🌅 Items tagged "technical"
-
Investigating Transferability in Pretrained Language Models — April 30, 2020
A techincal result showing how best to do transfer learning in NLP.
-
Consistent Video Depth Estimation — April 30, 2020
Making consistent predictions on video.
-
Editing in Style - Uncovering the Local Semantics of GANs — April 29, 2020
Editing generative images semantically
-
GANSpace - Discovering Interpretable GAN Controls — April 6, 2020
Semantically editing images
-
Adaptive Fractional Dilated Convolution Network for Image Aesthetics Assessment — April 6, 2020
Which makes an image beautiful?
-
Learning Object Permanence from Video — March 23, 2020
If you can't see it, is it really there?
-
Privacy-preserving Learning via Deep Net Pruning — March 4, 2020
A relationshpi between data privacy and pruning.
-
A Visual Analysis Tool to Explore Learned Representations in Transformers Models — October 11, 2019
A neat UI for interacting with text models.
-
Improving sample diversity of a pre-trained, class-conditional GAN by changing its class embeddings — October 10, 2019
Make your GAN generate more diverse pictures with this one simple trick!
-
Adversarial Learning of Deepfakes in Accounting — October 9, 2019
Using adversarial networks to trick financial auditors!
-
Learn to explain efficiently via neural logic inductive learning — October 6, 2019
Explaining model decisions with logic!
-
Addressing Failure Prediction by Learning Model Confidence — October 1, 2019
How can we think about and understand "confidence" in ML systems?
-
Gravity as a Reference for Estimating a Person’s Height from Video — September 5, 2019
Computing peoples heights by asking them to jump!
-
Enforcing Analytic Constraints in Neural-Networks Emulating Physical Systems — September 3, 2019
Encoding physical constraints into neural networks.
-
Story-oriented Image Selection and Placement — September 2, 2019
We can learn which images to display next to text by understanding both!
-
Visual Deprojection — September 1, 2019
Using deep learning to "deproject" data!
-
Physics Informed Data Driven model for Flood Prediction — August 23, 2019
Using GANs to speed up flood predictions.
-
Federated Learning — August 21, 2019
How we can train across many distributed devices in a privacy-preserving way.
-
Differentiable Reasoning — August 13, 2019
Using logic to provide context for classification.
-
Visual Search at Pinterest — August 5, 2019
Real-world example of visual search deployed at scale.
-
On the “steerability” of generative adversarial networks — July 16, 2019
Understanding what options levers we have for manipulating images coming from generative networks.
-
Mixed Integer Program as a Layer — July 12, 2019
Building MIP constraints into the optimisation of neural networks.
-
Learning to Self-Correct from Demonstrations — July 12, 2019
Technical result that let's networks generalise more conservatively from examples.
-
Out-of-Distribution Detection using Generative Models — July 10, 2019
Generative models can be used to work out if we're seeing some data that we were never trained on.
-
Detecting the Unexpected — April 16, 2019
How do we know what we don't know? Turns out we can know that.
-
Extreme Image Compression — April 8, 2019
Turns out it's possible to learn improved compression techniques for a given collection of images.
-
Image2StyleGan - aka Ryan Obama aka Oprah Johansson — April 5, 2019
How to make StyleGAN morph really well.
-
Learning how music and images relate — March 30, 2019
We can learn how to think about music and images in the same "space", and thus work out when they are related.
-
Women also Snowboard — March 26, 2018
Techniques to help models be less biased.
All tags:
