Masquer Hunter: Adversarial Occlusion-aware Face Detection
https://arxiv.org/abs/1709.05188v2

best paper awards at Conference NAACL 2019

What's in a Name? Reducing Bias in Bios without Access to Protected Attributes

https://arxiv.org/abs/1904.05233

best paper awards at Conference NAACL 2019

https://naacl2019.org/blog/best-papers/

Analysing Mathematical Reasoning Abilities of Neural Models
Saxton et al.: https://arxiv.org/abs/1904.01557
Code and data: https://github.com/deepmind/mathematics_dataset
#ArtificialIntelligence #MachineLearning #NeuralNetworks

Robots that Learn to Use Improvised Tools
Blog by Annie Xie: https://bair.berkeley.edu/blog/2019/04/11/tools/
#ArtificialIntelligence #DeepLearning #ReinforcementLearning #Robotics

Workshop on Fairness Accountability Transparency and Ethics in Computer Vision at CVPR 2019


https://sites.google.com/view/fatecv/home

Complete ML Study Path On Github
It contains links and resources to learn Tensorflow and Scikit-Learn
https://github.com/clone95/Virgilio @ArtificialIntelligenceArticles

A new Dataset from Stanford of 1000+ knee MRIs.

https://stanfordmlgroup.github.io/competitions/mrnet/

https://www.youtube.com/watch?v=guBrLgRKJuw&feature=share


Intro to TensorFlow for Deep Learning


https://eu.udacity.com/course/intro-to-tensorflow-for-deep-learning--ud187

Katie Bouman's talk at CVPR 2016 is a real treat to watch if you're a computer vision researcher! A nice overview of the CV challenges and a simplified, but clear description of the approach
https://www.youtube.com/watch?v=YgB6o_d4tL8 …


paper: https://arxiv.org/abs/1512.01413
#MachineLearning #DeepLearning

Comments on Michael Jordan’s Essay “Artificial Intelligence: The revolution hasn’t happened yet”
Emmanuel Candes, John Duchi, Chiara Sabatti. « We can summarize the points above with a slogan: cross-validation is not enough https://web.stanford.edu/~jduchi/projects/CandesDuSa19.pdf

CariGANs: Unpaired Photo-to-Caricature Translation
Cao et al.: https://ai.stanford.edu/~kaidicao/carigan.pdf
Code: https://cari-gan.github.io
#ArtificialIntelligence #Caricature #GAN #GenerativeAdversarialNetworks

A 2019 guide to Human Pose Estimation with Deep Learning
https://blog.nanonets.com/human-pose-estimation-2d-guide/

Artificial Intelligence Can Detect Alzheimer’s Disease in Brain Scans Six Years Before a Diagnosis
https://www.ucsf.edu/news/2019/01/412946/artificial-intelligence-can-detect-alzheimers-disease-brain-scans-six-years
paper:
A Deep Learning Model to Predict a Diagnosis of Alzheimer Disease by Using
F-FDG PET of the Brain

https://pubs.rsna.org/doi/10.1148/radiol.2018180958

TensorSpace: A Neural Network 3D Visualization Framework
Build interactive and intuitive model in browsers: https://tensorspace.org
#DeepLearning #MachineLearning #Keras #TensorFlow

@ArtificialIntelligenceArticles

Inferring the quantum density matrix with machine learning
Cranmer et al.: https://arxiv.org/abs/1904.05903
#QuantumPhysics #Physics #ArtificialIntelligence #MachineLearning

Linguistic Knowledge and Transferability of Contextual Representations
Liu et al.: https://arxiv.org/abs/1903.08855
#ArtificialIntelligence #DeepLearning #MachineLearning

CS294-158 Deep Unsupervised Learning
Ilya Sutskever @ilyasut guest lecture on GPT-2: https://youtu.be/X-B3nAN7YRM

#DeepLearning #MachineLearning #UnsupervisedLearning

Artificial Intelligence, deep learning, machine learning — whatever you’re doing if you don’t understand it — learn it. Because otherwise you’re going to be a dinosaur within 3 years . Mark Cuban

There's a new convolution operation in town!

For CNNs there's a proposed "Octave Convolution" (OctConv) which can be used as a direct replacement of plain vanilla convolutions without any adjustments in the network architecture.
The idea of OctConv is that, in images, information is conveyed at different frequencies i.e. high frequencies show fine details whereas low-frequencies show global structures.

The idea then is to factorize the feature maps into a high-frequency/low-frequency feature maps and then reduce the spatial resolutions of the low-frequency maps by an octave. This not only leads to lower memory/computation cost but also to better evaluation results such as accuracy in an image classification task.

Paper: https://export.arxiv.org/pdf/1904.05049