yconst

>Our brains have this giant section dedicated to nothing but recognizing characters in our native written language. Sure we have a lot of brain power going to shape recognition in general, but character recognition is so incredibly trained in our heads that it should be our go to.

Yea, but you still have to read them, i.e. focus this "machinery" on a single word/sentence. So you're working in sequence when trying to identify a bunch of unrelated labels. On the contrary, you can probably differentiate an entire group of icons just by a glance in their direction (provided they are appropriately designed, of course).

Plus reading speed is pretty much the same no matter how many times you repeat the task. Recognition on the other hand improves with repetition.

If you're using only text, you're pretty much stuck with being able to differentiate your icons either by location or by actually reading them. Icons add another component to this.

Totally agree with the color-related comment though.

IANANS btw, so your mileage may vary.

Interesting point. However, I'd like to argue that there is a fine line between the conventional definition of a utility function and the ability of an organism to survive in an environment. The latter is truly open-ended, and not really externally imposed, i.e. different animals/people fare throughout life "optimising" completely different functions (closure?). A quote from a Kozma paper comes to mind: "Intelligence is characterized by the flexible and creative pursuit of endogenously defined goals". I believe this quite well summarises the open-ended nature of the task at hand. As I understand it, indeed, goals, rewards, risks, hazards, they are all in the game and shape the decision-making of the agent, it's "policy". But the way they are formalized for each individual and situation, well, is probably subject to constant redefinition itself.

I found the first comment to this article quite interesting:

"Marcus has a point - even is some of what is said about deep neural networks is incorrect (for instance, they can learn and generalize from very few example, one shot learning).

However, he got it wrong with the answer. The key for machines to reach the symbolic abstraction level is the way we train them. All training algorithms, supervised, unsupervised or reinforcement learning with LSTM rely on the assumption that there is an "utility function" imposed by some external entity. Problem is, by doing so, we are taking away their capacity [of the machines] to make questions and create meaning.

The most important algorithm for learning is "meaning maximization" not utility maximization. The hard part is that we cannot define what is meaning - maybe we can't, I'm not sure. That is something I will be glad to discuss."

Precisely the same question I have as well. Especially given that the energy you could store in a spring or coil could be much more, and you are not limited by one's lifting capacity or maximum height. Anyone have an idea?

"Do I dare to say this somehow shows the relative unimportance of design?"

When design is about cramming some animation in just about everything you can show on the screen, and about how fancy your colors are going to be – yes, I agree, it's relatively unimportant.

But real design is not about that at all. In fact, functionality has been part of the core debate in design, for at least a century now. Ask any designer – not only the ones dealing with UI, and they'll tell you.

"We could also take other data into account, like the user who submitted the article, and generate features indicating things like the karma of the user..."

I'm wondering how helpful could that be in prediction though? Would it actually help if I wish to predict how many upvotes my headline would get, and I add my karma as a feature? I think in fact such features would degrade generalisation performance, as they stand in like placeholders (when training), i.e. high karma users are correlated to higher probability for a "hit story".

It does not elaborate on the "problem of crowds", at all. On the contrary, it offers a rather positive view to the whole spontaneous reaction thing. Which is great, don't get me wrong. However, in this case, I believe the solution could have been much much simpler to begin with. But that's just me :)

So that's all great and trying to help and all, but.. a crowd that yells at the top of their lungs to a driver till he goes in tears (in place of actually guiding him to properly move the vehicle), and thereafter tries to lift the bus using nothing but brute force.. isn't this like creating a problem and then trying to solve it?

A related thought might be, once this is in place, how much work would it be so support Appkit (OS X) apps as well?