Lightnews — Scholar-powered news

Sebastian Loeschcke @sloeschcke.bsky.social · Aug 13

🇳🇱 𝗤𝘂𝗮𝗹𝗰𝗼𝗺𝗺 𝗔𝗜 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗜𝗻𝘁𝗲𝗿𝗻𝘀𝗵𝗶𝗽 🇳🇱
Excited to join @qualcomm.bsky.social in Amsterdam as a research intern in the Model Efficiency group, where I’ll be working on quantization and compression of machine learning models.
I’ll return to Copenhagen in December to start the final year of my PhD.

6

Reposted by Sebastian Loeschcke

Stella Frank @scfrank.bsky.social · Jun 13

📯 Best Paper Award at CVPR workshop on Visual concepts for our (@doneata.bsky.social + @delliott.bsky.social) paper on probing vision/lang/ vision+lang models for semantic norms!

TLDR: SSL vision models (swinV2, dinoV2) are surprisingly similar to LLM & VLMs even w/o lang 👀
arxiv.org/abs/2506.03994

1 4 12

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

Thanks to my co-authors David Pitt, Robert Joseph George, Jiawwei Zhao, Cheng Luo, Yuandong Tian, Jean Kossaifi, @anima-anandkumar.bsky.social, and @caltech.edu for hosting me this spring!
Paper: arxiv.org/abs/2501.02379
Code: github.com/neuraloperat...

TensorGRaD: Tensor Gradient Robust Decomposition for Memory-Efficient Neural Operator Training

Scientific problems require resolving multi-scale phenomena across different resolutions and learning solution operators in infinite-dimensional function spaces. Neural operators provide a powerful fr...

arxiv.org

1 5

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

We also show strong results on other PDE benchmarks, including 𝐃𝐚𝐫𝐜𝐲 𝐟𝐥𝐨𝐰 and the 𝐁𝐮𝐫𝐠𝐞𝐫𝐬 equation, demonstrating TensorGRaD’s broad applicability across scientific domains.

1

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

We test TensorGRaD on large-scale Navier–Stokes at 1024×1024 resolution with Reynolds number 10e5, a highly turbulent setting. With mixed-precision and 75% optimizer state reduction, it 𝐦𝐚𝐭𝐜𝐡𝐞𝐬 𝐟𝐮𝐥𝐥-𝐩𝐫𝐞𝐜𝐢𝐬𝐢𝐨𝐧 𝐀𝐝𝐚𝐦 while cutting overall memory by up to 50%.

1

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

We also propose a 𝐦𝐢𝐱𝐞𝐝-𝐩𝐫𝐞𝐜𝐢𝐬𝐢𝐨𝐧 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠 strategy with weights, activations, and gradients in half precision and optimizer states in full precision, and empirically show that storing optimizer states in half precision hurts performance.

1 2

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

We extend low-rank and sparse methods to tensors via a 𝐫𝐨𝐛𝐮𝐬𝐭 𝐭𝐞𝐧𝐬𝐨𝐫 𝐝𝐞𝐜𝐨𝐦𝐩𝐨𝐬𝐢𝐭𝐢𝐨𝐧 that splits gradients into a low-rank Tucker part and an unstructured sparse tensor. Unlike matricized approaches, we prove our tensor-based method converges.

1

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

Recent methods reduce optimizer memory for matrix weights. This includes Low-rank and sparse methods from LLMs that work on matrices. But to use them for Neural Operators, we’d need to flatten tensors, which destroys their spatial/temporal structure and hurts performance.

1

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

These Neural Operators use tensor weights. However, optimizers like Adam store two full tensors per weight, making memory the bottleneck at scale.
TensorGRaD reduces this overhead by up to 75% (𝑑𝑎𝑟𝑘 𝑔𝑟𝑒𝑒𝑛 𝑏𝑎𝑟𝑠), without hurting accuracy.

1

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

Scientific computing operates on multiscale, multidimensional (𝐭𝐞𝐧𝐬𝐨𝐫) 𝐝𝐚𝐭𝐚. In weather forecasting, for example, inputs span space, time, and variables. Neural operators can capture these multiscale phenomena by learning an operator that maps between function spaces.

1

Sebastian Loeschcke @sloeschcke.bsky.social · Jun 3

Check out our new preprint 𝐓𝐞𝐧𝐬𝐨𝐫𝐆𝐑𝐚𝐃.
We use a robust decomposition of the gradient tensors into low-rank + sparse parts to reduce optimizer memory for Neural Operators by up to 𝟕𝟓%, while matching the performance of Adam, even on turbulent Navier–Stokes (Re 10e5).

