Francesco (@lakjfr) 's Twitter Profile
Francesco

@lakjfr

computer vision / ML / DL script k1ddie

ID: 899610635096424448

linkhttps://francesco-p.github.io/ calendar_today21-08-2017 12:34:24

61 Tweet

48 Followers

107 Following

Francesco (@lakjfr) 's Twitter Profile Photo

Woah another paper out! "Smaller Is Better: An Analysis of Instance Quantity/Quality Trade-off in Rehearsal-based Continual Learning" Accepted at #IJCNN 2022 :) Here's a preprint :D arxiv.org/abs/2105.14106

Francesco (@lakjfr) 's Twitter Profile Photo

Uh oh, it seems that these days are great...I published anoher paper!! In #CVPR2022 T4V workshop. Mmmm and this one is a 𝕊𝕠𝕝𝕠 paper, meaning single authored :D "Simpler is Better: off-the-shelf Continual Learning through Pretrained Backbones" arxiv.org/abs/2205.01586

Francesco (@lakjfr) 's Twitter Profile Photo

I was sleeping, but my co-author Saurav Jha updated me ASAP. I'm happy to say that we won the runner-up paper award at CLVISION workshop of #CVPR2022 We are among the first to investigate continual learning in transformers! link to the paper arxiv.org/abs/2203.13167

I was sleeping, but my co-author Saurav Jha updated me ASAP. 

I'm happy to say that we won the runner-up paper award at CLVISION workshop of #CVPR2022  

We are among the first to investigate continual learning in transformers!

link to the paper arxiv.org/abs/2203.13167
Matthias De Lange (@matthiasdelange) 's Twitter Profile Photo

Congrats to the runner-up paper award, winning $500 from the MATRIX project. "Towards Exemplar-Free Continual Learning in Vision Transformers: an Account of Attention, Functional and Weight Regularization"

Congrats to the runner-up paper award, winning $500 from the MATRIX project.

"Towards Exemplar-Free Continual Learning in Vision Transformers: an Account of Attention, Functional and Weight Regularization"
Francesco (@lakjfr) 's Twitter Profile Photo

w0ah big news! Glad that our paper "MIND: Multi-Task Incremental Network Distillation" has been accepted in the main conference track at #AAAI24 , one of the main main conferences in the world about artificial intelligence. arxiv.org/abs/2312.02916

w0ah big news!

Glad that our paper "MIND: Multi-Task Incremental Network Distillation" has been accepted in the main conference track at #AAAI24 , one of the main main conferences in the world about artificial intelligence. 

arxiv.org/abs/2312.02916
Jade Law (@jadel4w) 's Twitter Profile Photo

Kojima designing the maps and camera positions for Metal Gear Solid with Lego is one of my fave behind the scenes gamedev moments

Francesco (@lakjfr) 's Twitter Profile Photo

We're excited to announce that our paper, “Dynamic Label Injection for Imbalanced Industrial Defect Segmentation,” has been accepted to the VISION workshop at #ECCV2024 , hosted by Apple ! 🎉 Check out our preprint here: arxiv.org/abs/2408.10031 #Apple #industrial #AI

Francesco (@lakjfr) 's Twitter Profile Photo

Built a small Neural Gas playground to visualize how centroids adapt to a dataset and approximate its structure. Simple algorithm, surprisingly expressive. francesco-p.github.io/res/neural-gas…

Built a small Neural Gas playground to visualize how centroids adapt to a dataset and approximate its structure.

Simple algorithm, surprisingly expressive.
francesco-p.github.io/res/neural-gas…