Researchers at Johns Hopkins and OpenAI derived new equations that optimize training parameters, including parameter count, training corpus size, batch size, and training time on language model performance: https://hubs.ly/H0prWMd0

Metric-Learning-Assisted Domain Adaptation. https://arxiv.org/abs/2004.10963

Clova AI Research's StarGAN v2 (CVPR 2020 + code, pre-trained models, datasets) ⬇️

Paper: https://arxiv.org/abs/1912.01865
Code: https://github.com/clovaai/stargan-v2
Video: https://youtu.be/0EVh5Ki4dIY
Twitter: https://twitter.com/yunjey_choi

SQIL: Imitation Learning via Reinforcement Learning with Sparse Rewards

Reddy et al.:
https://arxiv.org/abs/1905.11108

SLIDES
lecture on "Graph Convolutional Networks"
for the NYU Deep Learning course

Xavier Bresson

https://drive.google.com/file/d/1oq-nZE2bEiQjqBlmk5_N_rFC8LQY0jQr/view

Papers with Code
ICLR 2020 : https://paperswithcode.com/conference/iclr-2020-1/official
#DeepLearning #ICLR #ICLR2020

H/T Cecile G. Tamura

"Attention the mechanism by which a person (or algorithm) focuses on a single element or a few elements at a time.@ArtificialIntelligenceArticles

It’s central both to machine learning model architectures like Google’s Transformer and to the bottleneck neuroscientific theory of consciousness, which suggests that people have limited attention resources, so information is distilled down in the brain to only its salient bits.

Models with attention have already achieved state-of-the-art results in domains like natural language processing, and they could form the foundation of enterprise AI that assists employees in a range of cognitively demanding tasks.
@ArtificialIntelligenceArticles
Bengio described the cognitive systems proposed by Israeli-American psychologist and economist Daniel Kahneman in his seminal book Thinking, Fast and Slow.
@ArtificialIntelligenceArticles
The first type is unconscious — it’s intuitive and fast, non-linguistic and habitual, and it deals only with implicit types of knowledge.

The second is conscious — it’s linguistic and algorithmic, and it incorporates reasoning and planning, as well as explicit forms of knowledge.

An interesting property of the conscious system is that it allows the manipulation of semantic concepts that can be recombined in novel situations, which Bengio noted is a desirable property in AI and machine learning algorithms."
@ArtificialIntelligenceArticles
https://venturebeat.com/2020/04/28/yoshua-bengio-attention-is-a-core-ingredient-of-consciousness-ai/

Bayesian Deep Learning and a Probabilistic Perspective of Generalization
Andrew Gordon Wilson, Pavel Izmailov : https://arxiv.org/abs/2002.08791
#Artificialintelligence #Bayesian #DeepLearning

Memristors -- from In-memory computing, Deep Learning Acceleration, Spiking Neural Networks, to the Future of Neuromorphic and Bio-inspired Computing
Mehonic et al.: https://arxiv.org/abs/2004.14942
#Memristor #Neuromorphic #DeepLearning

ข้อมูลวีดีโอ หาก Model รู้เข้าใจระดับความลึกและรูปทรงจะสามารถ ทำ Augmented เติมเข้าไปในวีดีโอได้อย่างน่าสนใจ

https://www.youtube.com/watch?v=51CQObCd_K0&feature=youtu.be&fbclid=IwAR3UHcxiphy2OnhHpcKZSf4zYB-nW8PHyPHBgxcltw-8SCpi8z0sQ8mGtaw

CNN Explainer: Learning Convolutional Neural Networks with Interactive Visualization
Wang et al.: https://arxiv.org/abs/2004.15004
Interactive visualization in the browser: https://poloclub.github.io/cnn-explainer/
#ConvolutionalNeuralNetworks #DeepLearning #NeuralNetworks

Memristors -- from In-memory computing, Deep Learning Acceleration, Spiking Neural Networks, to the Future of Neuromorphic and Bio-inspired Computing
Mehonic et al.: https://arxiv.org/abs/2004.14942
#Memristor #Neuromorphic #DeepLearning

The AI Economist: Improving Equality and Productivity with AI-Driven Tax Policies
‘It’s just a matter of time before we see “The AI Economist adversary: Using RL to find optimal tax evasion policies”’ — Denny Britz
Zheng et al.: https://arxiv.org/abs/2004.13332
Blog: https://blog.einstein.ai/the-ai-economist/
Q&A: https://salesforce.com/company/news-press/stories/2020/4/salesforce-ai-economist/
#AIGovernance #AIPolicy #DeepLearning

OpenAI releases Jukebox, a machine learning framework that generates music
Github: https://github.com/openai/jukebox/
Paper: https://cdn.openai.com/papers/jukebox.pdf
https://www.marktechpost.com/2020/05/02/openai-releases-jukebox-a-machine-learning-framework-that-generates-music/

TTNet: Real-time temporal and spatial video analysis of table tennis
Voeikov et al.: https://arxiv.org/abs/2004.09927
#ArtificialIntelligence #DeepLearning #MachineLearning

Should data scientists learn JavaScript?

https://www.google.com/amp/s/www.freecodecamp.org/news/should-data-scientists-learn-javascript-e611d45804b8/amp/

Deep learning with graph-structured representations
T.N. Kipf : https://dare.uva.nl/search?identifier=1b63b965-24c4-4bcd-aabb-b849056fa76d
#DeepLearning #Graph #NeuralNetworks