🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Are you amazed by NeRFs, SIRENs or FFNs but don’t know why they work? 🤔 Then you’re going to love our work 🚀: “A structured dictionary perspective on implicit neural representations” #CVPR2025 which explains how INRs manage to look so crisp! 🤓 📄: bit.ly/3adThjE 🧵 1/9

Tomorrow will be the first AI4Science Day at SwissTech Convention Center 🤝 #MeetInSwitzerland. Come to attend talks, and posters, and meet the EPFL community working on ML applications to science! More info at go.epfl.ch/ai4science

Deep learning phenomenon alert!🔬🚀 “Catastrophic overfitting is a bug but also a feature” 🐛 We find that single-step adversarial training fails due to the interaction between robust 💪 and non-robust 😈 features in the training set 🤯 📄: arxiv.org/abs/2206.08242 🧵 1/10

New *call for postdoctoral fellows* ai4science.epfl.ch for those who want to advance science with machine learning and combine their own ideas and projects with collaboration with laboratories EPFL that excel in the development and applications of ML tools for sciences.

Postdoc position opening in #AI for medicine jointly hosted by ⁦the labs of #EmmanuelAbbé and ⁦Pascal Frossard⁩ within the scope of a ⁦Center for Intelligent Systems CIS-EPFL⁩ supported research program. recruiting.epfl.ch/Vacancies/2529…

Excited to share our paper in nature: We revisit the 50+ year-old maths problem with AI: how efficiently can we multiply two matrices? Surprisingly, the answer is still not known - even for 3x3 matrices! With AI, we discover many new efficient and exact algorithms. 1/13

Neuro-X Institute calls for applications to its new Postdoctoral Fellowships Program. Join us to do pioneering research, fully funded for up to 24 months. go.epfl.ch/Neuro_X_Postdo… Stephanie P. Lacour

Bures-Wasserstein Means of Graphs. (arXiv:2305.19738v1 [stat.ML]) ift.tt/Vg0hbos

Are you amazed by weight interpolation techniques to edit large pre-trained models? 🤔 Then you’re going to love our work 🚀: “Task Arithmetic in the Tangent Space”🏄 which explains why they work and gives a simple way to make them better! 🤓 📄: rb.gy/q9djm 🧵1/9

Don't miss our latest preprint: Fine-tuning in the tangent space enhances the editing of large pre-trained models like CLIP! 🔥 Check it out!

EPFL's School of Engineering is proud of Dr. Isabel Haasler's award for best SICON paper and her presentation at the #SIAMCT23 actu.epfl.ch/news/dr-isabel…

Very happy to share that “Task Arithmetic in the Tangent Space” has been accepted as an oral presentation to NeurIPS Conference !! 🎉🎉 Thank you very much to the fantastic reviewers🙏 This is a great way to end a PhD 🤓

Glad that Task Arithmetic in the Tangent Space will be an Oral at NeurIPS! 😁 Huge thanks to Guillermo Ortiz-Jiménez, Pascal Frossard, and our reviewers! arxiv.org/abs/2305.12827

Wouldn't it be great if we could merge the knowledge of 20 specialist models into a single one without losing performance? 💪🏻 Introducing our new ICML paper "Localizing Task Information for Improved Model Merging and Compression". 🎉 📜: arxiv.org/pdf/2405.07813 🧵1/9

Having troubles with label noise in the training set? 🌧️ Our new ICML work "Pi-DUAL: Using privileged information to distinguish clean from noisy labels" provides an efficient solution! ☂️ 📜: arxiv.org/pdf/2310.06600 🧵1/8

Struggling to integrate structural constraints in graph generative models? 🧩 Our new paper "Generative Modelling of Structurally Constrained Graphs" presents ConStruct to your rescue! 🚀 📜: arxiv.org/abs/2406.17341 🧵1/9

Fine-tuning pre-trained models leads to catastrophic forgetting, gains on one task cause losses on others. These issues worsen in multi-task merging scenarios. Enter LiNeS 📈, a method to solve them with ease. 🔥 🌐: lines-merging.github.io 📜: arxiv.org/abs/2410.17146 🧵 1/11