MicrosoftAI raises the bar in text-to-speech with an “almost” unsupervised context, training ONLY 200 speech and text data to generate human-sounding speech for about 20mins - 99.84% world level intelligible rate.
Paper: https://arxiv.org/pdf/1905.06791.pdf
Sample: buff.ly/2X885F9
Paper: https://arxiv.org/pdf/1905.06791.pdf
Sample: buff.ly/2X885F9
Learning Sparse Networks Using Targeted Dropout
A new research paper from
Aidan N. Gomez, Ivan Zhang, Kevin Swersky, Yarin Gal, Geoffrey E. Hinton
https://arxiv.org/abs/1905.13678
@ArtificialIntelligenceArticles
A new research paper from
Aidan N. Gomez, Ivan Zhang, Kevin Swersky, Yarin Gal, Geoffrey E. Hinton
https://arxiv.org/abs/1905.13678
@ArtificialIntelligenceArticles
arXiv.org
Learning Sparse Networks Using Targeted Dropout
Neural networks are easier to optimise when they have many more weights than
are required for modelling the mapping from inputs to outputs. This suggests a
two-stage learning procedure that first...
are required for modelling the mapping from inputs to outputs. This suggests a
two-stage learning procedure that first...
"Automated Speech Generation from UN General Assembly Statements: Mapping Risks in AI Generated Texts"
Bullock et al.: https://arxiv.org/abs/1906.01946
#Computation #Language #AIEthics #AIGovernance #ArtificialIntelligence
@ArtificialIntelligenceArticles
Bullock et al.: https://arxiv.org/abs/1906.01946
#Computation #Language #AIEthics #AIGovernance #ArtificialIntelligence
@ArtificialIntelligenceArticles
A new research paper from Geoffry E.Hinton
Aidan N. Gomez, Ivan Zhang, Kevin Swersky, Yarin Gal
Learning Sparse Networks Using Targeted Dropout
https://arxiv.org/abs/1905.13678
@ArtificialIntelligenceArticles
Aidan N. Gomez, Ivan Zhang, Kevin Swersky, Yarin Gal
Learning Sparse Networks Using Targeted Dropout
https://arxiv.org/abs/1905.13678
@ArtificialIntelligenceArticles
SLIDES
GAUSSIAN PROCESSES
Marc Deisenroth
Department of Computing
Imperial College London
https://drive.google.com/file/d/1Ve_Jrn9f-4IcYxF2Jz5KrDnl1qE_IRqg/view?usp=drive_open
GAUSSIAN PROCESSES
Marc Deisenroth
Department of Computing
Imperial College London
https://drive.google.com/file/d/1Ve_Jrn9f-4IcYxF2Jz5KrDnl1qE_IRqg/view?usp=drive_open
Google Docs
lecture_gaussian_processes.pdf
Search engine for computer vision datasets
https://www.visualdata.io/
https://www.visualdata.io/
ICML Accepted Papers have been posted. https://icml.cc/Conferences/2019/AcceptedPapersInitial
icml.cc
ICML 2019 Schedule
ICML Website
Lecture Notes by Andrew Ng : Full Set
https://www.datasciencecentral.com/profiles/blogs/lecture-notes-by-ng-full-set
https://www.datasciencecentral.com/profiles/blogs/lecture-notes-by-ng-full-set
Datasciencecentral
Lecture Notes by Andrew Ng : Full Set
The following notes represent a complete, stand alone interpretation of Stanford's machine learning course presented by Professor Andrew Ng and originally post…
Mission Moon 3-D: A New Perspective on the Space Race https://www.aitribune.com/book/2018111096
Aitribune
Mission Moon 3-D: A New Perspective on the Space Race | AI Tribune
By: David J. Eicher, Brian May
The story of the lunar landing and the events that led up to it, told in text and visually stunning 3-D images.
The story of the lunar landing and the events that led up to it, told in text and visually stunning 3-D images.
This paper evaluates some of the methods in the context of computer vision, specifically when identifying different types of objects and predicting how far away an object is in images. The new method is called 3D- BoNet.
paper: [https://www.profillic.com/paper/arxiv:1906.01140]
(https://www.profillic.com/paper/arxiv:1906.01140)
paper: [https://www.profillic.com/paper/arxiv:1906.01140]
(https://www.profillic.com/paper/arxiv:1906.01140)
Profillic
Profillic: AI research & source code to supercharge your projects
Explore state-of-the-art in machine learning, AI, and robotics research. Browse papers, source code, models, and more by topics and authors. Connect with researchers and engineers working on related problems in machine learning, deep learning, natural language…
Make music with GANs
GANSynth is a new method for fast generation of high-fidelity audio.
