analyze-diffusion-dynamics
について
このClaude Skillは、確率微分方程式とフォッカー・プランク方程式を用いて拡散過程を解析します。確率密度の時間発展と平均初到達時間を計算し、パラメータ感度分析を実行します。閉形式解をシミュレーションで検証する必要がある場合や、ドリフト・拡散パラメータがシステムの挙動にどのように影響するかを研究する際にご利用ください。
クイックインストール
Claude Code
推奨npx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/analyze-diffusion-dynamicsこのコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします
ドキュメント
Analyze Diffusion Dynamics
Characterize behavior of diffusion processes. Specify SDEs. Derive Fokker-Planck equation. Compute first-passage time distributions analytic or numeric. Perform parameter sensitivity analysis. Validate against Monte Carlo simulation.
When Use
- Deriving probability density evolution of continuous-time diffusion process
- Computing mean first-passage times or full first-passage time distributions for bounded diffusion
- Analyzing how drift, diffusion coefficient, boundary parameters affect process behavior
- Validating closed-form solutions against stochastic simulation
- Building intuition for dynamics underlying drift-diffusion models or generative diffusion processes
Inputs
- Required: SDE specification (drift function, diffusion coefficient, domain/boundaries)
- Required: Parameter values or ranges for drift and diffusion functions
- Required: Boundary conditions (absorbing, reflecting, mixed)
- Optional: Time horizon for transient analysis (default: auto-detect from dynamics)
- Optional: Spatial discretization resolution for numerical PDE solvers (default: dx=0.001)
- Optional: Number of Monte Carlo trajectories for simulation validation (default: 10000)
Steps
Step 1: Specify SDE Model
Define drift function, diffusion coefficient, boundary conditions for process.
- Write SDE in standard Ito form:
dX(t) = mu(X, t) dt + sigma(X, t) dW(t)
where mu is drift function, sigma is diffusion coefficient, W(t) is standard Wiener process.
- Implement SDE components in code:
import numpy as np
class DiffusionProcess:
"""A one-dimensional diffusion process specified by drift and diffusion functions."""
def __init__(self, drift_fn, diffusion_fn, lower_bound=None, upper_bound=None,
boundary_type="absorbing"):
self.drift = drift_fn
self.diffusion = diffusion_fn
self.lower_bound = lower_bound
self.upper_bound = upper_bound
self.boundary_type = boundary_type
# Example: Ornstein-Uhlenbeck process on [0, a]
ou_process = DiffusionProcess(
drift_fn=lambda x, t: 2.0 * (0.5 - x), # mean-reverting drift
diffusion_fn=lambda x, t: 0.1, # constant diffusion
lower_bound=0.0,
upper_bound=1.0,
boundary_type="absorbing"
)
# Example: Standard DDM (constant drift and diffusion)
ddm_process = DiffusionProcess(
drift_fn=lambda x, t: 0.5, # drift rate v
diffusion_fn=lambda x, t: 1.0, # unit diffusion (s=1, convention)
lower_bound=0.0, # lower absorbing boundary
upper_bound=1.5, # upper absorbing boundary (a)
boundary_type="absorbing"
)
- Define initial condition:
# Point source at x0
x0 = 0.75 # starting point (e.g., midpoint between boundaries for DDM with z=a/2)
# Or a distribution
initial_distribution = lambda x: np.exp(-50 * (x - 0.75)**2) # narrow Gaussian
- Validate parameter consistency:
def validate_process(process, x0):
"""Check that the SDE specification is self-consistent."""
assert process.lower_bound < process.upper_bound, "Lower bound must be less than upper bound"
assert process.lower_bound <= x0 <= process.upper_bound, \
f"Initial position {x0} outside bounds [{process.lower_bound}, {process.upper_bound}]"
test_drift = process.drift(x0, 0)
test_diff = process.diffusion(x0, 0)
assert np.isfinite(test_drift), f"Drift is not finite at x0={x0}"
assert test_diff > 0, f"Diffusion coefficient must be positive, got {test_diff}"
print(f"Process validated: drift={test_drift:.4f}, diffusion={test_diff:.4f} at x0={x0}")
validate_process(ddm_process, x0=0.75)
Got: Fully specified SDE with finite drift values, strictly positive diffusion coefficient, initial condition within domain boundaries.
If fail: Diffusion coefficient zero or negative at any point in domain? Process degenerate -- check functional form. Drift infinite at boundary? Consider whether reflecting boundary more appropriate.
