#representationlearning — Public Fediverse posts

🧠 New preprint by Fabian A. Mikulasch & @fzenke: Understanding Self-Supervised #Learning via #LatentDistribution Matching proposes a unifying theoretical framework for #SelfSupervisedLearning.

The paper reframes #SSL as latent distribution matching, connecting contrastive, non-contrastive, predictive, and stop-gradient methods through a common probabilistic principle linking alignment, uniformity, and latent entropy.

📝 https://arxiv.org/abs/2605.03517

#MachineLearning #RepresentationLearning #AI

🧠 New preprint by Fabian A. Mikulasch & @fzenke: Understanding Self-Supervised #Learning via #LatentDistribution Matching proposes a unifying theoretical framework for #SelfSupervisedLearning.

The paper reframes #SSL as latent distribution matching, connecting contrastive, non-contrastive, predictive, and stop-gradient methods through a common probabilistic principle linking alignment, uniformity, and latent entropy.

📝 https://arxiv.org/abs/2605.03517

#MachineLearning #RepresentationLearning #AI

🧠 New preprint by Fabian A. Mikulasch & @fzenke: Understanding Self-Supervised #Learning via #LatentDistribution Matching proposes a unifying theoretical framework for #SelfSupervisedLearning.

The paper reframes #SSL as latent distribution matching, connecting contrastive, non-contrastive, predictive, and stop-gradient methods through a common probabilistic principle linking alignment, uniformity, and latent entropy.

📝 https://arxiv.org/abs/2605.03517

#MachineLearning #RepresentationLearning #AI

🧠 New preprint by Fabian A. Mikulasch & @fzenke: Understanding Self-Supervised #Learning via #LatentDistribution Matching proposes a unifying theoretical framework for #SelfSupervisedLearning.

The paper reframes #SSL as latent distribution matching, connecting contrastive, non-contrastive, predictive, and stop-gradient methods through a common probabilistic principle linking alignment, uniformity, and latent entropy.

📝 https://arxiv.org/abs/2605.03517

#MachineLearning #RepresentationLearning #AI

🧠 New preprint by Fabian A. Mikulasch & @fzenke: Understanding Self-Supervised #Learning via #LatentDistribution Matching proposes a unifying theoretical framework for #SelfSupervisedLearning.

The paper reframes #SSL as latent distribution matching, connecting contrastive, non-contrastive, predictive, and stop-gradient methods through a common probabilistic principle linking alignment, uniformity, and latent entropy.

📝 https://arxiv.org/abs/2605.03517

#MachineLearning #RepresentationLearning #AI

Big congratulations to all authors! 🚀

#ICLR2026 #MachineLearning #AIResearch #RepresentationLearning #InformationRetrieval #DenseRetrieval #SelfSupervisedLearning #LanguageModels #NLP #UKPLab #ICLR

Big congratulations to all authors! 🚀

#ICLR2026 #MachineLearning #AIResearch #RepresentationLearning #InformationRetrieval #DenseRetrieval #SelfSupervisedLearning #LanguageModels #NLP #UKPLab #ICLR

Big congratulations to all authors! 🚀

#ICLR2026 #MachineLearning #AIResearch #RepresentationLearning #InformationRetrieval #DenseRetrieval #SelfSupervisedLearning #LanguageModels #NLP #UKPLab #ICLR

Big congratulations to all authors! 🚀

#ICLR2026 #MachineLearning #AIResearch #RepresentationLearning #InformationRetrieval #DenseRetrieval #SelfSupervisedLearning #LanguageModels #NLP #UKPLab #ICLR

Representation learning often emphasizes metric preservation. We instead build Symplectic structural invariance directly into the representation.

https://arxiv.org/abs/2512.19409

We embed Hamiltonian/symplectic geometry by making the RNN state dynamics a symplectomorphism, which preserves Legendre duality (information geometry) through time. This yields structure-preserving representations enforced by the latent dynamics, rather than imposed indirectly via the output.

#ReservoirComputing #RepresentationLearning #InformationGeometry #SymplecticGeometry #HamiltonianDynamics #GeometricDeepLearning #DynamicalSystems #PhysicsInformedML

We still have free slots in our KIT summer semester 2025 seminar on "Large Language Model-Enhanced Representation Learning for Knowledge Graphs" supervised by @GenAsefa, Mary Ann Tan and @lysander07

https://portal.wiwi.kit.edu/ys/8600

#teaching #knowledgegraphs #llms #generativeai #representationlearning #seminar @fiz_karlsruhe @KIT_Karlsruhe #AI

'Desiderata for Representation Learning: A Causal Perspective', by Yixin Wang, Michael I. Jordan.

http://jmlr.org/papers/v25/21-107.html

#representationlearning #representations #representation

In today's #ise2023 lecture, we discussed neural networks, from the very simply McCulloch-Pitts Neuron up to Convolutional Neural Networks and Generative Adversarial Networks. A lot of content for only 90 minutes of lecture ;-)
Slides: https://drive.google.com/file/d/1KsAGBViEzu7UDkz-F2QJJJ23ce-AQ9pl/view?usp=drive_link
#machinelearning #deeplearning #representationlearning #knowledgegraphs #graphembeddings #embeddings #lecture @fizise @enorouzi #aiart #stablediffusionart

Hi everybody #introduction, this is FIZ ISE (Information Service Engineering) research group at #FIZKarlsruhe and #AIFB/KIT, switching from the birdcage to this lovely new environment. We will be tooting about our latest research in #semanticweb #knowledgegraph #deeplearning #knowledgeextraction #researchdatamanagement #representationlearning #semanticsearch #exploratorysearch and many more.

