- Not surprising that if you take the negative of an image, a deep ANN can't classify it. ANN aren't magic: http://www.kdnuggets.com/2017/04/negative-results-images-flaw-deep-learning.html
- Three interesting real-world applications of machine learning: http://www.informationweek.com/big-data/3-cool-ai-projects/a/d-id/1328666
- At least they admit it is because of biased data-FaceApp apologises for making a racist selfie filter: https://techcrunch.com/2017/04/25/faceapp-apologises-for-building-a-racist-ai/
- What is a "whole brain approach" in ANN? https://www.datanami.com/2017/04/25/startup-patents-whole-brain-ai-approach/
- Probably best to make this technology openly available, might make it easier to develop ways of detecting it: https://techcrunch.com/2017/04/25/lyrebird-is-a-voice-mimic-for-the-fake-news-era/
- A cheat-sheet for Python deep learning libraries: http://www.datasciencecentral.com/profiles/blogs/deep-learning-cheat-sheet-using-python-libraries
- PDF cheat sheet for Python deep learning libraries: https://s3.amazonaws.com/assets.datacamp.com/blog_assets/Keras_Cheat_Sheet_Python.pdf
- On the need for a standard intermediate language for machine learning frameworks: http://www.kdnuggets.com/2017/04/deep-learning-virtual-machine-rule-all.html A role for @ieeecis?
- Machine learning is a cloud thing: https://www.theregister.co.uk/2017/04/27/ai_cloud_vendors_race/
- A crash-course in using machine learning in Python: http://www.kdnuggets.com/2017/05/guerrilla-guide-machine-learning-python.html
- Some of the languages in which you can utilise machine learning: https://www.theregister.co.uk/2017/04/25/building_the_machine_languages_for_you/
- Building a recurrent ANN in TensorFlow: https://medium.com/@erikhallstrm/hello-world-rnn-83cd7105b767
- Half of all jobs will be replaced by AI in ten years: http://www.cnbc.com/2017/04/27/kai-fu-lee-robots-will-replace-half-of-all-jobs.html
Monday, May 1, 2017
Weekly Review 1 May 2017
Below are some of the interesting links I Tweeted about recently.