Step 2: Derive Fokker-Planck Equation
Convert SDE to its equivalent partial differential equation for probability density.
- Write Fokker-Planck equation (FPE) for transition density p(x, t):
dp/dt = -d/dx [mu(x,t) * p(x,t)] + (1/2) * d^2/dx^2 [sigma(x,t)^2 * p(x,t)]
- For constant coefficients (standard DDM case), this simplifies to:
dp/dt = -v * dp/dx + (s^2 / 2) * d^2p/dx^2
- Implement numerical solution of FPE via finite differences:
from scipy.sparse import diags
from scipy.sparse.linalg import spsolve
def solve_fokker_planck(process, x0, t_max, dx=0.001, dt=None):
"""Solve the FPE numerically using Crank-Nicolson scheme."""
x_grid = np.arange(process.lower_bound, process.upper_bound + dx, dx)
N = len(x_grid)
if dt is None:
max_sigma = max(process.diffusion(x, 0) for x in x_grid)
dt = 0.4 * dx**2 / max_sigma**2 # CFL-like stability condition
# Initial condition: narrow Gaussian centered at x0
p = np.exp(-((x_grid - x0)**2) / (2 * (2*dx)**2))
p[0] = 0 # absorbing boundary
p[-1] = 0 # absorbing boundary
p = p / (np.sum(p) * dx)
t_steps = int(t_max / dt)
survival = np.zeros(t_steps)
density_snapshots = []
for step in range(t_steps):
mu_vals = np.array([process.drift(x, step*dt) for x in x_grid])
sigma_vals = np.array([process.diffusion(x, step*dt) for x in x_grid])
D = 0.5 * sigma_vals**2
# Finite difference operators (interior points)
advection = -mu_vals[1:-1] / (2 * dx)
diffusion_coeff = D[1:-1] / dx**2
main_diag = 1 + dt * 2 * diffusion_coeff
upper_diag = dt * (-diffusion_coeff[:-1] - advection[:-1])
lower_diag = dt * (-diffusion_coeff[1:] + advection[1:])
A = diags([lower_diag, main_diag, upper_diag], [-1, 0, 1], format="csc")
p[1:-1] = spsolve(A, p[1:-1])
p[0] = 0
p[-1] = 0
survival[step] = np.sum(p[1:-1]) * dx
if step % (t_steps // 10) == 0:
density_snapshots.append((step * dt, p.copy()))
return x_grid, survival, density_snapshots
- Run and plot evolving density:
import matplotlib.pyplot as plt
x_grid, survival, snapshots = solve_fokker_planck(ddm_process, x0=0.75, t_max=5.0)
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))
for t_val, density in snapshots:
ax1.plot(x_grid, density, label=f"t={t_val:.2f}")
ax1.set_xlabel("x")
ax1.set_ylabel("p(x, t)")
ax1.set_title("Fokker-Planck Density Evolution")
ax1.legend()
t_vals = np.linspace(0, 5.0, len(survival))
ax2.plot(t_vals, survival)
ax2.set_xlabel("Time")
ax2.set_ylabel("Survival probability")
ax2.set_title("Survival Probability S(t)")
fig.tight_layout()
fig.savefig("fokker_planck_solution.png", dpi=150)
Got: Density starts as narrow peak at x0, spreads and drifts according to SDE coefficients, gradually decays as probability absorbed at boundaries. Survival probability decreases monotonic from 1 toward 0.
If fail: Density develops oscillations or negative values? Time step too large -- reduce dt. Density does not decay (survival stays near 1)? Boundaries may be too far from x0 or drift pushes away from both boundaries. Check boundary conditions in solver.
Step 3: Compute First-Passage Time Distributions
Derive distribution of times at which process first reaches boundary.
- Compute first-passage time density from survival function:
def first_passage_time_density(survival, dt):
"""FPT density is the negative derivative of survival probability."""
fpt_density = -np.gradient(survival, dt)
fpt_density = np.maximum(fpt_density, 0) # enforce non-negativity
return fpt_density
- For standard DDM with constant drift, use known analytic solution:
def ddm_fpt_upper(t, v, a, z, s=1.0, n_terms=50):
"""Analytic FPT density at the upper boundary for constant-drift DDM.
Uses the infinite series representation (large-time expansion).