Application areas: #culturalheritage #digitalhumanities #materialsscience
#datascience #mathematics #ai

Looking to follow (and be followed!) #mlbio #bioml #biotech #bayes #bayesian #generativemodels #startup #startups #medicine #biology #representationlearning #techbio #immunology #neurips #icml #iclr #cambridge #nyc

Time for an #introduction I guess :)

I'm an assistant professor at VU Amsterdam in machine learning, with a focus on reinforcement learning and representation learning.

I don't post very often on social media but when I do it's usually about research.

I'm looking forward to seeing how Mastodon with its decentralized approach can take off!

#machinelearning, #reinforcementlearning, #representationlearning, #deeplearning

'Desiderata for Representation Learning: A Causal Perspective', by Yixin Wang, Michael I. Jordan.

http://jmlr.org/papers/v25/21-107.html

#representationlearning #representations #representation

'Desiderata for Representation Learning: A Causal Perspective', by Yixin Wang, Michael I. Jordan.

http://jmlr.org/papers/v25/21-107.html

#representationlearning #representations #representation

'Desiderata for Representation Learning: A Causal Perspective', by Yixin Wang, Michael I. Jordan.

http://jmlr.org/papers/v25/21-107.html

#representationlearning #representations #representation

'Desiderata for Representation Learning: A Causal Perspective', by Yixin Wang, Michael I. Jordan.

http://jmlr.org/papers/v25/21-107.html

#representationlearning #representations #representation

Hi everybody #introduction, this is FIZ ISE (Information Service Engineering) research group at #FIZKarlsruhe and #AIFB/KIT, switching from the birdcage to this lovely new environment. We will be tooting about our latest research in #semanticweb #knowledgegraph #deeplearning #knowledgeextraction #researchdatamanagement #representationlearning #semanticsearch #exploratorysearch and many more.

Application areas: #culturalheritage #digitalhumanities #materialsscience
#datascience #mathematics #ai

Hi everybody #introduction, this is FIZ ISE (Information Service Engineering) research group at #FIZKarlsruhe and #AIFB/KIT, switching from the birdcage to this lovely new environment. We will be tooting about our latest research in #semanticweb #knowledgegraph #deeplearning #knowledgeextraction #researchdatamanagement #representationlearning #semanticsearch #exploratorysearch and many more.

Application areas: #culturalheritage #digitalhumanities #materialsscience
#datascience #mathematics #ai

Hi everybody #introduction, this is FIZ ISE (Information Service Engineering) research group at #FIZKarlsruhe and #AIFB/KIT, switching from the birdcage to this lovely new environment. We will be tooting about our latest research in #semanticweb #knowledgegraph #deeplearning #knowledgeextraction #researchdatamanagement #representationlearning #semanticsearch #exploratorysearch and many more.

Application areas: #culturalheritage #digitalhumanities #materialsscience
#datascience #mathematics #ai

Hi everybody #introduction, this is FIZ ISE (Information Service Engineering) research group at #FIZKarlsruhe and #AIFB/KIT, switching from the birdcage to this lovely new environment. We will be tooting about our latest research in #semanticweb #knowledgegraph #deeplearning #knowledgeextraction #researchdatamanagement #representationlearning #semanticsearch #exploratorysearch and many more.

Application areas: #culturalheritage #digitalhumanities #materialsscience
#datascience #mathematics #ai

Looking to follow (and be followed!) #mlbio #bioml #biotech #bayes #bayesian #generativemodels #startup #startups #medicine #biology #representationlearning #techbio #immunology #neurips #icml #iclr #cambridge #nyc

Looking to follow (and be followed!) #mlbio #bioml #biotech #bayes #bayesian #generativemodels #startup #startups #medicine #biology #representationlearning #techbio #immunology #neurips #icml #iclr #cambridge #nyc

Looking to follow (and be followed!) #mlbio #bioml #biotech #bayes #bayesian #generativemodels #startup #startups #medicine #biology #representationlearning #techbio #immunology #neurips #icml #iclr #cambridge #nyc

Looking to follow (and be followed!) #mlbio #bioml #biotech #bayes #bayesian #generativemodels #startup #startups #medicine #biology #representationlearning #techbio #immunology #neurips #icml #iclr #cambridge #nyc