AroraBarak-errata

From CRYPTUTOR

Jump to: navigation, search

Contributed Errata and Suggestions for Arora-Barak webdraft

The version referred to is "Web draft 2007-01-08 21:59."

Contents

Errata

Chapter 1

  • Section 1.2.1: In the definition of a TM, Γ,Q are finite sets. After Figure 1.2, it should be \delta(q,(\sigma_1,\ldots,\sigma_k)) (not \delta(q,(\sigma_1,\ldots,\sigma_{k+1}))) and z\in\{\mathsf{L},\mathsf{S},\mathsf{R}\} (not \{\mathsf{L},\mathsf{SR}\}. On the next page, right at the top it says "The machine will be in state q', and the k + 1 heads will move Left, Right or Stay..." The machine only has k heads. Also, may want to explicity say that k\ge 2 (first tape being input, and last being output).

Chapter 2

  • Page 2.10: At the bottom of the page (item 2), it should be (\triangleright,\triangleright,q_{\mathsf{start}}) (not (\triangleright,\square,q_{\mathsf{start}})).
  • Definition 2.21: \textrm{coNP} = \{ L : \overline{L} \in \textrm{NP} \}, not \overline{L} \in \textrm{P}.
  • Problem 2.2: The theorem to be proved is 2.9 not 2.
  • Problem 2.11: "...certificates of f is equal to..." should be "...certificates of x is equal to..."

Chapter 3

  • Proof of Thm 3.3: Overloading of f.

Chapter 4

  • Example 4.9: x = y (not x\not=y).
  • Section 4.3.1: First paragraph, "PSPACE = NSPACE" should read "PSPACE = NPSPACE".
  • Theorem 4.11: p4.8 line 2 - the second C' should be C", line 6 - the C' s should be C" s and the last C" should be C'.
  • Proof of Theorem 4.12: line -3 - the algorithm makes "M" recursive invocations, not "n". (Also M is overloaded in the proof).
  • Proof of Theorem 4.18: line 3 - reachable from "s", not "u".

Chapter 5

  • Remark 5.5: \Pi_1^p = \textrm{coNP} (not \Pi_2^p)
  • Proof of Theorem 5.15, page 98: "using choices c_1,\dots,x_m" should be "using choices c_1,\dots,c_m"

Chapter 6

  • Definition 6.21: depth O(login) (not O(logdn)), and constant fan-in. Constant fan-in was not mentioned in the definition (perhaps that was the purpose of d?).
  • Before Theorem 6.24: "classes AC, NC" should just be NC. Should point out that AC = NC before that.
  • Problem 6.11: NC^1 \subseteq L (not NC1 = L).
  • Definition 6.28: "i'th but" should be "i'th bit".

Chapter 7

  • Proof of Theorem 7.10: k = 8 | x | 2c + d (not 8 | x | 2d + c). Also, Corollary 7.11 should have e^{-\frac{\delta^2}{2p}k} (not e^{-\frac{\delta^2}{4}pk}). Then plugging in p\le 1 (not p\ge\frac12), (and δ = nc / 2), the bound is e^{-\frac{n^{-2c}}{8}8n^{2c+d}} (not e^{-\frac1{4n^{-2c}} \frac12 8n^{2c+d}}).
  • Proof of Theorem 7.18: probability that Br happens is at most 2nk (not (1 − 2n)k).

Chapter 8

  • Definition 8.10: The value of the probability expression should be 2 − 2k not 2 − 2n.
  • Proof of Claim 8.13.1: In the last line, \frac12 \frac{|S|^2}{2^{2k}} (not \frac12 \frac{|S|^2}{2^{k}}).

Chapter 9

  • Example 9.1: Should be "induced subgraphs" (not "subgraphs").
  • Proof of Theorem 9.8: Clause gadget uses a pair of parallel edges (between the two lower nodes). Needs to be fixed by inserting a node with self-loop in the inner edge. XOR gadget seems to have errors: v' should have an outgoing edge and v should have an incoming edge to be consistent with the symbolic representation.
  • pg 9.12 (before Proof of Lemma 9.18): φ + ψ(z) should be \left( (z_0=z_{n+1}=\ldots=z_m=0 \wedge \phi(z_1,\ldots, z_n)=0) \vee (z_0=1 \wedge \phi(z_1,\ldots, z_m)=0) \right) (assuming m\ge n).

Other Chapters

  • Section 18.2.1, Equation (2): \varphi \not\in {\mathsf{3SAT}} (not \varphi\not\in L.
  • Example 11-6: f_{k-1}(x_1, \ldots, x_{2^{k-1}-1}) and f_{k-1}(x_{2^{k-1}}, \ldots, x_{2^k}) should be f_{k-1}(x_1, \ldots, x_{2^{k-1}}) and f_{k-1}(x_{2^{k-1}+1}, \ldots, x_{2^k}), respectively (i.e, off-by-one). Base case should also be k = 0, not 1.
  • Theorem 22.1: Should be "μ(f) − 1 is a lowerbound" (instead of μ(f)).

Appendices

  • A.4.1 Non-Prime Fields, second paragraph, where "For every prime q and k..." should be "For every prime q and every k \in \N... "

Some Suggestions

  • Complexity Zoo: See examples from lectures: 1 2 3
  • Section 3.3: Suggested figure/explanation for non-deterministic time hierarchy: slide.
  • Chapter 8: May want to add more details on AM. (e.g., Lecture adapted from Du and Ko; using "Alternating Threshold TMs.")
Personal tools