"""
if t <= 0:
return 0.0
density = 0.0
for k in range(1, n_terms + 1):
density += (k * np.pi * s**2 / a**2) * \
np.exp(-v * (a - z) / s**2 - 0.5 * v**2 * t / s**2) * \
np.sin(k * np.pi * z / a) * \
np.exp(-0.5 * (k * np.pi * s / a)**2 * t)
return density
- Compute summary statistics of FPT distribution:
def fpt_statistics(fpt_density, dt):
"""Compute mean, variance, and quantiles of the FPT distribution."""
t_vals = np.arange(len(fpt_density)) * dt
total_mass = np.sum(fpt_density) * dt
# Normalize
fpt_normed = fpt_density / total_mass if total_mass > 0 else fpt_density
mean_fpt = np.sum(t_vals * fpt_normed) * dt
var_fpt = np.sum((t_vals - mean_fpt)**2 * fpt_normed) * dt
# Quantiles via CDF
cdf = np.cumsum(fpt_normed) * dt
quantile_10 = t_vals[np.searchsorted(cdf, 0.1)]
quantile_50 = t_vals[np.searchsorted(cdf, 0.5)]
quantile_90 = t_vals[np.searchsorted(cdf, 0.9)]
return {
"mean": mean_fpt,
"std": np.sqrt(var_fpt),
"q10": quantile_10,
"q50": quantile_50,
"q90": quantile_90,
"total_probability": total_mass
}
- For two-boundary problems, separate FPT by boundary using probability flux at each absorbing wall (finite difference of density at boundary grid points).
Got: FPT density is right-skewed unimodal distribution. For DDM with positive drift, upper boundary FPT has more mass and shorter mode than lower boundary FPT. Mean FPT for typical DDM parameters (v=1, a=1.5, z=0.75) approximately 0.5-2.0 seconds.
If fail: FPT density has negative values? Numerical differentiation noisy -- apply small Gaussian smoothing kernel. Total probability at both boundaries does not sum to approximately 1.0? Either time horizon too short (increase t_max) or probability leakage in solver.
Step 4: Analyze Parameter Sensitivity
Quantify how changes in each parameter affect first-passage time distribution.
- Define parameter grid for sensitivity analysis:
param_ranges = {
"v": np.linspace(0.2, 3.0, 15), # drift rate
"a": np.linspace(0.5, 2.5, 15), # boundary separation
"z_ratio": np.linspace(0.3, 0.7, 9) # starting point as fraction of a
}
base_params = {"v": 1.0, "a": 1.5, "z_ratio": 0.5}
- Sweep each parameter while holding others at baseline:
sensitivity_results = {}
for param_name, param_values in param_ranges.items():
means = []
accuracies = []
for val in param_values:
params = base_params.copy()
params[param_name] = val
z = params["z_ratio"] * params["a"]
process = DiffusionProcess(
drift_fn=lambda x, t, v=params["v"]: v,
diffusion_fn=lambda x, t: 1.0,
lower_bound=0.0,
upper_bound=params["a"],
boundary_type="absorbing"
)
_, survival, _ = solve_fokker_planck(process, x0=z, t_max=10.0)
fpt = first_passage_time_density(survival, dt=10.0/len(survival))
stats = fpt_statistics(fpt, dt=10.0/len(survival))
means.append(stats["mean"])
accuracies.append(stats["total_probability"]) # proxy for upper boundary
sensitivity_results[param_name] = {
"values": param_values,
"mean_fpt": np.array(means),
"accuracy": np.array(accuracies)
}
- Plot sensitivity curves:
fig, axes = plt.subplots(1, 3, figsize=(18, 5))
for idx, (param_name, result) in enumerate(sensitivity_results.items()):
ax = axes[idx]
ax.plot(result["values"], result["mean_fpt"], "b-o", label="Mean FPT")
ax.set_xlabel(param_name)
ax.set_ylabel("Mean FPT")
ax.set_title(f"Sensitivity to {param_name}")
ax2 = ax.twinx()
ax2.plot(result["values"], result["accuracy"], "r--s", label="P(upper)")
ax2.set_ylabel("P(upper boundary)")
ax.legend(loc="upper left")
ax2.legend(loc="upper right")
fig.tight_layout()
fig.savefig("parameter_sensitivity.png", dpi=150)
- Compute partial derivatives (local sensitivity at baseline):
for param_name, result in sensitivity_results.items():
idx_base = np.argmin(np.abs(result["values"] - base_params[param_name]))
if idx_base > 0 and idx_base < len(result["values"]) - 1:
d_mean = (result["mean_fpt"][idx_base+1] - result["mean_fpt"][idx_base-1]) / \
(result["values"][idx_base+1] - result["values"][idx_base-1])
print(f"d(mean_FPT)/d({param_name}) at baseline: {d_mean:.4f}")
Got: Drift rate (v) has strong negative effect on mean FPT and strong positive effect on accuracy. Boundary separation (a) has strong positive effect on mean FPT (speed-accuracy tradeoff). Starting point (z) shifts accuracy with smaller effect on mean FPT.
