Top 10: Data Science and Machine Learning Articles in Aug
Top 10 most popular posts in Data Science and Machine Learning made in the month of Aug.
- Detecting Money Laundering bit.ly/2a7JHh1 #MachineLearning #DataScience pic.twitter.com/hXZVnsjUGX
- #MachineLearning #DataScience to Predict Out-Of-Sample Performance of Trading Algorithms bit.ly/2bw6IbN pic.twitter.com/NEU2zmetkG
- Massive #DataScience study of chess games reveals how and why humans make mistakes https://bit.ly/29U8zKK pic.twitter.com/2a4Bjl7sKa
- The Theorem Every Data Scientist Should Know https://bit.ly/2bhzFsK #MachineLearning #DataScience pic.twitter.com/BamNlGdFB5
- Bayesian Statistics explained to Beginners in Simple English https://bit.ly/29UgR5g #MachineLearning #DataScience pic.twitter.com/aCuQ02kbpc
- Teaching an #AI to write Python code with Python code https://bit.ly/2bhxFAX #MachineLearning #DeepLearning pic.twitter.com/Kjaose0nsG
- Deep Reinforcement Learning: Pong from Pixels https://bit.ly/29WKC1q #MachineLearning #DataScience pic.twitter.com/zcTitGeQlC
- Amazon goes open source with #MachineLearning #DataScience tech, competing with #TensorFlow https://bit.ly/29WLL9q pic.twitter.com/fIDf1vBvPA
- “Adversarial images” that fool a #MachineLearning #AI vision algorithms https://bit.ly/2bKW6pU pic.twitter.com/pFqCVg7KdR
- The Theorem Every Data Scientist Should Know (Part 2) https://bit.ly/2ajPBLJ #MachineLearning #DataScience pic.twitter.com/WlzN4xiyOt
Retired. Formerly worked with Marvell Semiconductor, Electronic Warfare Associates, Imaginic Inc, McAndrews, Held & Malloy Ltd, Integrated Yield Solutions, and other great companies and people.
8 年Hi Mike. About “Adversarial images” that fool a #MachineLearning #AI vision algorithms: I looked at the images in the related paper. The authors modify the image set to be "grainy," compared to the original images, and then they apply those modified images to a neural network with very small convolution filters (2x2 or 3x3) [Google's Inception-v3]. The grainy character of the images seems to prevent edges to be detected correctly for each color plane since the grain size is the same or even larger than the filter size. My guess is that the original GoogLeNet network would have performed significantly better on the grainy image set. In general, object detection in low-light and night vision applications faces similar problems.
#TheFunctionalFuturist: Exploring the spaces between #webAI and #tinyML. An #Educator for #Makers everywhere to learn #tensorflowJS, #machineLearning, #robotics and #IoT using #edgeAI.
8 年Hi Mike. This looks good but can you make the links clickable