Nicolas Audebert
@nshaud.bsky.social
Neuron trainer and Earth Observation mage at IGN France. Kicking stuff until it works.
🌐 https://nicolas.audebert.at
🐘 http://mastodon.social/@nsh
(he/him)
🌐 https://nicolas.audebert.at
🐘 http://mastodon.social/@nsh
(he/him)
Félicitations au nouveau docteur Maxime Mérizette qui a soutenu avec brio ce matin sa thèse sur la segmentation sémantique de petits objets dans des nuages de points 3D.
Un vaste travail mêlant données synthétiques, pseudo-annotations de modèles profonds. Bravo ! 👏🏻
Un vaste travail mêlant données synthétiques, pseudo-annotations de modèles profonds. Bravo ! 👏🏻
October 8, 2025 at 3:46 PM
Félicitations au nouveau docteur Maxime Mérizette qui a soutenu avec brio ce matin sa thèse sur la segmentation sémantique de petits objets dans des nuages de points 3D.
Un vaste travail mêlant données synthétiques, pseudo-annotations de modèles profonds. Bravo ! 👏🏻
Un vaste travail mêlant données synthétiques, pseudo-annotations de modèles profonds. Bravo ! 👏🏻
MusicXML is a weirdly lax format that allows for various mistakes and inconsistencies.
Léo designed a state machine to detect errors such as:
- incomplete/overcomplete bars
- duplicate notes in chords,
- overlapping voices,
- etc.
AFAIK, this is the first comprehensive checker for digital scores!
Léo designed a state machine to detect errors such as:
- incomplete/overcomplete bars
- duplicate notes in chords,
- overlapping voices,
- etc.
AFAIK, this is the first comprehensive checker for digital scores!
October 6, 2025 at 3:01 PM
MusicXML is a weirdly lax format that allows for various mistakes and inconsistencies.
Léo designed a state machine to detect errors such as:
- incomplete/overcomplete bars
- duplicate notes in chords,
- overlapping voices,
- etc.
AFAIK, this is the first comprehensive checker for digital scores!
Léo designed a state machine to detect errors such as:
- incomplete/overcomplete bars
- duplicate notes in chords,
- overlapping voices,
- etc.
AFAIK, this is the first comprehensive checker for digital scores!
Breaking news from #AI4EO symposium: cows confirmed as extraterrestrial beings. 👽 #CowsFromSpace
More information to come.
More information to come.
September 12, 2025 at 8:11 AM
Breaking news from #AI4EO symposium: cows confirmed as extraterrestrial beings. 👽 #CowsFromSpace
More information to come.
More information to come.
Although I shouldn't be that surprised, I worked with two co-authors of the FlowNet paper that introduced the Flying Chairs dataset.
lmb.informatik.uni-freiburg.de/resources/da...
lmb.informatik.uni-freiburg.de/resources/da...
June 20, 2025 at 4:19 PM
Although I shouldn't be that surprised, I worked with two co-authors of the FlowNet paper that introduced the Flying Chairs dataset.
lmb.informatik.uni-freiburg.de/resources/da...
lmb.informatik.uni-freiburg.de/resources/da...
Browsing old mails, I found one from 2015 during my internship at TUM working on facial expression recognition.
I had completely forgotten that I was considering using synthetic data at that time! We dabbled with MakeHuman for a while but gave up.
Funny I'm still into that 10 years later.
I had completely forgotten that I was considering using synthetic data at that time! We dabbled with MakeHuman for a while but gave up.
Funny I'm still into that 10 years later.
June 20, 2025 at 4:13 PM
Browsing old mails, I found one from 2015 during my internship at TUM working on facial expression recognition.
I had completely forgotten that I was considering using synthetic data at that time! We dabbled with MakeHuman for a while but gave up.
Funny I'm still into that 10 years later.
I had completely forgotten that I was considering using synthetic data at that time! We dabbled with MakeHuman for a while but gave up.
Funny I'm still into that 10 years later.
📣 Author notifications for TerraBytes are OUT!
We've received more than 50 submissions from all around the world. Thanks to our amazing reviewers who have worked on very short notice to help us build an amazing schedule. 🫂
See you all in Vancouver!
We've received more than 50 submissions from all around the world. Thanks to our amazing reviewers who have worked on very short notice to help us build an amazing schedule. 🫂
See you all in Vancouver!
June 10, 2025 at 12:56 PM
📣 Author notifications for TerraBytes are OUT!
We've received more than 50 submissions from all around the world. Thanks to our amazing reviewers who have worked on very short notice to help us build an amazing schedule. 🫂
See you all in Vancouver!
We've received more than 50 submissions from all around the world. Thanks to our amazing reviewers who have worked on very short notice to help us build an amazing schedule. 🫂
See you all in Vancouver!
🚨 Workshop alert!
🌍💿 TerraBytes: Towards global datasets and models for Earth Observation
Co-located with ICML 2025 (Vancouver, Canada)
🔗 terrabytes-workshop.github.io
📝 We accept short 4-pages and full 8-pages paper on remote sensing, Earth Observation and machine learning.
📆 Deadline: May 16th
🌍💿 TerraBytes: Towards global datasets and models for Earth Observation
Co-located with ICML 2025 (Vancouver, Canada)
🔗 terrabytes-workshop.github.io
📝 We accept short 4-pages and full 8-pages paper on remote sensing, Earth Observation and machine learning.
