Abstract:In the paper, the occurrence of zeros and ones in the binary expansion of the primes is studied. In particular the statement in the title is established. The proof is unconditional.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to study the occurrence frequencies of 0 and 1 in the binary expansions of prime numbers and prove that there is no correlation between the "majority function" and prime numbers. Specifically, the author aims to prove the following theorem:
\[ \sum_{1 \leq x < N} \Lambda(x) f(x) \approx \frac{N}{2} \]
where:
- \( \Lambda(x) \) is the Von Mangoldt function, which is used to measure the distribution of prime numbers.
- \( f(x) \) is the majority function, which is a Boolean function defined as follows: in the binary representation, if the number of 1s exceeds half, it outputs 1; otherwise, it outputs 0.
### Problem Background
The research motivation of the paper comes from a question proposed by G. Kalai, that is, to study whether the digit patterns in the binary expansions of prime numbers have some special properties. By analyzing the correlation between the Von Mangoldt function and the majority function, the author attempts to answer this question.
### Main Contributions
The author has proven that there is no significant correlation between the majority function and prime numbers, which means that in the binary expansions of prime numbers, the distribution of 0 and 1 is relatively random. To prove this, the author uses Fourier - Walsh spectral analysis, symmetrization techniques, and some relevant estimates about Walsh functions.
### Key Formulas
The key formulas involved in the paper include:
1. **Fourier - Walsh Expansion of the Majority Function**:
\[
f(x) = \sum_{S \subset \{0, \ldots, n - 1\}} \hat{f}(S) w_S(x)
\]
where \( w_S(x) = \prod_{j \in S} \epsilon_j \), \( \epsilon_j = 1 - 2x_j \).
2. **Symmetrization of the Von Mangoldt Function**:
\[
\Lambda_s = \sum_{k = 1}^n \sum_{x \in \Omega_k} \frac{\Lambda(x)}{\binom{n}{k}} 1_{\Omega_k}
\]
where \( \Omega_k = \{x \in \{0, 1\}^n; \sum x_j = k\} \).
3. **Exact Form of the Main Theorem**:
\[
\sum_{1 \leq x < N} \Lambda(x) f(x) = \frac{N}{2} + O(n^{-c} N)
\]
Through these formulas and techniques, the author successfully proves that there is no significant correlation between the majority function and prime numbers, thus solving the problem proposed by G. Kalai.