If fail: Sensitivity curves flat or non-monotonic? Check parameter range wide enough and solver time horizon captures full FPT distribution. Non-monotonic mean FPT with respect to drift rate would indicate solver bug.
Step 5: Validate Analytics Against Numerical Simulation
Run Monte Carlo simulations of SDE to confirm analytical and numerical PDE results.
- Implement Euler-Maruyama simulation of SDE:
def simulate_sde(process, x0, dt_sim=0.0001, t_max=10.0, n_trajectories=10000):
"""Simulate SDE paths and record first-passage times."""
n_steps = int(t_max / dt_sim)
fpt_upper = np.full(n_trajectories, np.nan)
fpt_lower = np.full(n_trajectories, np.nan)
x = np.full(n_trajectories, x0)
sqrt_dt = np.sqrt(dt_sim)
for step in range(n_steps):
t = step * dt_sim
active = np.isnan(fpt_upper) & np.isnan(fpt_lower)
if not active.any():
break
mu = np.array([process.drift(xi, t) for xi in x[active]])
sigma = np.array([process.diffusion(xi, t) for xi in x[active]])
dW = np.random.randn(active.sum()) * sqrt_dt
x[active] += mu * dt_sim + sigma * dW
# Check boundary crossings
hit_upper = active & (x >= process.upper_bound)
hit_lower = active & (x <= process.lower_bound)
fpt_upper[hit_upper] = (step + 1) * dt_sim
fpt_lower[hit_lower] = (step + 1) * dt_sim
return fpt_upper, fpt_lower
- Run simulation and compute empirical FPT distribution:
fpt_upper_sim, fpt_lower_sim = simulate_sde(ddm_process, x0=0.75, n_trajectories=50000)
# Empirical statistics
valid_upper = fpt_upper_sim[~np.isnan(fpt_upper_sim)]
valid_lower = fpt_lower_sim[~np.isnan(fpt_lower_sim)]
total_absorbed = len(valid_upper) + len(valid_lower)
accuracy_sim = len(valid_upper) / total_absorbed
print(f"Simulated accuracy: {accuracy_sim:.4f}")
print(f"Mean FPT (upper): {valid_upper.mean():.4f} +/- {valid_upper.std()/np.sqrt(len(valid_upper)):.4f}")
print(f"Mean FPT (lower): {valid_lower.mean():.4f} +/- {valid_lower.std()/np.sqrt(len(valid_lower)):.4f}")
- Compare simulation against analytical or numerical PDE solution:
fig, ax = plt.subplots(figsize=(10, 6))
# Empirical histogram
ax.hist(valid_upper, bins=100, density=True, alpha=0.5, label="Simulation (upper)")
ax.hist(valid_lower, bins=100, density=True, alpha=0.5, label="Simulation (lower)")
# Analytical solution overlay
t_vals_analytic = np.linspace(0.01, 5.0, 500)
v, a, z = 0.5, 1.5, 0.75
fpt_analytic = [ddm_fpt_upper(t, v, a, z) for t in t_vals_analytic]
ax.plot(t_vals_analytic, fpt_analytic, "k-", linewidth=2, label="Analytic (upper)")
ax.set_xlabel("First-passage time")
ax.set_ylabel("Density")
ax.set_title("FPT Distribution: Simulation vs. Analytic")
ax.legend()
fig.savefig("fpt_validation.png", dpi=150)
- Quantify agreement between methods:
from scipy.stats import ks_2samp
# Kolmogorov-Smirnov test between simulated and analytically-derived samples
analytic_cdf = np.cumsum(fpt_analytic) * (t_vals_analytic[1] - t_vals_analytic[0])
sim_sorted = np.sort(valid_upper)
sim_cdf = np.arange(1, len(sim_sorted)+1) / len(sim_sorted)
# Interpolate analytic CDF at simulation quantiles
from scipy.interpolate import interp1d
analytic_interp = interp1d(t_vals_analytic, analytic_cdf, bounds_error=False, fill_value=(0, 1))
max_diff = np.max(np.abs(sim_cdf - analytic_interp(sim_sorted)))
print(f"Max CDF difference (simulation vs. analytic): {max_diff:.4f}")
assert max_diff < 0.05, f"Simulation and analytic FPT differ by {max_diff:.4f} (threshold: 0.05)"
Got: Simulation histograms close match analytical FPT curves. KS-test maximum CDF difference below 0.05 for 50,000 trajectories. Mean FPT from simulation within 2 standard errors of analytical value.
