Weekly Review 6 May 2022

1) Controlling fusion plasma with #AI: https://www.wired.co.uk/article/deepmind-ai-nuclear-fusion

2) I agree, people need to put more effort into preparing the data for their #AI, and we need lighter-weight algorithms as well: https://spectrum.ieee.org/andrew-ng-data-centric-ai

3) An older article, on building a chatbot with Python: https://www.kdnuggets.com/2019/12/build-intelligent-chatbot.html

4) Not so much wanting to list the #AI as the inventor for patent purposes, as the company that owns the AI. This makes more sense to me: https://www.odt.co.nz/news/national/legal-challenge-ai-test-case

5) Combining optical diffraction with #DeepLearning to encrypt and decrypt images: https://spectrum.ieee.org/deep-learning-facial-recognition-technology

6) Self-driving cars is one area where explainable #AI is very important: https://spectrum.ieee.org/at-last-a-self-driving-car-that-can-explain-itself

7) The use of #MachineLearning can improve business, but it's not a silver bullet: https://www.kdnuggets.com/2021/12/machine-learning-produce-value-business.html

8) An open source #Ai group who think OpenAI is not open enough: https://spectrum.ieee.org/eleutherai-openai-not-open-enough

9) A smart agent for selecting which cookies to accept: https://www.theregister.com/2022/04/12/cookie_consent_is_broken_and/

10) 12 graphs about the state of #AI in 2022, and what they show: https://spectrum.ieee.org/artificial-intelligence-index

11) #MachineLearning models can leak personal data, but only if the training data set is accessible: https://www.theregister.com/2022/04/12/machine_learning_poisoning/

12) Ordering companies to destroy their #AI models constructed with illegally gathered data: https://spectrum.ieee.org/ai-concerns-algorithmic-destruction

13) More work on the limitations of #DeepLearning: https://spectrum.ieee.org/deep-neural-network