5 ms·
Frankly, I doubt that neural networks can "reason" in an abstract way. Correct me if I am wrong but it looks like you got what you ordinarily except from a neur
by foxes 8y ago
Frankly, I doubt that neural networks can "reason" in an abstract way. Correct me if I am wrong but it looks like you got what you ordinarily except from a neural network. The network was successfully trained to pick out certain progressions (or properties) in the training images. It's not doing any reasoning at all - just the usual thing neural networks do.
- thomasfedb 8y agoThe big neural network I keep behind my eyes feels otherwise.
- olalonde 8y agoThat's not a very convincing argument given that modern ANNs are only vaguely inspired by the biological neural network behind your eyes.
- thomasfedb 8y agoI think the ultimately important similarity is that of a complex system that consists of (relatively) simple repeating units.
- eivarv 8y agoWhat is reasoning, really? Does the internal operations of "artificial" systems necessarily differ the (automatic) processes of our own mind? And how do we know that we're _really_ reasoning? I don't suffer under some illusion that current machine learning algorithms have reached AGI, consciousness or sentience – but we need to define these terms before arguing about them. It seems to me that there's a general problem of definition when people talk about what systems can and can't do - and I think it all might just be a convoluted form of an "Appeal to nature" [0] in the case of these endless, unoriginal criticisms of current AI-efforts. [0]: https://en.wikipedia.org/wiki/Appeal_to_nature https://en.wikipedia.org/wiki/Appeal_to_nature
- foxes 8y agoI think they roughly define what reasoning is. Deducing something from abstract information. I don't think they have really captured that with this work. They generated some data according to some rules and fed it to a recurrent neural network, which is meant to be good at this. It seems pretty standard. I don't think it gives much insight to how "reasoning" works, or how neural networks work. These things were originally inspired as a crude approximation to how a real brain might work. However a real brain probably has far more interconnections and recurrence. Should I be surprised if it can capture some of the real behavior?
- bhouston 8y ago> I think they roughly define what reasoning is. Deducing something from abstract information. Reasoning is not prediction learned by seeing things happen before exactly in the same way. Moving a hand to catch a ball is not reasoning, it is recognition and prediction based on doing it a bunch of times. Reasoning to know that 4 comes after 3 after seeing that happen a lot is not reasoning, it is prediction again, and even easier than predicting the motion of a ball. Reasoning involves finding a pattern that you haven't seen before a thousand times but you use the elements you know and look to see how they combine to fit the existing problem. Reasoning by its nature is the ability to succeed in a novel situation that requires logic and exploration in search of a problem. Reasoning is successive cycles of experimental prediction with recognition and running an evaluator on top to see which way to explore. (I think this can be coded probably in the near term, but it is more than just a single NN.) That said, if you provide enough situations to a NN it may be able to solve a lot of "reasoning" situations via recognition and prediction, but it will sort of be stupid if anything falls out of its training data -- similar to how NNs can be fooled completely by slightly modified pictures of animals.
- stared 8y agoAbstraction is looking at some properties, while abstracting others (removing, ignoring). Even simple artificial neural networks for computer vision can do so - e.g. ignoring noise, small rotations, reflections or the background. If, or when, it gets to human-level (for a given task) is another question. But saying that computers cannot learn how to abstract data is demonstrably false. Vide https://distill.pub/2017/feature-visualization/ https://distill.pub/2017/feature-visualization/ for internal representations.
- guntars 8y agoI’d say neural networks abstract over successive layers until the answer becomes self evident. That’s not the same as reasoning though.
- trump_supporter 8y agoYou doubt that because you’re a pontificating pseudo intellectual. Abstract reasoning is the entire point. If it is a human, it has an ear, no matter where it appears, except, let’s encode the logic as a convolution/cross correlation. It’s the entire point of teacher/curriculum/etc. It may as well be learning the properties and topology of an algebraic structure using deductive reasoning instead.
- ArtWomb 8y ago>>> I doubt that neural networks can "reason" in an abstract way Unless there is of course some underlying "structure" to our phenomenological reality that affords us these insights ;) I'd take a look at the recent CV paper on "Taskonomy" from the Stanford Graphics Lab. Instead of semantic labeling of training data by human operators. We can begin to apply learning to high resolution sensor data. Such as RGBA, motion and depth images. And uncover hidden emergent relationships inherent in their geometric and physical properties. Since these "laws" are universal. They can essentially be transferred to myriad other tasks in vision. http://taskonomy.stanford.edu/ http://taskonomy.stanford.edu/ It is precisely this type of learning that DeepMind seeks to quantify with its tool. No one can say precisely why we as humans can look at a thing for the first time and comprehend its "thingness": its utility, its purpose, its meaning and social context. But I believe we are far closer than we imagine to robust meta-learning for intelligent machines on a grand scale!
- dimensionality 8y agoI'd say that semantic embeddings already capture abstract conceptual relations, what remains is to figure out how to extract and relate this information unsupervised.