📆 Deadline: May 16th
March 27, 2025 at 1:50 PM
🚨 Workshop alert!
🌍💿 TerraBytes: Towards global datasets and models for Earth Observation
Co-located with ICML 2025 (Vancouver, Canada)
🔗 terrabytes-workshop.github.io
📝 We accept short 4-pages and full 8-pages paper on remote sensing, Earth Observation and machine learning.
📆 Deadline: May 16th
🌍💿 TerraBytes: Towards global datasets and models for Earth Observation
Co-located with ICML 2025 (Vancouver, Canada)
🔗 terrabytes-workshop.github.io
📝 We accept short 4-pages and full 8-pages paper on remote sensing, Earth Observation and machine learning.
📆 Deadline: May 16th
Violets are blue,
Roses are red,
Roses are red,
March 5, 2025 at 8:48 AM
Violets are blue,
Roses are red,
Roses are red,
1. We introduce an alignment algorithm to label automatically a point cloud based on a CAD model.
We show that these pseudo-labels are nearly as effective than real labels, especially when used to annotate more points.
We show that these pseudo-labels are nearly as effective than real labels, especially when used to annotate more points.
January 27, 2025 at 1:30 PM
1. We introduce an alignment algorithm to label automatically a point cloud based on a CAD model.
We show that these pseudo-labels are nearly as effective than real labels, especially when used to annotate more points.
We show that these pseudo-labels are nearly as effective than real labels, especially when used to annotate more points.
Intéressante cette question du DGRI du CNRS sur LinkedIn: que dit la science concernant les modes de financement ?
Une brève recherche m'indique que la concentration des moyens est moins efficace que la dispersion :
direct.mit.edu/qss/article/...
link.springer.com/chapter/10.1...
Dingue !
Une brève recherche m'indique que la concentration des moyens est moins efficace que la dispersion :
direct.mit.edu/qss/article/...
link.springer.com/chapter/10.1...
Dingue !
January 27, 2025 at 9:47 AM
Intéressante cette question du DGRI du CNRS sur LinkedIn: que dit la science concernant les modes de financement ?
Une brève recherche m'indique que la concentration des moyens est moins efficace que la dispersion :
direct.mit.edu/qss/article/...
link.springer.com/chapter/10.1...
Dingue !
Une brève recherche m'indique que la concentration des moyens est moins efficace que la dispersion :
direct.mit.edu/qss/article/...
link.springer.com/chapter/10.1...
Dingue !
📣 Offre de thèse
🧑💻 Je recrute un⋅e doctorant⋅e en génération de procédurale directe et inverse de plans de ville et de bâtiments. 🏙️
Venez combiner apprentissage profond et grammaires pour créer une nouvelle classe de modèles génératifs hybrides et interprétables ! 🤖
📄 audebert.at/files/2025_T...
🧑💻 Je recrute un⋅e doctorant⋅e en génération de procédurale directe et inverse de plans de ville et de bâtiments. 🏙️
Venez combiner apprentissage profond et grammaires pour créer une nouvelle classe de modèles génératifs hybrides et interprétables ! 🤖
📄 audebert.at/files/2025_T...
January 6, 2025 at 4:36 PM
📣 Offre de thèse
🧑💻 Je recrute un⋅e doctorant⋅e en génération de procédurale directe et inverse de plans de ville et de bâtiments. 🏙️
Venez combiner apprentissage profond et grammaires pour créer une nouvelle classe de modèles génératifs hybrides et interprétables ! 🤖
📄 audebert.at/files/2025_T...
🧑💻 Je recrute un⋅e doctorant⋅e en génération de procédurale directe et inverse de plans de ville et de bâtiments. 🏙️
Venez combiner apprentissage profond et grammaires pour créer une nouvelle classe de modèles génératifs hybrides et interprétables ! 🤖
📄 audebert.at/files/2025_T...
« On ne voit aucune bonne raison pour qu’un chercheur académique en sociologie, en économie ou en ingénierie soit dispensé d’obligation d’enseignement. »
November 26, 2024 at 12:29 PM
« On ne voit aucune bonne raison pour qu’un chercheur académique en sociologie, en économie ou en ingénierie soit dispensé d’obligation d’enseignement. »
Si vous êtes prof de CPGE (ou que vous en connaissez), sachez que l'ENSG-Géomatique propose via le LIESSE une formation :
« Intelligence artificielle pour la cartographie » 🤖🗺️ (21/10)
➡️ Inscriptions : liesse.minesparis.psl.eu/course/view....
« Intelligence artificielle pour la cartographie » 🤖🗺️ (21/10)
➡️ Inscriptions : liesse.minesparis.psl.eu/course/view....
August 28, 2024 at 10:20 AM
Si vous êtes prof de CPGE (ou que vous en connaissez), sachez que l'ENSG-Géomatique propose via le LIESSE une formation :
« Intelligence artificielle pour la cartographie » 🤖🗺️ (21/10)
➡️ Inscriptions : liesse.minesparis.psl.eu/course/view....
« Intelligence artificielle pour la cartographie » 🤖🗺️ (21/10)
➡️ Inscriptions : liesse.minesparis.psl.eu/course/view....
July 30, 2024 at 12:15 PM