🎵 Examples: https://goo.gl/magenta/gansynth-examples
⏯ Colab: https://goo.gl/magenta/gansynth-demo
📝 Paper: https://goo.gl/magenta/gansynth-paper
💻 Code: https://goo.gl/magenta/gansynth-code
⌨️ Blog: https://magenta.tensorflow.org/gansynth
#artificialintelligence #deeplearning #generativeadversarialnetworks
GANSynth is a new method for fast generation of high-fidelity audio.
🎵 Examples: https://goo.gl/magenta/gansynth-examples
⏯ Colab: https://goo.gl/magenta/gansynth-demo
📝 Paper: https://goo.gl/magenta/gansynth-paper
💻 Code: https://goo.gl/magenta/gansynth-code
⌨️ Blog: https://magenta.tensorflow.org/gansynth
#artificialintelligence #deeplearning #generativeadversarialnetworks
Google
Google Colaboratory
SAVOIAS: A Diverse, Multi-Category Visual Complexity Dataset
By Elham Saraee, Mona Jalal, Margrit Betke : https://arxiv.org/abs/1810.01771v1
GitHub : https://github.com/esaraee/Savoias-Dataset
By Elham Saraee, Mona Jalal, Margrit Betke : https://arxiv.org/abs/1810.01771v1
GitHub : https://github.com/esaraee/Savoias-Dataset
Geometric deep learning is a very exciting new field, but its mathematics is slowly drifting into the territory of algebraic topology and theoretical physics.
Background story: https://towardsdatascience.com/an-easy-guide-to-gauge-equivariant-convolutional-networks-9366fb600b70
Paper: https://arxiv.org/abs/1902.04615
Background story: https://towardsdatascience.com/an-easy-guide-to-gauge-equivariant-convolutional-networks-9366fb600b70
Paper: https://arxiv.org/abs/1902.04615
Machine learning programs created by the Robotics Systems Lab at ETH Zurich teaches the dog-like robot "ANYmal" to run fast and "roll over" when it falls. As seen in Science Robotics:
https://robotics.sciencemag.org/content/4/26/eaau5872
https://robotics.sciencemag.org/content/4/26/eaau5872
Science Robotics
Learning agile and dynamic motor skills for legged robots
Legged robots pose one of the greatest challenges in robotics. Dynamic and agile maneuvers of animals cannot be imitated by existing methods that are crafted by humans. A compelling alternative is reinforcement learning, which requires minimal craftsmanship…
Using machine-learning and sensory hardware, Assistant Professor Alberto Rodriguez and members of MIT's MCube lab have developed a robot that is learning how to play the game Jenga®
Learn more about how this robot combines vision and touch to learn the game of Jenga: https://mitsha.re/fQP630nwST1
Learn more about how this robot combines vision and touch to learn the game of Jenga: https://mitsha.re/fQP630nwST1
MIT News
MIT robot combines vision and touch to learn the game of Jenga
Machine-learning approach could help robots assemble cellphones and other small parts in a manufacturing line.
Wonder how the most powerful #opensource #DeepLearning #AI frameworks are tied together by #PowerAI?
Watch video to find out:
https://mediacenter.ibm.com/media/Tanmay+BakshiA+Power+AI+-+Why+I'm+fascinated+by+the+combination+of+AI+and+deep+learning+frameworks/0_347tcpe2
#PowerAI is a great technology with huge potential in challenges like Call for Code!
Watch video to find out:
https://mediacenter.ibm.com/media/Tanmay+BakshiA+Power+AI+-+Why+I'm+fascinated+by+the+combination+of+AI+and+deep+learning+frameworks/0_347tcpe2
#PowerAI is a great technology with huge potential in challenges like Call for Code!
IBM MediaCenter
Tanmay Bakshi: Power AI - Why I'm fascinated by the combination of AI and deep learning frameworks
Tanmay Bakshi talks about how IBM Power AI can help solve some key challenges with deep learning and allows domain or subject matter experts (SMEs) to create their own deep learning models – without needing coding or deep learning expertise. "With machine…
Drop an Octave: Reducing Spatial Redundancy in Convolutional Neural Networks with Octave Convolution
arxiv.org/abs/1904.05049
arxiv.org/abs/1904.05049