Maybe 3.5 stars. Interesting stuff about how the learning process works within the brain, with some contrast/compare of computer learning algorithms. Some suggestions about how schools should restructure educational techniques in light of knowledge about how learning works.

I felt disappointed that there was very little discussion about how what is learned is actually encoded in the brain for future retrieval. The simple fact is that scientists just don’t know how this works, but I would have loved to hear some intelligent informed speculation. Not sure if this is the cautious scientist sticking to scientific facts, or if the mysteries of how knowledge is stored is still so unfathomable that it’s just not worth speculating about. But I guess I expect that a book about how the mind learns would addressthis situation, if at least to say that nobody has any real idea how information is encoded in the brain. ( )

Very interesting. Many of the studies he refers to are also used in similar books like Make It Stick and Why We Sleep. This isn't a bad thing; it's helpful for remembering because it spaces out learning (as all these texts advocate). ( )

It's so fascinating how our brains work! Our bitty baby brains are equipped with learning software built in. We're constantly taking in information and testing our assumptions about the world. Like Dehaene keeps saying, we're all budding scientists!

This book contains good looks at the research underpinning our knowledge of learning. What's important? Attention, engagement, error feedback, consolidation. Basically you need to pay attention to what you want to learn, actually engage with it (don't just let it wash over you - wrestle with it!), understand what your mistakes are and why you made them, and then REST to let it all linger in your brain.
Rest is actually such an important part of learning, and not understanding that is why I thought procrastinating on studying would be fine. Actually all the sleeps you have in between studying help your brain keep knowledge inside and use it for problem solving later. ( )

A machine whatever it does is programmed by humans. So how on earth can it possible be more creative some people like to say? Answer: Our children are created and taught by us; how could they ever be more creative?
It asks a wider question of humanity. Individuals can be creative and able to compute information, ideas and emotions entirely uniquely - but that's not many individuals. For most of us everything we think, feel and do is the composite of external influence, be it upbringing, media, social circles, etc. If humanity's creativity is defined by the individual experience and derived from free will, how is the individual who seems unable to exercise that free will any different from an AI or a machine - programmed to make a predetermined set of responses to various stimuli?
Modern statistical machine learning algorithms are notorious for being incapable of generalising beyond their training set. They require many hundreds of thousands of examples of a concept to learn to represent it, they must "see" these examples many thousands of times and even then they are constitutionally incapable of recognising new instances of the concept that have attributes they have never seen before in their example set.
Compare that with how humans learn: If you so much as describe the idea of a giraffe to a child who has never seen one, the child will then be able to recognise a giraffe when it first sees one, even if it's just a stuffed toy giraffe, a cartoon giraffe, a 3d-animated giraffe, a pink giraffe with green spots, a winged giraffe, a giraffe doctor or a giraffe aviator, etc., etc. We are capable of understanding and generating an extremely broad range of representations of the same object in many different levels of abstraction - like I say, from the plainest caricature to the most intricately detailed image of the real thing. Statistical machine learning algorithms don't even come close.
It could be argued that the creative act of abstraction of new concepts to their elementary components and invention of new representations are essential abilities of the human mind. There is not a shred of a reason to say that machines are more original creators than humans. There is no comparison.
Ultimately, art requires experience. (It requires other things as well, but that would take us off the point.) I think Klingemann is right, in some ways, but he's also wrong, because he is twisting the words. "re-inventing" or "making connections" is simply a way of producing a false equivalence between data inputted into a machine, and "things we have seen." But the machine only has the context of the data being fed into it - we experience things differently - in a mood, as the result of a personal loss or triumph, in an emotional state which can be brought by anything from breaking up with your partner to seeing the results of a famine, and those are channeled through our own self-awareness. If we could give computers the ability to experience, as opposed to simply learn or process data, then there's no reason why eventually they would not feel the need to express themselves in some way that is more than functional. But that's going to take a while.
How many people are creative as Francis Bacon? If we assume an AI reaches the highest point of technological advancement, what differentiates the human from the AI? I'd say it was free will - free will being key to creativity. Brilliant minds are few and far between, many of the individuals who make up humanity are simply composites of external influences - be it upbringing, religion, media, social circle, etc. If you're a composite of these influences, do you exercise free will? Or are your emotional and intellectual responses to the world and the self-predetermined.
(Or more, programmed) by those influences - in which case, everything is learned and nothing you think or feel is unique, creative or indeed human. If we assume that to be the case, what separates the bulk of humanity from an AI or a machine? Nevertheless, Brian Eno gave the best definition of Art that I've heard in his John Peel lecture - Art is anything you don't have to do (drawing being a good case in point: https://manuelaantao.blogspot.com/p/uskp.html - and I get the distinct feeling that the 'you' he was referring to was definitely human. After all, the machine has little choice in doing what it is doing.
Skynet will be operational soon! ( )

"How We Learn" ought to be a text studied for a degree in education, and would be great for new parents as well.. Dehaene draws conclusions from the latest brain studies done on infants and children and argues that education requires more one-on-one interaction between the instructor and the student (or parent and child) and less lecturing--and no grades! Frequent testing and consistent error feedback are essential for learning to take hold but grades are simply demoralizing.

I particularly found the organization of the book helpful. I learned about the neurological function of sleep in memory retention, what is true and false about brain plasticity, the grim realities of trauma and addiction, the affect of music education, the benefit of bilingualism from an early age, and the enormous potential of the small child's mind. In fact, I'll never look at small children or even tiny babies in the same way again after reading this book. When I go to bed, I'm also going to make it a habit to dwell on the positive aspects of the day, or something neat I have learned during the day, just before I go to sleep, in order to maximize the chances that my brain will hold onto it better. Adults can't learn as easily as children, but we can still learn quite a bit with the right techniques.

Machine learning, on the other hand, has a long way to go. As the author points out, some human person must input massive amounts of data in order to get AI to do a few things. We're the opposite: with a few pieces of data, we can do massive numbers of things.

I received an advanced readers copy from the publisher and was encouraged to submit an honest review. ( )