Video Question Generation via Cross-Modal Self-Attention Networks Learning. arxiv.org/abs/1907.03049

Dependency-aware Attention Control for Unconstrained Face Recognition with Image Sets. arxiv.org/abs/1907.03030

One of the hardest problems in #AI is common sense reasoning. This paper by
Nazneen Fatema Rajani, Bryan McCann, Caiming Xiong, Richard Socher
arxiv.org/abs/1906.02361
Github: (link: https://github.com/salesforce/cos-e)
Blog: (link: https://blog.einstein.ai/leveraging-language-models-for-commonsense/)

15 Of The Best Machine Learning Courses On Coursera For Free
https://hackernoon.com/15-of-the-best-machine-learning-courses-on-coursera-for-free-d83ad23f54d7

Si os interesa, también han publicado un estudio en el que detallan el experimento.
Es este: L. Broussard, K. Bailey, W. Bailey et al.; "New Search for Mirror Neutrons at HFIR" y está disponible en arXiv: arxiv.org/pdf/1710.00767

Blind Universal Bayesian Image Denoising with Gaussian Noise Level Learning. arxiv.org/abs/1907.03029

Financial Time Series Data Processing for Machine Learning. arxiv.org/abs/1907.03010

Get ready for the new upcoming book from Machine Learning Mastery on Generative Adversarial Network (GAN)!
(The image from Image-to-Image Translation with Conditional Adversarial Nets site:
https://phillipi.github.io/pix2pix/)

Real-Time Hair Segmentation and Recoloring on Mobile GPUs

Real-time inference speed on mobile GPUs with high accuracy:

Full size (512×512) in 5.7 ms on iPhone XS with 81.0% IOU accuracy
Small size (256×256) in 6 ms on Pixel 3 with 80.2% IOU accuracy

https://static1.squarespace.com/static/5c3f69e1cc8fedbc039ea739/t/5d0291ea06eb89000122c4b9/1560449515878/24_CVPR2019_Hair_Segmentation_v2.pdf

tf.keras for Researchers: Crash Course

"That's all you need to get started with reimplementing most deep learning research papers in TensorFlow 2.0 and Keras!"

Code by François Chollet: https://colab.research.google.com/drive/14CvUNTaX1OFHDfaKaaZzrBsvMfhCOHIR

#deeplearning #keras #tensorflow #tutorial

Why Responsible
AI Development
Needs Cooperation
on Safety

By Amanda Askell, Miles Brundage and Jack Clark: https://openai.com/blog/cooperation-on-safety/

#ArtificialIntelligence #AIEthics #AISafety #AIGovernance #Governance

SLIDES - Cornell Tech, Learning Machines Seminar, New York City, March 2019
Anticipating the Unseen and Unheard for Embodied Perception
Kristen GraumanUniversity of Texas at Austin
Facebook AI Research

https://www.cs.utexas.edu/%7Egrauman/slides/grauman-cornell2019.pdf

Machine Learning for Health Postodoctoral Position


Open Postdoctoral Positions In Machine Learning for Health
The ML4H lab in in the Department of Computer Science at the University of Toronto, and at the Vector Institute is seeking motivated postdoctoral researcher for the Fall 2019 with a strong background in machine learning. The goal is to push the state-of-the-art in machine learning on the major challenges arising in health and health care.

Researchers will have the potential to participate in and create projects that target basic, translational science that bring novel machine learning techniques towards meaningful applications. Machine learning topics of interest include, but are not limited to, probabilistic modeling, representation learning, deep learning, time-series modelling, generative models, integrating multi-modal data, model interpretability, convex and non-convex optimization.

These topics are inspired by the challenges posed by biomedical data: high-dimensional, multi-modal datasets with missing data, collected under noisy and imperfect conditions, with complex temporal dynamics and a sensitive nature.

Applying
Candidates should send a research proposal, a CV and a cover letter/personal statement including the names of three referees to Dr. Marzyeh Ghassemi and use “ML4H Postdoc Application” in the subject line.

Qualifications
Interested postdoc applicants should have a Ph.D. in machine learning, fairness, applied causality, RL, or statistics with a strong publication record in top conferences such as NeurIPS, ICML, ICLR, AISTATS, AAAI, KDD, AMIA, MLHC, FAT*, etc.

Prior experience working on health-related data is not required, but you must be interested in meaningful applications of your work.

Fast Estimating Pedestrian Moving State Based on Single 2D Body Pose by Shallow Neural Ne... arxiv.org/abs/1907.04361

Bilevel Integrative Optimization for Ill-posed Inverse Problems. arxiv.org/abs/1907.03083

Video Question Generation via Cross-Modal Self-Attention Networks Learning. arxiv.org/abs/1907.03049

Head animation from single shot by #SamsungAI team

Samsung researchers have released a model that can generate faces in new poses from just a single image/frame (for each of face, pose). Done by building a well-trained landmark model in advance & one-shotting from that, using keypoints, adaptive instance norms and GANs. Model performs no 3D face modelling!

ArXiV: https://arxiv.org/abs/1905.08233v1
Youtube: https://www.youtube.com/watch?v=p1b5aiTrGzY

#GAN #CV #DL