is not perfect, but a way to improve is by contributing to it in a diligent manner.

Proud to have an #ICML'26 gold reviewer among us.
Congrats Benjamin!

@icmlconf
#AI #ML #Science #PeerReview #Award
#UAntwerp #IDLab

[sqIRL lab @sqIRLlab]

Need more evidence on the potential of #AI #interpretability to support scientific #discovery?

ICYMI: during his fellowship at @GoodfireAI Thomas (@thomasdooms) contributed to the understanding of #genetic variants using interpretability methods

Proud to have had you onboard!

[Goodfire @GoodfireAI]

We achieved state-of-the-art performance in predicting which of 4.2 million genetic variants cause diseases by interpreting a genomics model, in a new preprint with @MayoClinic.

We're now releasing an open source database for all variants in the NIH's clinvar database. 🧵(1/8)

[sqIRL lab @sqIRLlab]

Want to know more about Benjamin's work?
https://t.co/vWVa2mTQam

[sqIRL lab @sqIRLlab]

We are celebrating at the sqIRL lab
Some days ago Benjamin successfully defended hi PhD on parameter-efficient models

Thanks for the cool discussions and all your contributions to #sqIRL/#IDLab.

Congratulations and lots of success dr. Vandersmissen!

#LTH #TNA #AI #ML #UAntwerp

[sqIRL lab @sqIRLlab]

Arian presenting our work on measuring the #uncertainty of visual #explanations from autonomous navigation systems at the #ACRAI research day.

A paper with more details is in the works. A link will follow soon.

#InterpretableML #XAI #AI #ML #UAI #sqIRL #IDLab #UAntwerp #imec

[sqIRL lab @sqIRLlab]

Interested in training hyperspectral image analysis models with reduced annotated data?
Salma (@Salma_h) explores this question in her recent paper.
https://t.co/Fc9bDhZllS

#hyperspectral #HSI #AI #ML #ContrastiveLearning
#sqIRL #IDLab #UAntwerp #imec

[sqIRL lab @sqIRLlab]

Time to celebrate!
Last week @Salma_h successfully defended her PhD on #Representation #Learning for #Hyperspectral Image Analysis.

Thanks for the cool research, your support and all your contributions to #sqIRL/#IDLab.

Congratulations Dr. Haidar!

#HSI #AI #ML #UAntwerp

[sqIRL lab @sqIRLlab]

Kudos to Thomas and the involved collaborators for the solid contributions to the field.

Curious about our work, have a look at our website: https://t.co/fvFX17C11O

#Interpretability #mechinterp #compinterp #xai #AI #ML
#sqIRL #UAntwerp #IDLab

[sqIRL lab @sqIRLlab]

Have a look at the work our lab will be presenting at #NeurIPS '25.
On the main track, SimpleStories, a dataset full of simple yet diverse stories which has the potential of becoming the MNIST for language.
https://t.co/1gRoHDDsPc

#Interpretability #mechinterp #xai #AI #ml

[sqIRL lab @sqIRLlab]

At the MI workshop (spotlight), we show how Bilinear Autoencoders ease the analysis of neural representations through their decomposition into polynomial latents.
Paper and the cool demos at https://t.co/Y1B5hkto0o

#Interpretability #mechinterp #compinterp #xai #AI #ML

[sqIRL lab @sqIRLlab]

We just launched a #linkedin page. Please help us spread the word and share it with people that might be interested.
https://t.co/XOd9WvAwuk

#RepresentationLearning #interpretability #explainability #XAI #mechinterp #AI #ML #sqIRL #ComputerVision #HSI #IDLab #UAntwerp

[sqIRL lab @sqIRLlab]

Saja's survey on the #intepretability/#explainability of Capsule Networks was accepted at #Neurocomputing. Give it a look while it is free to access.
https://t.co/wmzvuu8eEQ

#CapsNets #AI #ML #XAI #sqIRL #UAntwerp #IDLab

[sqIRL lab @sqIRLlab]

Thanks for the #Flanders AI Research Program (FAIR) for supporting this collaboration and to the involved persons for the fruitful collaboration.

#Neuromorphic #hyperdimensional #interpretability #Explainability #VSA #xai #AI #ML #FAIR #UAntwerp #imec

[sqIRL lab @sqIRLlab]

Our work on Interpretable #Hyperdimensional Computing (HDC) classifiers for #tabular data is now available at #Neurocomputing.

https://t.co/oeiEcShD4I

#Neuromorphic #interpretability #Explainability #VSA #xai #ML #FAIR #UAntwerp #imec

[sqIRL lab @sqIRLlab]

#HDC models aim to be an energy-efficient alternative to current #AI systems and thanks to the efforts of our collaborators, their decision-making process is now more interpretable.

#Neuromorphic #interpretability #Explainability #VSA #xai #AI #ML #FAIR #UAntwerp #imec

[sqIRL lab @sqIRLlab]

Thanks to our collaborators from the #VUB, Ward Gauderis and Geraint Wiggins; as well as #sqIRL members Thomas Dooms (@thomasdooms) and José Oramas (@jaom7) for the nice collaboration. #UAntwerp #FlandersAI #FAIR #AIFlanders

[sqIRL lab @sqIRLlab]

This week our lab was present at the Flanders AI Research day. There we contributed with a deep dive session on #Compositional #Interpretability.
More details at: https://t.co/CrRJL61hnB
#CompInterp #interpretableML #XAI #explainability #aisafety #ML #AI