2 7 29

Reposted by Sebastian Loeschcke

Serge Belongie @serge.belongie.com · Mar 30

Would you present your next NeurIPS paper in Europe instead of traveling to San Diego (US) if this was an option? Søren Hauberg (DTU) and I would love to hear the answer through this poll: (1/6)

NeurIPS participation in Europe

We seek to understand if there is interest in being able to attend NeurIPS in Europe, i.e. without travelling to San Diego, US. In the following, assume that it is possible to present accepted papers ...

docs.google.com

6 160 280

Sebastian Loeschcke @sloeschcke.bsky.social · Mar 8

Visited the beautiful UC Santa Barbara yesterday.

2

Reposted by Sebastian Loeschcke

Zhaochong An @zhaochongan.bsky.social · Feb 11

Thrilled to announce "Multimodality Helps Few-shot 3D Point Cloud Semantic Segmentation" is accepted as a Spotlight (5%) at #ICLR2025!

Our model MM-FSS leverages 3D, 2D, & text modalities for robust few-shot 3D segmentation—all without extra labeling cost. 🤩

arxiv.org/pdf/2410.22489

More details👇

1 7 26

Sebastian Loeschcke @sloeschcke.bsky.social · Jan 28

While Pasadena will be my home, I’ll also be making trips to Austin, the Bay Area, and San Diego. If you’re nearby and up for a chat, reach out—let’s meet up!

Sebastian Loeschcke @sloeschcke.bsky.social · Jan 28

☀️ Moved to Pasadena, California! ☀️
For the next five months, I’ll be a Visiting Student Researcher at Anima Anandkumar's group at Caltech, collaborating with her team and Jean Kossaifi from NVIDIA on Efficient Machine Learning and AI4Science.

2 8

Reposted by Sebastian Loeschcke

Nico Lang @nicolang.bsky.social · Jan 24

Recordings of the SSL4EO-2024 summer school are now released!

This blog post summarizes what has been covered:
langnico.github.io/posts/SSL4EO...

Recordings: www.youtube.com/playlist?lis...

Course website: ankitkariryaa.github.io/ssl4eo/
[1/3]

Screenshot of the course website for "SSL4EO: Self-Supervised Learning for Earth Observation"

3 15 36

Reposted by Sebastian Loeschcke

Serge Belongie @serge.belongie.com · Dec 16

New Starter Pack: Pioneer Centre for AI researchers

5 33

Sebastian Loeschcke @sloeschcke.bsky.social · Dec 12

Come by our poster session tomorrow!
🗓️ West Ballroom A-D #6104
🕒 Thu, 12 Dec, 4:30 p.m. – 7:30 p.m. PST
@madstoftrup.bsky.social and I are presenting LoQT: Low-Rank Adapters for Quantized Pretraining: arxiv.org/abs/2405.16528
#Neurips2024

3 7

Sebastian Loeschcke @sloeschcke.bsky.social · Dec 11

Copenhagen University and Aarhus University meet-up in Vancouver 🇩🇰🇨🇦
#NeurIPS2024

6

Sebastian Loeschcke @sloeschcke.bsky.social · Dec 8

On my way to NeurIPS in Vancouver 🇨🇦
Looking forward to reconnecting with friends and meeting new people. Let me know if you are interested in efficient training, quantization, or grabbing a coffee!
#NeurIPS2024

1 18

Sebastian Loeschcke @sloeschcke.bsky.social · Dec 3

Check out the work our lab in Copenhagen will be presenting at #NeurIPS2024 🌟
@neuripsconf.bsky.social @belongielab.org

Belongie Lab @belongielab.org · Dec 3

With @neuripsconf.bsky.social right around the corner, we’re excited to be presenting our work soon! Here’s an overview

(1/5)

2

Reposted by Sebastian Loeschcke

Belongie Lab @belongielab.org · Nov 25

Here’s a starter pack with members of our lab that have joined Bluesky

Belongie Lab

Join the conversation

go.bsky.app

4 13

Sebastian Loeschcke @sloeschcke.bsky.social · Nov 22

Pre-NeurIPS Poster Session in Copenhagen.
Thanks to the Pioneer Centre for AI and @ellis.eu for sponsoring.
@neuripsconf.bsky.social
#neurips2024

1 13

Reposted by Sebastian Loeschcke

ELLIS @ellis.eu · Nov 22

Check out the ELLIS Pre-NeurIPS Fest event today in...🇩🇰Copenhagen!

ELLIS Unit Copenhagen is holding their event at the Pioneer Center for AI showcasing #NeurIPS posters and other Denmark-affiliated papers in #AI and #ML.

More info: bit.ly/4fRFrAh

ELLIS Pre-NeurIPS Fest 2024: Celebrate, Connect, Collaborate

The ELLIS mission is to create a diverse European network that promotes research excellence and advances breakthroughs in AI, as well as a pan-European PhD program to educate the next generation of AI...

bit.ly

5 24