61 votes
Accepted

How is it possible that deep neural networks are so easily fooled?

First up, those images (even the first few) aren't complete trash despite being junk to humans; they're actually finely tuned with various advanced techniques, including another neural network. The ...
Ben N's user avatar
  • 2,599
53 votes
Accepted

What is fuzzy logic?

As complexity rises, precise statements lose meaning and meaningful statements lose precision. ( Lofti Zadeh ). Fuzzy logic deals with reasoning that is approximate rather than fixed and exact. This ...
Franck Dernoncourt's user avatar
32 votes
Accepted

How is a deep neural network different from other neural networks?

The difference is mostly in the number of layers. For a long time, it was believed that "1-2 hidden layers are enough for most tasks" and it was impractical to use more than that, because training ...
Disenchanted Lurker's user avatar
28 votes

How is it possible that deep neural networks are so easily fooled?

The images that you provided may be unrecognizable for us. They are actually the images that we recognize but evolved using the Sferes evolutionary framework. While these images are almost impossible ...
Vishnu JK's user avatar
  • 1,072
22 votes

What is fuzzy logic?

Fuzzy logic is based on regular boolean logic. Boolean logic means you are working with truth values of either true or false (or 1 or 0 if you prefer). Fuzzy logic is the same apart from you can have ...
god of llamas's user avatar
17 votes

How is it possible that deep neural networks are so easily fooled?

All answers here are great, but, for some reason, nothing has been said so far on why this effect should not surprise you. I'll fill the blank. Let me start with one requirement that is absolutely ...
Maxim's user avatar
  • 1,947
13 votes

Is the pattern recognition capability of CNNs limited to image processing?

Convolutional Nets (CNN) rely on mathematical convolution (e.g. 2D or 3D convolutions), which is commonly used for signal processing. Images are a type of signal, and convolution can equally be used ...
Eric Platon's user avatar
  • 1,500
12 votes
Accepted

What kind of problems require more than 2 hidden layers?

Formally, a single hidden layer is sufficient to approximate a continuous function to any desired degree of accuracy, so in that sense, you never need more than 1. This is called the Universal ...
NietzscheanAI's user avatar
12 votes
Accepted

What is a deep neural network?

A deep neural network (DNN) is nothing but a neural network which has multiple layers, where multiple can be subjective. IMHO, any network which has 6 or 7 or more layers is considered deep. So, the ...
Dawny33's user avatar
  • 1,371
11 votes

How is it possible that deep neural networks are so easily fooled?

An important question that does not yet have a satisfactory answer in neural network research is how DNNs come up with the predictions they offer. DNNs effectively work (though not exactly) by ...
Harsh's user avatar
  • 1,315
10 votes

How is a deep neural network different from other neural networks?

A deep neural network is just a (feed-forward) neural network with many layers. However, deep belief networks, Deep Boltzmann networks, etc., are not considered (debatable) deep neural networks, as ...
Dawny33's user avatar
  • 1,371
10 votes

What is fuzzy logic?

It's analogous to analogue versus digital, or the many shades of gray in between black and white: when evaluating the truthiness of a result, in binary boolean it's either true or false (0 or 1), but ...
Jeff Puckett's user avatar
8 votes
Accepted

What is the "dropout" technique?

Dropout means that every individual data point is only used to fit a random subset of the neurons. This is done to make the neural network more like an ensemble model. That is, just as a random ...
Matthew Gray's user avatar
  • 4,262
8 votes

What could an oscillating training loss curve represent?

Overview As it has already been observed, your main problem, beside the training related issues like fixing the learning rate, is you have basically no chance to learn such a big model woth such a ...
Nicola Bernini's user avatar
7 votes
Accepted

Are there any learning algorithms as powerful as "deep" architectures?

Have you read the book The Master Algorithm: by Pedro Domingos? He discusses the present day machine learning algorithms... Their strengths, weaknesses and applications... Deep Neural Network ...
conor's user avatar
  • 554
7 votes
Accepted

How many neurons would a network have after a training of 100k small images?

