Abstract:Over the past decades, a fascinating computational phase transition has been identified in sampling from Gibbs distributions. Though, the computational complexity at the critical point remains poorly understood, as previous algorithmic and hardness results all required a constant slack from this threshold.
In this paper, we resolve this open question at the critical phase transition threshold, thus completing the picture of the computational phase transition. We show that for the hardcore model on graphs with maximum degree $\Delta\ge 3$ at the uniqueness threshold $\lambda = \lambda_c(\Delta)$, the mixing time of Glauber dynamics is upper bounded by a polynomial in $n$, but is not nearly linear in the worst case.
For the Ising model (either antiferromagnetic or ferromagnetic), we establish similar results. For the Ising model on graphs with maximum degree $\Delta\ge 3$ at the critical temperature $\beta$ where $|\beta| = \beta_c(\Delta)$, with the tree-uniqueness threshold $\beta_c(\Delta)$, we show that the mixing time of Glauber dynamics is upper bounded by $\tilde{O}\left(n^{2 + O(1/\Delta)}\right)$ and lower bounded by $\Omega\left(n^{3/2}\right)$ in the worst case. For the Ising model specified by a critical interaction matrix $J$ with $\left \lVert J \right \rVert_2=1$, we obtain an upper bound $\tilde{O}(n^{3/2})$ for the mixing time, matching the lower bound $\Omega\left(n^{3/2}\right)$ on the complete graph up to a logarithmic factor.
Our mixing time upper bounds are derived from a new interpretation and analysis of the localization scheme method introduced by Chen and Eldan (2022), applied to the field dynamics for the hardcore model and the proximal sampler for the Ising model. As key steps in both our upper and lower bounds, we establish sub-linear upper and lower bounds for spectral independence at the critical point for worst-case instances.
What problem does this paper attempt to address?
The core problem that this paper attempts to solve is the mixing time of Gibbs distribution sampling problems at the critical point (i.e., the tree - uniqueness threshold) in computational complexity theory. Specifically, for the hardcore model and the Ising model on graphs with maximum degree ∆, at the critical parameters (the critical fugacity λ = λc(∆) in the hardcore model and the critical temperature β = βc(∆) in the Ising model), whether the mixing time of Glauber dynamics can still be efficiently achieved.
### Problem Background
1. **Hardcore Model**:
- For graphs with maximum degree ∆, when fugacity λ < λc(∆), the mixing time of Glauber dynamics is O(n log n).
- When λ > λc(∆), the mixing time of Glauber dynamics grows exponentially, and unless RP = NP, there is no polynomial - time algorithm for approximate sampling.
2. **Ising Model**:
- In the sub - critical region (|β| < βc(∆)), the optimal mixing time of Glauber dynamics is O(n log n).
- In the super - critical region (|β| > βc(∆)), the mixing time of Glauber dynamics slows down, and for the antiferromagnetic Ising model, the sampling problem is computationally difficult; while for the ferromagnetic Ising model, there are efficient approximate sampling algorithms for all β > 0.
### Main Contributions of the Paper
1. **Mixing Time of the Hardcore Model at the Critical Point**:
- **Upper Bound**: For any graph with n vertices and maximum degree ∆ ≥ 3, the upper bound of the Glauber dynamics mixing time of the hardcore model at the critical fugacity is \( O\left(n^{(2 + 4e)}+\frac{4e}{\Delta - 2}\log\Delta\right) \).
- **Lower Bound**: For any fixed maximum degree ∆ ≥ 3, there exists an infinite family of graphs such that for any graph with n vertices, the lower bound of the Glauber dynamics mixing time of the hardcore model at the critical fugacity is \( \Omega(n^{4/3}) \).
2. **Mixing Time of the Ising Model at the Critical Temperature**:
- **Upper Bound**: For any graph with n vertices and maximum degree ∆ ≥ 3, the upper bound of the Glauber dynamics mixing time of the Ising model at the critical temperature is \( O\left(n^{2+\frac{2}{\Delta - 2}}\log n\right) \).
- **Lower Bound**: For any fixed maximum degree ∆ ≥ 3, there exists an infinite family of graphs such that for any graph with n vertices, the lower bound of the Glauber dynamics mixing time of the Ising model at the critical temperature is \( \Omega(n^{3/2}) \).
3. **Ising Model with Interaction Matrix J**:
- For the interaction matrix J with critical 2 - norm ∥J∥2 = 1, the upper bound of the Glauber dynamics mixing time of the Ising model in the critical case is \( O(n^{3/2}\log n) \), and the lower bound of the mean - field Ising model on the complete graph is \( \Omega(n^{3/2}) \).
### Conclusion
Through rigorous mathematical analysis, this paper solves for the first time the mixing time problems of the hardcore model and the Ising model at the critical points, filling an important gap in this field. The research results not only provide new theoretical insights but also offer important references and tools for future research.