23 Simple Things About DL  

Deep Learning

Deep Learning is a technology that uses computational networks to reduce the amount of data required to perform an analysis.

23 Simple Things What About Deep Learning

1. Deep Learning is not so new. <a style="color: #000000;" href=" removed link " target="true">Deep Learning has been around for a long time.
2. DL systems use vast amounts of computer hardware and software resources ( GPUs, FPGAs, cloud services) to learn.
3. DL is starting to get useful, and increasingly so.
4. DL is a way of artificially generating natural language.
5. DL is not making decisions, it is making the neural network read data to understand how to make a decision.
6. DL is the science of building models to learn and understand the world as it exists.
7. DL models are typically executed on GPU hardware to massively parallelize and achieve the scale required for AI capabilities.
8. DL works in regions of the brain that do not need to work for something like processing speech, finger movements, or lip-reading.
9. DL is an integral component of AI; it has to be. We can't just be adding in new categories as we build products, we need to learn through our products, and that takes learning in time.
10. DL methods are nonlinear models of multiple data types.
11. DL models are typically executed on GPU hardware to massively parallelize and achieve the scale required for AI capabilities.
12. DL is applying the idea of generative models to artificial neural networks and creating artificial neural networks that learn.
13. DL systems will be able to do pretty much everything a human-level AI can do today, and in a far cheaper and smaller size than a human brain.
14. DL does not teach you how to run it, or how to design the models. It does not explain the algorithms and not even teach you how to measure them.
15. DLMs are not on-ramps into the AI/ML revolution but instead provide an extension of our currently limited models of intelligence.
16. DL is a new way of thinking about the world.
17. DL predicts semantic representations and contains conditional probabilistic logic.
18. DL has no limits to its capability.
19. DL is too big of a risk to make yet. But what I think is real is that we are exploring a very powerful new technique that can potentially solve some of the biggest questions that humanity faces and that makes us understand better.
20. DL is clearly impressive. DL techniques can, for example, break humans on a given task (see above) or outperform them in tasks that we have little idea about (see below).
21. DL is a future word of the future when it comes to implementing deep learning on current architecture.
22. DLs are going to change the way we design our systems, and not because we think they're a bad idea.
23. DL intelligence and creativity is based on its power to understand data, unify and combine disparate sources and kinds of data.

Source : <a style="color: #000000;" title="Deep Learning Co UK" href=" removed link " target="true">Deep Learning Co Uk