@FHKPetersen

Machine learning researcher, postdoc. Investigating differentiable algorithms etc. @Stanford @StanfordAILab.

 Bay Area
petersen.ai
Joined November 2016
  • Tweets171
  • Following298
  • Followers2.3K
91Media
Excited to share our NeurIPS 2024 Oral, Convolutional Differentiable Logic Gate Networks, leading to a range of inference efficiency records, including inference in only 4 nanoseconds 🏎️. We reduce model sizes by factors of 29x-61x over the SOTA. Paper: arxiv.org/abs/2411.04732
19
243
37
1,538
157,818
Join us today for Ubiquitous Ultra-Efficient Mobile Video Super-resolution + Denoising with Logic Gates at #ICCV2025 booth 247. 🌺
1
9
866
I'm excited to share that our work on convolutional differentiable logic gate networks was covered by MIT Technology Review. 🎉 technologyreview.com/2024/12… @StefanoErmon @HildeKuehne

The next generation of neural networks could live in hardware

Researchers have devised a way to make computer vision systems more efficient by building networks out of computer chips’ logic gates.

technologyreview.com
2
23
107
10,431
The next generation of neural networks could live in hardware trib.al/65TrkHz

The next generation of neural networks could live in hardware

Researchers have devised a way to make computer vision systems more efficient by building networks out of computer chips’ logic gates.

technologyreview.com
1
7
1
32
7,559
Felix Petersen retweeted
Alfredo Canziani
@alfcnz13 Dec 2024
Convolutional Differentiable Logic Gate Networks @FHKPetersen
10
123
7
1,568
105,315
Join us today for our Oral presentation at 10:20am in East Ballroom A, B and our poster at 11am-2pm in East Exhibit Hall A-C #2411.
Felix Petersen
@FHKPetersen11 Nov 2024
Excited to share our NeurIPS 2024 Oral, Convolutional Differentiable Logic Gate Networks, leading to a range of inference efficiency records, including inference in only 4 nanoseconds 🏎️. We reduce model sizes by factors of 29x-61x over the SOTA. Paper: arxiv.org/abs/2411.04732
1
1
13
1,894
Join us now at our poster #1502 in East Exhibit Hall A-C 1502.
Felix Petersen
@FHKPetersen3 Dec 2024
Ever wondered how training dynamics differ between LLMs 🖋️ and Vision 👁️ models? We explore this and close the gap between VMs and LLMs in our #NeurIPS2024 paper "TrAct: Making First-layer Pre-Activations Trainable". Paper📜 arxiv.org/abs/2410.23970 Video🎥 youtu.be/ZjTAjjxbkRY
1
18
1,601
Join us on Wednesday, 11am-2pm for our poster on Newton Losses in West Ballroom A-D #6207. neurips.cc/virtual/2024/post…
Felix Petersen
@FHKPetersen28 Nov 2024
I'm excited to share our NeurIPS 2024 paper "Newton Losses: Using Curvature Information for Learning with Differentiable Algorithms" 🤖. Paper link 📜: arxiv.org/abs/2410.19055 Video link 🎥: youtube.com/watch?v=7aFPIMqS… @StanfordAILab 1/6 🧵
11
1,153
Ever wondered how training dynamics differ between LLMs 🖋️ and Vision 👁️ models? We explore this and close the gap between VMs and LLMs in our #NeurIPS2024 paper "TrAct: Making First-layer Pre-Activations Trainable". Paper📜 arxiv.org/abs/2410.23970 Video🎥 youtu.be/ZjTAjjxbkRY
2
64
2
309
20,991
more replies
...and it speeds up overall training by factors ranging from 1.25x (for large ViT pre-training) to 4x (for ConvNets). We benchmark TrAct on a suite of 50 experimental settings.
1
7
693
Learn more in our paper 📑 (arxiv.org/abs/2410.23970) and check out our paper video directly here on 𝕏! --- Thanks to my co-authors Christian Borgelt and @StefanoErmon. @StanfordAILab
5
617
I'm excited to share our NeurIPS 2024 paper "Newton Losses: Using Curvature Information for Learning with Differentiable Algorithms" 🤖. Paper link 📜: arxiv.org/abs/2410.19055 Video link 🎥: youtube.com/watch?v=7aFPIMqS… @StanfordAILab 1/6 🧵
5
42
1
203
19,140
more replies
Also a shout-out to the methods we build on: Perturbed Optimizers @qberthet @mblondel_ml @BachFrancis NeuralSort @adityagrover_ @ecjwg SoftSort @eisenjulian 5/6 🧵
1
3
791
Check out our 5 minute paper video directly here on 𝕏! 6/6 🧵
1
3
11
2,566
Load more