Lecture 11 : Circuit Lower

نویسندگان

  • Shachar Lovett
  • Dongcai Shen
چکیده

There are specific kinds of circuits for which lower bounds techniques were successfully developed. One is small-depth circuits, the other is monotone circuits. For constant-depth circuits with AND,OR,NOT gates, people proved that they cannot compute simple functions like PARITY [3, 1] or MAJORITY. For monotone circuits, Alexander A. Razborov proved that CLIQUE, an NP-complete problem, has exponential-sized circuit complexity [5]. We won’t talk about lower bounds for monotone circuits. Definition 1 (NC) NC is the class of languages solved using circuits of depth O(log n), size poly(n), with AND, OR, NOT gates of fan-in ≤ 2. NC stands for “Nick’s Class” named by Steven Cook in honor of Nick Pippenger. NC includes functions f(x1, · · · , xn) which depend on O(1) inputs. Clearly, NC circuits cannot compute all functions. For example, PARITY cannot be computed by NC because the parity value depends on all n inputs, instead of a constant number of inputs. However, we still don’t know the answer yet: Question 2 P = NC? Note: obviously, NC ⊆ P/poly. Definition 3 (AC) AC is the same as NC except that AND, OR gates allow unbounded fan-ins.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Lecture 1 : Course Overview and Turing machine complexity

1. Basic properties of Turing Machines (TMs), Circuits & Complexity 2. P, NP, NP-completeness, Cook-Levin Theorem. 3. Hierarchy theorems, Circuit lower bounds. 4. Space complexity: PSPACE, PSPACE-completeness, L, NL, Closure properties 5. Polynomial-time hierarchy 6. #P and counting problems 7. Randomized complexity 8. Circuit lower bounds 9. Interactive proofs & PCP Theorem Hardness of approxi...

متن کامل

Lecture 5: Razborov-Smolensky Lower Bounds for Constant-Depth Circuit with MODp Gates

In this lecture, we will talk about circuit lower bound for constant-depth circuit with MODp gates. Using switching lemma, we can prove exponential size lower bound for constant-depth circuit computing parity and majority. What if parity (=MOD2 gates) are allowed? It was conjectured that majority still needs exponential size to compute in constant-depth circuit. Razborov (1987) solves this conj...

متن کامل

2 Monotone Functions and Monotone Circuits

In the last lecture we looked at lower bounds for constant-depth circuits, proving that PARITY cannot be computed by constant-depth circuits, i.e. PARITY / ∈ AC0. General circuit lower bounds for explicit functions are quite weak: the best we can prove after years of effort is that there is a function, which requires circuits of size 5n − o(n). In this lecture we will examine what happens if we...

متن کامل

18.405J S16 Lecture 22: Derandomization Implies Circuit Lower Bounds

Q: Why are we using circuit lower bounds here, as opposed to a claim such as E ̸⊆ P for example? A: The proof of the Nisan-Wigderson pseudorandom generator relies on nonuniformity, by showing that distinguishing a pseudorandom generator implies a circuit for solving a hard problem – this reduction involves hardwiring advice into a circuit in order to solve the hard problem. A contradiction requi...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2013