fun to see all the discussion re the future of #deeplearning. if you or someone you love has a substantive critique of my recent critique of deep learning, reply with a link here, please. I would love to hear it and plan respond soon. RTs appreciated. #ai #ml
-
I agree about the Berkeley systems paper. I have no idea why it was cited. That paper was about engineering issues for these newer kinds of systems. Here is a similar summary:https://medium.com/intuitionmachine/biologically-inspired-software-architecture-for-deep-learning-e64db295bb2f …
Although Berkeley system paper that I mentions is framed quite differently than my own recent appraisal, it acknowledges challenges in robustness, interpretability, generalization, and spoofability, all central to my own list.
-
-
You are conflating the issues of software architecture with the problems of AGI. Berkley's Ray is a new software architecture to tackle ML work loads. It has nothing to do with the flaws of Deep Learning.
-
This is like saying that the complexities found in Google's software architecture ( https://medium.com/intuitionmachine/google-and-ubers-best-practices-for-deep-learning-58488a8899b6 … ) is evidence that Deep Learning is flawed. This kind of reasoning is absolutely absurd.
- 1 more reply
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.