These slides will be explaining decidability and Rice’s theorem. You can ask if anything is not clear.
Part 2 is added now...
Decidability Slides Updated
Gotham City suffers from something, then BATMAN rises !
If GO people suffers from something, then ARJUN SURESH sir rises !
@Arjun Sir, @Subarna Das & @Sukanya Das Thanks for all your efforts. It was helpful :)
Just a correction (though very minor)
Slide 11 (Property of set) :
Is 1/x an element of N?- satisfied by NO x
But its satisfied only when x=1
@MiNiPanda,
that'll be $\dfrac{1}{2x}$ -- typing mistake
@Arjun Sir, Is this correct interpretation.
2.L = { | L(M) has at most 10000 strings}
Lyes=lang having <= 10000 string
Lno= any lang having more than 10000string
3.L = { | (L(M) is finite}
yes= lang contain finite no of strings
no= any lang contain infinite lang
3.L = { | L(M) = Σ* }
Lyes=?
Lno=?
4.L = { | L(M) is infinite}
yes= lang which produce infinite string
no= any lang having finite no of strings.
5.L = { | L(M) is recursive}
@Arjun
Sir, My understanding is that, membership problem for a language is undecidable, if we cannot write an yes/no algorithm for checking the membership of a string in that language (such a language should be 'not RE' right ?). If a language is RE, then there exist a TM (algorithm) that would surely say at least 'yes', if that string is present in the language.
My doubt is, consider this: L = {<M> | L(M) has at least 10000 strings} (slide 17) , this means that , L is the set of all TM's which accept at least 10000 strings. So, if it is RE , then there should exist an algorithm for checking that condition , i.e we could recursively enumerate all TM's which satisfies this condition. That is , Should it be semi-decidable ? I cannot understand how is it undecidable. Also, I cannot understand how it is different from the semi-decidable problem of 'halting problem' . Here , we could simulate 'M' on a universal TM and check if a TM contains 10000 strings(dovetailing can be used for simultaneous executions) . So UTM would say 'yes' for all the TM's which satisfies this condition. We have an algorithm that at least say 'yes' right ? How is it undecidable ?
I cannot understand where I am going wrong. Please help.
its really great thanks @Arjun @Subarna Das @Sukannya Ma'am
one doubt slide 30 eg of Trivial property
"L(M) can be accepted by TM with even no of states" ==>do mean 2 state universal TM having multi-tape with stay option??
Hi, I was referring this ( https://gatecse.in/rices-theorem/) post on Rice theorem by Arjun sir.
The Rice I theorem is given as:
Any non-trivial property of the LANGUAGE recognizable by a Turing machine (recursively enumerable language) is undecidable
Then in problem, it says:
$L(M)$ has at least 10 strings We can have $T_{yes}$ for $Σ^∗$ and $T_{no}$ for $ϕ$. Hence, L={M∣L(M) has at least 10 strings} is not Turing decidable (not recursive).
$L(M)$ has at least 10 strings
We can have $T_{yes}$ for $Σ^∗$ and $T_{no}$ for $ϕ$. Hence, L={M∣L(M) has at least 10 strings} is not Turing decidable (not recursive).
The theorem talks about recursively enumerable language and empty language $\phi$ is not recursively enumerable (just checked in Ullman). So, is it correct to consider $\phi$ during answer? (Even the slides do consider $\phi$)
ϕ is a Regular language then it should be Recursive enumerable language, right @raj+singh
@Shaik
Ullman says on page 362:
$L_e$ is not RE
It defines $L_e=\{M | L(M) = \phi\}$
I guess this is different from $L=\phi$ and $\phi$ is regular. Seems that I unnecessarily got confused. Right?
@Arjun sir @Subarna Das while doing reductions,A<=B does A follows properties of B or B follows properties of A.These are from same slides.
@Astha Take ANY ONE "yes" case and any one "no" case and see if you can get the SUBSET relation.
Arjun sir plzz make blog how to avoid lil mistakes and different startegies to try perfection what we will do in exam .....means those we dont know dont care but sometimes the questions we know we are making mistakes so if there are 70% ques i know prefectly then only 50% -60 % of them are correct 10- 20 % everytime mistake ....plzz add something to avoid mistakes
Can someone help me understand examples on slide page 23. Not able to get meaning of solutions. For example:
L = { <M> | L(M) = {}] - We can say “no” is M accepts some string, but can we every say “yes” as we need to check infinite number of strings?
@Arjun Sir,Thanks a lot
One doubt regarding Slide no. 30 : Some trivially decidable properties
L = {<M> | L(M) is countable} --> for this language is it correct to argue that since we have diagonalisation procedure, so this problem becomes decidable?
@Arjun sir confused in this one (https://gateoverflow.in/83694/turing-machine).
Understood your answer given there but here can't we create Tyes= {epsilon} Tno= {a,aa,aaa...} ?
For semi-deciding we can say, we may consider a TM having initial state as final state. So for yes instance it can output Yes but cannot say for any NO instance.
These two are different right? The question there is for the second one. For 1, what you told is correct.
64.3k questions
77.9k answers
243k comments
79.7k users