The neural network is typically a set size once it's created. You'd have to create a network big enough for your data-set.
Zakk Diaz's user avatar
  • 366
7 votes

Why are deep neural networks and deep learning insufficient to achieve general intelligence?

Everyone dealing with neural networks misses an important point when comparing systems with human like intelligence. A human takes many months to do anything intelligible, let alone being able to ...
Cem Kalyoncu's user avatar
7 votes
Accepted

How much of a problem is white noise for the real-world usage of a DNN?

The white noise that fools DNNs isn't really white noise. It has been altered in the same way as the synthetic misclassified pictures have been altered. You have to change many input pixels in exactly ...
BlindKungFuMaster's user avatar
7 votes

Why is no activation function used at the final layer of super-resolution models?

I am not into the field of super-resolution, but I think this question applies to general neural network construction. Usually, you try to solve a classification problem or a regression problem with ...
Marcel_marcel1991's user avatar
7 votes

Why do we need target network in deep Q learning?

In DQN that was presented in the original paper the update target for the Q-Network is $\left(r_t + \max_aQ(s_{t+1},a;\theta^-) - Q(s_t,a_t; \theta)\right)^2$ were $\theta^-$ is some old version of ...
David's user avatar
  • 4,770
6 votes

What is the "dropout" technique?

The original paper1 that proposed neural network dropout is titled: Dropout: A simple way to prevent neural networks from overfitting. That tittle pretty much explains in one sentence what Dropout ...
Tshilidzi Mudau's user avatar
6 votes
Accepted

How can the generalization error be estimated?

Generalization error is the error obtained by applying a model to data it has not seen before. So, if you want to measure generalization error, you need to remove a subset from your data and don't ...
rcpinto's user avatar
  • 2,119
6 votes
Accepted

Has any research been done on DNN Music?

The first thing is to define what is a «good» and a «bad» sound. This is an extremely tricky issue, since the networks need numeric inputs. And music is whole bunch of numbers. I know from people ...
Luis's user avatar
  • 538
6 votes
Accepted

Should deep residual networks be viewed as an ensemble of networks?

Imagine a genie grants you three wishes. Because you are an ambitious deep learning researcher your first wish is a perfect solution for a 1000-layer NN for Image Net, which promptly appears on your ...
BlindKungFuMaster's user avatar
6 votes

Why are deep neural networks and deep learning insufficient to achieve general intelligence?

Deep Learning is mostly successful in supervised learning, whereas the brain builds categories mostly in an unsupervised way. We don't yet know how to do that. (Take a look at google brain: 16,000 ...
BlindKungFuMaster's user avatar
6 votes
Accepted

Deep Q-Learning poor convergence on Stochastic Environment

The inputs that you describe seem like they should be sufficient for a DQN-based agent to learn a good strategy for playing Minesweeper, regardless of whether or not the starting layout changes. The ...
Dennis Soemers's user avatar
  • 10.2k
6 votes

What could an oscillating training loss curve represent?

Try lowering the learning rate. Such a loss curve can be indicative of a high learning rate. Due to a high learning rate the algorithm can take large steps in the direction of the gradient and miss ...
Sabyasachi Ghosh's user avatar
6 votes
Accepted

When should you not use the bias in a layer?

The most usual case of bias=False is in layers before/after Batch Normalization with no activators in between. The BatchNorm layer will re-center the data anyway, ...
Kostya's user avatar
  • 2,516
5 votes

Is the pattern recognition capability of CNNs limited to image processing?

The simple answer is "no, they aren't limited to images": CNNs are also being used for natural language processing. (See here for an introduction.) I haven't seen them applied to graphical data yet, ...
Matthew Gray's user avatar
  • 4,262
5 votes
Accepted

What is the purpose of the hidden layers?

"Hidden" layers really aren't all that special... a hidden layer is really no more than any layer that isn't input or output. So even a very simple 3 layer NN has 1 hidden layer. So I think the ...
mindcrime's user avatar
  • 3,757

Only top scored, non community-wiki answers of a minimum length are eligible