If fail: Simulation disagrees with analytics? First check Euler-Maruyama step size -- dt_sim should be small enough that boundary crossings not missed (try dt_sim=0.00001). Analytical series does not converge? Increase n_terms. For non-constant coefficients where no analytic solution exists, compare two numerical methods (PDE solver vs. simulation) against each other.
Checks
- SDE specification passes consistency checks (finite drift, positive diffusion, x0 in domain)
- Fokker-Planck density integrates to value that decreases monotonic over time (survival function)
- Fokker-Planck solution shows no numerical artifacts (oscillations, negative values)
- FPT density non-negative and integrates to approximately 1.0 across both boundaries
- Sensitivity analysis shows expected monotonic relationships (v vs. accuracy, a vs. mean FPT)
- Monte Carlo simulation mean FPT within 2 standard errors of PDE/analytic solution
- KS-test maximum CDF difference between simulation and analytics below 0.05
Pitfalls
- Euler-Maruyama step size too large: Large dt_sim causes trajectories to overshoot boundaries. Leads to biased FPT estimates. Use dt_sim at most 1/10 of expected mean FPT, or use boundary-corrected scheme.
- Truncating FPT series too early: Analytic DDM FPT density uses infinite series. Too few terms (< 20) causes visible artifacts, especially at short times. Use at least 50 terms and check convergence.
- Ignoring numerical diffusion in PDE solver: First-order finite difference schemes introduce artificial diffusion that broadens FPT distribution. Use Crank-Nicolson or higher-order schemes for accuracy.
- Confusing Ito and Stratonovich forms: Fokker-Planck equation differs depending on SDE convention. Standard form above assumes Ito calculus. SDE written in Stratonovich form? Add noise-induced drift correction term.
- Not accounting for both boundaries: In two-boundary problems, total absorption probability must sum to 1.0. Reporting only upper boundary FPT without accounting for lower boundary gives incorrect statistics.
See Also
fit-drift-diffusion-model- applies these dynamics to estimate parameters from behavioral dataimplement-diffusion-network- generative diffusion models discretize same SDE frameworkwrite-testthat-tests- testing numerical solvers and analytical implementationscreate-technical-report- documenting diffusion analysis results
GitHub リポジトリ
関連スキル
llamaguard
その他LlamaGuardは、暴力やヘイトスピーチなど6つの安全性カテゴリーにおいて、LLMの入力と出力をモデレートするMetaの70-80億パラメータモデルです。94〜95%の精度を提供し、vLLM、Hugging Face、Amazon SageMakerを使用してデプロイ可能です。このスキルを使用して、AIアプリケーションにコンテンツフィルタリングと安全策を簡単に統合できます。
cost-optimization
その他このClaudeスキルは、リソースの適正サイジング、タグ付け戦略、支出分析を通じて、開発者がクラウドコストを最適化することを支援します。AWS、Azure、GCPにわたるクラウド支出の削減とコストガバナンスの実施のためのフレームワークを提供します。インフラコストの分析、リソースの適正サイジング、または予算制約への対応が必要な際にご利用ください。
quantizing-models-bitsandbytes
その他このスキルは、bitsandbytesを使用してLLMを8ビットまたは4ビット精度に量子化し、精度の低下を最小限に抑えつつ50〜75%のメモリ削減を実現します。限られたGPUメモリでより大規模なモデルを実行したり、推論を高速化するのに理想的で、INT8、NF4、FP4などのフォーマットをサポートしています。HuggingFace Transformersと統合され、QLoRAトレーニングや8ビットオプティマイザーを可能にします。
dispatching-parallel-agents
その他このClaudeスキルは、複数のエージェントを配備し、3つ以上の独立した問題を並行して調査・修正します。共有状態や依存関係がなく解決可能な、無関係な障害が発生するシナリオ向けに設計されています。中核となる機能は並列問題解決であり、効率を最大化するために独立した問題領域ごとに1つのエージェントを割り当てます。
