✓✓ Beats tuned baseline 2026

Confidence-Tube Neural Rollouts

Implementation & benchmark of arXiv:2607.12343 — Learning-based Homothetic Tube MPC with Non-Asymptotic Guarantees

Usefulness8/10

Difficulty6/10

Novelty6/10

Dynamics Stability Accuracy Sample efficiency World model Rl Optimizer Control theory Dynamical systems Probability Spectral theory Optimization

Source paper: Learning-based Homothetic Tube MPC with Non-Asymptotic Guarantees arXiv:2607.12343 ⓘ · analyzed Aug 30, 2026

AI-generated research hypothesis, automatically tested. Not peer-reviewed.

Idea description

Augment a learned neural state-space model with an online regularized least-squares confidence set for its local linearization or last-layer dynamics, then propagate a homothetic uncertainty tube around every predicted trajectory. Use the tube to tighten RL action constraints, reject unsafe imagined rollouts, or weight training examples by certified prediction reliability. The mechanism should improve long-horizon behavior specifically when model uncertainty is large, rather than acting as an undifferentiated regularizer.

Formulas

$$V_t=\lambda I+\sum_{s=0}^{t-1}z_sz_s^\top,\qquad \hat\theta_t=V_t^{-1}\sum_{s=0}^{t-1}z_sx_{s+1}^\top,\qquad z_s=\begin{bmatrix}x_s\\u_s\end{bmatrix}.$$

$$\Theta_t=\{\theta\mid H_t\theta\le h_t\}=\bar\Theta_t^{\mathrm p}\cap\Theta_{t-1},\qquad \bar\Theta_t^{\mathrm p}=\{\theta\mid |\theta_j-\hat\theta_{t,j}|\leq\epsilon_{t,j}\ \forall j\}.$$

$$\alpha_{i+1}\mathcal S\supseteq\operatorname{co}_{\theta\in\mathcal V(\Theta_t)}\left((A_\theta+B_\theta K)\alpha_i\mathcal S\right)\oplus\mathcal W,\qquad \alpha_{i+1}\leq\gamma_t\alpha_i+w_{\mathcal S},\quad \gamma_t=\max_{\theta\in\mathcal V(\Theta_t)}\|A_\theta+B_\theta K\|_{\mathcal S}<1.$$

$$\mathcal X_i^{\mathrm{safe}}=\mathcal X\ominus\alpha_i\mathcal S,\qquad \mathcal U_i^{\mathrm{safe}}=\mathcal U\ominus K(\alpha_i\mathcal S).$$

Mathematical statement

Assume the true local dynamics are $x_{t+1}=A_*x_t+B_*u_t+w_t$, with bounded disturbance $w_t\in\mathcal W$ and unknown parameter vector $\theta_*=[\operatorname{vec}(A_*)^\top,\operatorname{vec}(B_*)^\top]^\top$. Given regressors $z_t=[x_t^\top,u_t^\top]^\top$ and targets $x_{t+1}$, compute the regularized least-squares estimate $\hat\theta_t=V_t^{-1}\sum_{s<t}z_sx_{s+1}^\top$ with $V_t=\lambda I+\sum_{s<t}z_sz_s^\top$, where $\lambda>0$. Construct a high-probability polytope $\Theta_t=\{\theta:H_t\theta\le h_t\}$ containing $\theta_*$; coordinate-wise intervals are a practical approximation, with half-width $\epsilon_{t,j}$ determined by the non-asymptotic confidence radius and disturbance bound. For nominal feedback rollout $\bar x_{i+1}=(\hat A_t+\hat B_tK)\bar x_i+\hat B_t v_i$, represent the actual state as $x_i=\bar x_i+e_i$. If $e_i\in\alpha_i\mathcal S$ for a template set $\mathcal S$, choose $\alpha_{i+1}$ so every model vertex and disturbance satisfy $e_{i+1}\in\alpha_{i+1}\mathcal S$. If $\gamma_t=\max_{\theta\in\mathcal V(\Theta_t)}\|A_\theta+B_\theta K\|_{\mathcal S}<1$ and $w_{\mathcal S}$ bounds the disturbance, then $\alpha_{i+1}\le\gamma_t\alpha_i+w_{\mathcal S}$ and $\alpha_i\le\gamma_t^i\alpha_0+w_{\mathcal S}(1-\gamma_t^i)/(1-\gamma_t)$.

