New Caelifera

New Caelifera

modern methow cabin

Book: On Intelligence

Summary: Jeff Hawkins is one of the creators of the Palm Pilot. He is also someone who harbored a strong desire to understand intelligence throughout his life, trying and failing to get into MIT to study AI, and then going on to entrepreneurial success with the Palm Pilot and Handspring. In 2004 he published “On Intelligence” with science writer Sandra Blakeslee, which is his attempt to come up with a theory of intelligence (note that I refer to the theory as Hawkins’ even though the book is by Hawkins and Blakeslee.) There is much here that is likely controversial. But there is also a lot that, reading this in 2016 when multi-layered neural nets are all the rage, now sounds very prescient.

One of Hawkins’ central ideas is the memory prediction framework. Essentially the brain stores memories and then uses those memories to make predictions which then propagate out to actions and influence memories. He talks a lot about “invariant representations” and hierarchies, with a special focus on cortical columns in the neocortex for carrying all this out.

Rating: Strongly recommend. While there are things I found myself disagreeing with (an argument about parallel computers that Hawkins gives seemed way off base to me), or wishing for more details, the book should spawn your own neurons to fire in interest at the wide range of topics the authors attempts to bring together in support of Hawkins’ theory. There is a tone of outsider seeing things clearly when all the narrow academics couldn’t, but this personally didn’t grate on me much (contrast this with “A New Kind of Science” for example).

Speculative: If you must know, I have my own views on predicting the future. The essential idea is that local entities cannot predict their own future due to the local nature of the laws of physics. But in that discussion I fail to point out that this is only true to the extent that mixing in the laws and uniformity of priors about outside-the-light-cone information causes us to lack predictability. But regularity in law and non trivial priors muck with this and make prediction better than chance. In a sense this is deeply tied with Hawkins’ ideas and a core component of modern machine learning. While there is no free lunch, lunch does seem to come in particular packages and opening the bag results in nearly the same results (sometimes turkey, sometimes egg salad). The brain certainly can use this both in perceiving, but also in creating the models it uses to predict the future.

So let’s push this even further. If prediction is essential, and the local laws of physics limit, but do no eliminate, this, then might one be able to derive fundamental constraints on the memory-prediction framework from basic physics? This is along the lines of David Deutsch’s ideas about deriving the Church-Turing thesis from physics. Indeed maybe the memory-prediction framework provides us with a sort of Church-Turing thesis for intelligence. All intelligence arises from prediction feeding back to memory and action, no matter what the substrate. Hmm, seems not quite hashed out, but interesting to contemplate.

admin February 16, 2016 Leave A Comment Permalink

Book: The Engelbart Hypothesis

Summary: The late Doug Engelbart was the inventor of the mouse, but that is the least interesting thing about him. He is the creator of “The Mother of All Demos”, a 1968 demo that included hypertext, teleconferencing, word processing, hypermedia, and, yes, the mouse. The 1968 demo was so amazing that he was accused of faking the whole thing.

“The Engelbart Hypothesis: Dialogs with Douglas Engelbart” by Valerie Landau, Eileen Clegg, and, yes, Douglas Engelbart, is an interview of Engelbart followed by a series of short pieces by luminaries who knew and were motivated by Engelbart’s vision (Alan Kay and Vint Cerf among others.) The interviews cover a lot of Engelbart’s thoughts on machines augmenting the human intellect and enabling groups of humans to achieve more than they could individually. If you’ve read Augmenting Human Intellect: A Conceptual Framework, Engelbart’s famous whitepaper which ended up leading to the mother of all demos, there is a lot that is presented along those lines. There is also a lot of biographical material here of interest, for example the story about his talk that perhaps indirectly led to Moore’s work on Moore’s law. All in all a fine introduction to these ideas and emphasizes how much the human part of the story motivated Engelbart. The second half interviews give perspective on how others have taken what Engelbart conceived and moved it to the ‘real world’.

Rating: Read Augmenting Human Intellect: A Conceptual Framework. If you find that fascinating, you’ll love this book. Note: the Kindle version of the book has some severe formatting problems. These problems don’t make it totally unreadable but there are footnotes/asides that need to be un-winded from the main flow of the text.

Speculative: Out speculating one of the greatest speculators of the twentieth century seems like a huge task. One of the central original ideas of Engelbart’s augmentation strategy was to work towards augmenting programmers. There are a lot of reasons that this was a great starting point, but ever since I can remember reading about this I’ve wondered what sorts of tools I would use to augment doing theoretical physics. When I try to take apart what I am doing when thinking about a theory problem, it is amazing how much of it is just involved in creating the structures to conceptualize the problem. Each new problem, at least for me, leads to a new domain specific language for the problem. Frustratingly some of the components of this step should be handled more adeptly by a computer. Some people do this by resorting to tools like Mathematica to “play” with examples. But it feels like there is a missing tool which is not so tightly coupled to our math notation, but is both geometric and automatic. Somewhere between a CAD and Mathematica it feels like there is a space for visualized mathematics that a theorist could use to think through their problems (I will admit that I am a geometry guy, not an algebraist. When I meet one the later my mind can barely grasp how they are thinking. Likely they have a tool they would like too.)

admin February 5, 2016 2 Comments Permalink

Book: The Sports Gene

Summary: In grad school a small group of friends would often debate “Nature versus Nurture” over beers. Thing is, I could never remember which side I wanted to defend and so had to make my arguments “flexible”. The Sports Gene: Inside the Science of Extraordinary Athletic Performance by David Epstein feels like that argument restricted to the world of extreme athletic ability. Epstein wants to understand what makes elite athletes and what role genetics plays in these extraordinary athletes. How much does practice matter (are there naturals)? Are there specific genes that one can point to or is it a confluence of multiple genes or does genetics play no role? More difficult, what about regional effects like certain Kenyans being amazing long distance runners? Epstein romps through a large number of sports trying to flexibly suss out the answer to the role of genetics in athletics.

Rating: Worth your time if you like sports. Causality is difficult in this field, and at times Epstein drops back on a data machine gun with anecdotal grenades (those this data and these anecdotes are mostly interesting.) At other times he is careful to not push the causality bar too far. Even in cases like the distinction between woman and men in sports convincing evidence is surprising hard to come by. Discussion of the variance of the 10,000 hour rule was my favorite part (along with observations about why baseball hitters are good: hint it is not reaction time!)

Speculation: Thinking about the difficulty of experiments in this field: will there someday be advertisements in magazines “Couples: Participate in a Study Identifying Genes for Long Distance Running?”

admin February 1, 2016 Leave A Comment Permalink