18 votes
Accepted

Did 'Eugene Goostman' really pass the Turing test?

There is no "official Turing test" so there's no concept of "officially pass[ing] the test". Turing described a methodology that one might use to evaluate artificial intelligences. The organizers of ...
user avatar
18 votes
Accepted

Does Church-Turing thesis also apply to artificial intelligence?

The Church-Turing thesis says that the informal notion of an algorithm as a sequence of instructions coincides with Turing machines. Equivalently, it says that any reasonable model of computation has ...
user avatar
18 votes
Accepted

Water Jug Problem in AI

Recoverable, as we can pour all water back to the 12L jug to restore the original state, hence any state derived thereof (by following the same steps from the start). The problem is, also, solvable: <...
user avatar
17 votes
Accepted

Why can't we solve the Halting Problem by using Artificial Intelligence?

The halting problem is not a statement about intelligence (human or artificial) it is a statement about the limits of mathematics. It is an historically important example of an undecidable problem. ...
user avatar
10 votes
Accepted

How is a Turing Test defined?

Although Alan Turing is of course a very important computer scientist, the Turing test is only superficially related to computer science. It is more related to philosophy. As long as machines exist ...
user avatar
  • 1,090
9 votes

Over-fitting Always Occurs?

Roughly speaking, over-fitting typically occurs when the ratio $\frac{\text{complexity of the model}}{\text{training set size}} $ is too high. Think of over-fitting as a situation where your model ...
user avatar
9 votes

In what ways can we distinguish between a human and bot behavior?

The most common/obvious way is a challenge-response test that is easy for humans but hard for computers (of course, but not only, CAPTCHA). This kind of test is very effective{1} but falls under the ...
user avatar
  • 1,927
9 votes

Did 'Eugene Goostman' really pass the Turing test?

I think the prizes you're referring to are the Loebner Prize. According to the Wikipedia page (see prior link), the winner for 2014 is 'Rose' by Bruce Wilcox. That program did not win one of the one-...
user avatar
8 votes
Accepted

In principle, what is the relation between Artifical Intelligence and Turing machine?

Turing machines are a model of computation, one way of formally defining the concept of an algorithm. While Turing machines are usually defined using barebones input/output capabilities, it is not ...
user avatar
7 votes

Google DeepDream Elaborated

The idea of DeepDream is this: pick some layer from the network (usually a convolutional layer), pass the starting image through the network to extract features at the chosen layer, set the gradient ...
user avatar
  • 640
7 votes
Accepted

Bidirectional Dijkstra vs Dijkstra

Absolutely yes! your arguments are correct. And, as matter of fact, it is very easy to come up with a graph where Bidirectional Dijkstra would expand more nodes than Unidirectional Dijkstra, following ...
user avatar
7 votes

Does the the undecidability of the Halting Problem eliminate the possibility of 'Hard AI'?

You're understanding of "solve all instances of the Halting Problem" is flawed. All instances means ALL instances. Every single one, ever. For example, no human knows if a Turing Machine verifying ...
user avatar
  • 29.1k
7 votes

Water Jug Problem in AI

I couldn't in a quick Google through the textbooks tell whether recoverability requires immediate, single-move recoverability (ctrl-Z-style) with each undo step being O(1) complexity, and which can be ...
user avatar
6 votes
Accepted

Is it possible to implement a Neural Network using a graph data structure?

Many implementations you can find out in the web are done on matrices (MATLAB for instance) since it provides a compact notation. Haykin's textbook on neural networks takes this approach. Matrices ...
user avatar
  • 1,574
6 votes
Accepted

Cognitive Computing vs Artificial Intelligence?

A brief definition would be: Cognitive computing is the simulation of human thought processes in a computerized model. more detailed explanation: Cognitive computing involves self-learning ...
user avatar
  • 99
6 votes
Accepted

What are the practical uses of ontologies?

I don't know about robotics, but ontologies are part of the standard toolkit for modern expert systems, especially those with a natural language processing component. For example, consider the ...
user avatar
  • 18.9k
6 votes
Accepted

Is there any published paper about automated planning software used in the ESA Rosetta mission?

Well, as far as I can tell, there was no specific automated planning tool for the Rosetta mission as we understand automated planning systems in ICAPS, which is the most prominent conference on ...
user avatar
6 votes

The meaning of discount factor on reinforcement learning

The discount factor does not represent the likelihood to reach the state $s′ $from the state $s$. That would be $p(s'|s,a)$, which is not used in Q-Learning, since it is model-free (only model-based ...
user avatar
  • 481
6 votes

Comparison between IDA* and Recursive best first search

Let me please start by succintly summarizing the behaviour of RBFS. For a thorough explanation of the algorithm refer to the original journal paper: Richard Korf. Linear-space best-first search. ...
user avatar
6 votes
Accepted

Why does an admissible heuristic mean A* is optimal?

With an admissible heuristic The heuristic defines which nodes will be explored first, but does not change the final path found. In your example, the heuristic will cause the path to Z to be ...
user avatar
6 votes
Accepted

How do neural networks create results like its inputs?

These are known as Autoencoders. As you said, these neural networks are trained to produce output that is similar to the input, rather than output a classification of some kind. Internally, they do ...
user avatar
6 votes

What's the input to the decoder in a sequence to sequence autoencoder?

I was wondering the same and just stumbled across a nice tutorial by Quoc V. Le. The following explanation deals with the conditional case since this seems to be the common case. My explanation is ...
user avatar
  • 161
6 votes
Accepted

Why do we use the log in gradient-based reinforcement algorithms?

We often take the logarithm because: Maximizing $\log \Phi(x)$ is equivalent to maximizing $\Phi(x)$, so in maximum-likelihood problems, we can maximize the log of the likelihood instead of ...
user avatar
  • 141k
6 votes

Is the halting problem claim true with the advent of AI

Even if we can design better AI solving problems we did not think possible even 10 years ago, they will always be computer programs. And as you said it yourself, "Alan Turing states that, there can't ...
user avatar
  • 761
6 votes

How does machine learning relate to artificial intelligence?

Machine Learning is a subset of (the scientific field of) Artificial Intelligence. What is ML? Machine Learning is defined by Tom Mitchel: A computer program is said to learn from experience E ...
user avatar
  • 2,300
6 votes

Artificial intelligence and undecidibility

No. The Post correspondence problem is undecidable. That means that no computer program can solve it (in all cases). "Artificial intelligence" is just a computer program. Actually, "artificial ...
user avatar
  • 141k
6 votes

Water Jug Problem in AI

Suppose the leftmost jug had 3 liters, the center jug had 2 liters, the right jug is empty. Suppose you poured the leftmost jug to the center jug. Now the leftmost jug is empty, and the center jug ...
user avatar
5 votes

When should I learn artificial intelligence?

I would say right away. Of course you'll need lots of different subjects, like the one Dave Clarke mentioned. Which ones you need really depends on which flavor of AI you go for. If you for towards ...
user avatar
  • 1,475
5 votes
Accepted

Closed form solution for a single layer linear perceptron

Yes, there is a closed form solution. In the most general terms, $WX = Y$ is a linear equation, so it can be solved as $W = X^{-1}Y$. If $X$ has no inverse, using the pseudo-inverse $X^\dagger = X^T(...
user avatar
  • 1,475

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