Implementation notes

First, integrate the method into a neural state-space model $f_\phi(x,u)$ used by MPC or model-based RL. Treat the final linear layer, or a local Jacobian pair $(A,B)$, as the uncertain dynamics object. Around the current nominal state, compute Jacobians with automatic differentiation and maintain RLS statistics from observed transitions. Initialize $V=\lambda I$, $b=0$, and a parameter polytope $\Theta$. For each transition, form $z=[x;u]$, update $V\leftarrow V+zz^\top$ and $b\leftarrow b+zx_+^\top$, calculate $\hat\theta=V^{-1}b$, obtain confidence widths $\epsilon_j$, and intersect the new interval box with the previous $\Theta$. During planning, enumerate box vertices $\theta^{(q)}$, calculate $\gamma=\max_q\|A^{(q)}+B^{(q)}K\|_{\mathcal S}$, propagate $\alpha_{i+1}=\gamma\alpha_i+w_{\mathcal S}$, and optimize only over $\bar x_i\in\mathcal X\ominus\alpha_i\mathcal S$ and $v_i$ such that the feedback input remains in $\mathcal U\ominus K(\alpha_i\mathcal S)$. RLS updates and vertex propagation are computed directly; disturbance bounds and confidence calibration are estimated from residual quantiles with a safety multiplier; support functions of $\mathcal S$ can be evaluated by linear programs. Start with a 2-D linear system plus a small MLP residual, then test CartPole or constrained pendulum. Compare nominal MPC and ensemble-MPC. The predicted signature is one-step residual coverage at the declared confidence level, confidence widths decreasing approximately like $1/\sqrt{\lambda_{\min}(V_t)}$, and tube radii approaching $w_{\mathcal S}/(1-\gamma)$. Sweeping feedback gain should reveal a rollout-divergence boundary near $\gamma=1$.

Verification

Beats tuned baseline

Stage 1 · Toy mechanism gate: Passed ✓

Stage 2 · Mechanism transferred to benchmark: Confirmed ✓

Stage 2 · Practical benchmark result: Significant win ✓

Methodology: 8 paired seeds, permutation test (p<0.05) against a learning-rate-tuned baseline. How verification works

Stage 1 — Mechanism check agent confidence 9/10

Built a deterministic scalar linear-system MVP implementing RLS confidence widths, interval uncertainty vertices, homothetic tube propagation, and tightened safety filtering. The core predictions manifested: confidence-width scaling had observed log-log exponents 1.000 and 1.000 versus predicted 1; the tube boundary straddled the predicted gamma=1 transition (stable up to gamma=0.9745, unstable from gamma=1.021); and exact vertex rollout matched the scalar recurrence within 8.6e-9. In the uncertainty sweep, nominal filtering accepted all 9 candidates while tube filtering rejected the highest-uncertainty case, which had 3.2% sampled unsafe successors.

Agent confidence: 9/10

Baseline: Nominal constraint check accepted 9/9 uncertainty levels; sampled unsafe rates reached 0.85% and 3.2% at uncertainty scales 0.175 and 0.2.
Idea: Tube constraint check accepted 8/9 levels, rejected scale 0.2; confidence coverage 96.375%; width exponents [1.000, 1.000]; stable tube terminal relative error median 2.3e-16; unstable geometric rate 1.0791 observed versus 1.0791 predicted.

Limitations: This is a scalar/low-dimensional numerical MVP rather than a neural state-space model, MPC optimizer, CartPole, or RL benchmark. Confidence widths use an empirically calibrated multiplier rather than a formal theorem-derived radius, and the safety test uses sampled model uncertainty rather than exhaustive multi-step closed-loop violation statistics.

How to run: python3 confidence_tube_experiment.py

Stage 2 — Benchmark vs. tuned baseline (arXiv:2607.12343) Sep 2, 2026

Beats the tuned baseline by 84.6% (p=0.0081, wins 8 / 8 paired seeds; metric: lower is better).

Benchmark: Dynamics forecasting
Model: rnn_small
Paired seeds: 8
Baseline mean: 0.0002
Idea mean: 0
Effect (Δ): -0.0002 (−84.6%; negative = idea better)
Wins: 8 / 8 paired seeds
p-value: 0.0081 (permutation test, 20 000 shuffles)
Smallest detectable effect: ±52.2%
Mechanism: Confirmed ✓
Practical effect: Helps
Baseline tuning: swept over 6 configs

Limitations:

The registered dynamics benchmark measures supervised one-step prediction rather than full MPC planning, action constraint tightening, multi-step closed-loop safety violations, or exhaustive uncertainty-polytope vertex propagation. The implementation uses a diagonal/mean RLS confidence approximation and an empirically conservative disturbance proxy.

How to run: python3 stage2_confidence_tube_bench.py

Verdict computed by deterministic test code from paired-seed statistics — not by the language model.

Stage 2 — Benchmark Aug 30, 2026 latest Worked ✓

Agent confidence: 8/10

Confidence-Tube Neural Rollouts produced a significant win on the structurally matched registered dynamics track. Using the same rnn_small architecture and paired seeds, the idea reduced mean test MSE from 2.2203499e-4 to 3.4294854e-5, winning on all 8 seeds with permutation p=0.0081. The trained-model signature confirmed decreasing RLS widths and conservative empirical coverage, though this does not establish full closed-loop MPC safety.

Baseline: Tuned rnn_small dynamics baseline: mean test MSE 0.00022203498974704416.
Idea: RLS confidence-tube reliability-weighted rnn_small: mean test MSE 0.0000342948537763732; paired delta -0.00018774013597067096; p=0.0081; 8/8 wins.

Benchmark result