Tweet

@mcstammi

@wightmanr

@DanHendrycks

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

More from @paperswithcode

Papers with Code

@paperswithcode

14 Dec 21

📊 Robustness Reports on ImageNet!

We've indexed AugLy's robustness reports, which show model vulnerabilities to different manipulations. Check it out below:

paperswithcode.com/sota/image-cla…

1/4

Reports shows that EfficientNet not only has higher accuracy on ImageNet, but is also significantly more robust than older models.

Accuracy is one measure of performance, but it doesn’t capture how robust the model is to simple manipulations such as noise and rotation...

2/4

Using AugLy’s image augmentations it's possible to test each model’s vulnerability to different manipulations in the robustness reports.

You can also produce your own robustness reports for models using AugLy and add them to the Papers with Code leaderboards.

3/4

Read 4 tweets

Papers with Code

@paperswithcode

7 Dec 21

Text-Driven Neural Stylization for 3D Meshes

This new paper proposes a text-driven, neural network based approach (Text2Mesh) for stylizing 3D meshes.

Paper & Code: paperswithcode.com/paper/text2mes…

(A short thread) 1/6

Text2Mesh is a framework to stylize meshes by predicting color and geometric details which conform to a target text prompt. Optimization occurs through rendered 2D images & augmentations, then using CLIP similarity to train the network to produce consistent stylized meshes.

2/6

The technique leverages the representational power of CLIP and is capable of generating a variety of different local geometric displacements to synthesize a wide range of styles.

3/6

Read 6 tweets

Papers with Code

@paperswithcode

10 Nov 21

Deep learning models for tabular data continue to improve. What are the latest methods and recent progress?

Let’s have a look ↓

1) Wide&Deep jointly trains wide linear models and deep neural networks to combine the benefits of memorization and generalization for real-world recommender systems. The model was productionized and evaluated on Google Play.

paperswithcode.com/method/wide-de…

2) TaBERT is a pretrained LM that jointly learns representations for natural language sentences and (semi-)structured tables. TaBERT works well for semantic parsing and is trained on a large corpus of 26 million tables and their English contexts.

paperswithcode.com/method/tabert

Read 7 tweets

Papers with Code

@paperswithcode

27 Oct 21

Graph neural networks are driving lots of progress in machine learning by extending deep learning approaches to complex graph data and applications.

Let’s take a look at a few methods ↓

1) A Graph Convolutional Network, or GCN, is an approach for semi-supervised learning on graph-structured data. It’s based on an efficient variant of CNNs which operates directly on graphs and is useful for semi-supervised node classification.

paperswithcode.com/method/gcn

2) Diffusion-convolutional neural networks (DCNN) introduce a diffusion-convolution operation to extend CNNs to graph data. This enables learning of diffusion-based representations. It's used as an effective basis for node classification.

paperswithcode.com/method/dcnn

Read 7 tweets

Papers with Code

@paperswithcode

12 Oct 21

StyleGAN3 is out and results are 🤯!

It proposes architectural changes that suppress aliasing and forces the model to implement more natural hierarchical refinement which improves its ability to generate video and animation.

paperswithcode.com/paper/alias-fr…

1/8

In the cinemagraph below, we can see that in StyleGAN2 the texture (e.g., wrinkles and hairs) appears to stick to the screen coordinates. In comparison, StyleGAN3 (right) transforms details coherently:

2/8

The following example shows the same issue with StyleGAN2: textural details appear fixed. As for alias-free StyleGAN3, smooth transformations with the rest of the screen can be seen.

3/8

Read 8 tweets

Papers with Code

@paperswithcode

4 Oct 21

@wightmanr

In a new paper from @wightmanr et al. a traditional ResNet-50 is re-trained using a modern training protocol. It achieves a very competitive 80.4% top-1 accuracy on ImageNet without using extra data or distillation.

[mini-thread]

paperswithcode.com/paper/resnet-s…

The paper catalogues the exact training settings to provide a robust baseline for future experiments:

It also records training costs and inference times on ImageNet classification between other architectures trained with the proposed ResNet-50 optimized training procedure:

Read 6 tweets

Share this page!

Papers with Code

Try unrolling a thread yourself!

More from @paperswithcode

Papers with Code

Papers with Code

Papers with Code

Papers with Code

Papers with Code

Papers with Code

Did Thread Reader help you today?

Like this author's thread?