✓✓ Beats tuned baseline 2026

Projected Absolute-Residual Compensation for Neural State-Space Models

Implementation & benchmark of arXiv:2607.28142 — Data-Driven Dead-Zone Compensation via Projection in Predictive Control Setting

Usefulness7/10

Difficulty5/10

Novelty8/10

Dynamics Stability Accuracy Rnn Ssm World model Control theory Dynamical systems Linear algebra Optimization

Source paper: Data-Driven Dead-Zone Compensation via Projection in Predictive Control Setting arXiv:2607.28142 ⓘ · analyzed Aug 31, 2026

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 Not peer-reviewed.

Idea description

Augment a recurrent or state-space neural model with two predictors: an absolute predictor using raw command and output histories, and an incremental predictor using differences. Use the absolute prediction residual, projected onto an offline-learned mismatch subspace, to estimate persistent actuator bias or dead-zone effects and compensate the next command or latent transition. The incremental branch provides a diagnostic because a constant mismatch should vanish there while the absolute residual remains nonzero and proportional to the mismatch.

Formulas

$$\hat{y}_k^{\mathrm{abs}}=P_u U_k+P_y Y_k+P_c c_k,\qquad e_k^{\mathrm{abs}}=y_k-\hat{y}_k^{\mathrm{abs}}.$$

$$\Delta\hat{y}_k^{\mathrm{inc}}=Q_u\Delta U_k+Q_y\Delta Y_k+Q_c\Delta c_k,\qquad e_k^{\mathrm{inc}}=\Delta y_k-\Delta\hat{y}_k^{\mathrm{inc}}.$$

$$\Pi_d=H_d(H_d^{\mathsf T}H_d+\lambda I)^{-1}H_d^{\mathsf T},\qquad \hat e_{d,k}=\Pi_d e_k^{\mathrm{abs}},\qquad \delta u_k=-\gamma G_d^\dagger \hat e_{d,k}.$$

$$\|z_k\|\le c\rho(A_{\mathrm{cl}})^k\|z_0\|+\frac{c\|B_e\|}{1-\rho(A_{\mathrm{cl}})}\sup_j\|(I-\Pi_d)P_d d_j\|,\qquad \rho(A_{\mathrm{cl}})<1.$$

Mathematical statement

Let u_k be the commanded input, u_k^eff the unknown effective input after a dead-zone or quantizer, and y_k the observed output. Fit an absolute one-step predictor yhat_k^abs = P_u U_k + P_y Y_k + P_c c_k, where U_k and Y_k are stacked input and output histories and c_k contains current neural features. Fit an incremental predictor Delta-yhat_k^inc = Q_u Delta-U_k + Q_y Delta-Y_k + Q_c Delta-c_k. For a persistent mismatch d_k = u_k^eff - u_k, the absolute residual e_k^abs = y_k - yhat_k^abs is approximately P_d d_k, while the incremental residual tends to zero for constant d_k. Let H_d contain calibration residual directions and define the regularized orthogonal projector Pi_d = H_d (H_d^T H_d + lambda I)^(-1) H_d^T. The estimated mismatch signature is ehat_d,k = Pi_d e_k^abs. If G_d maps a command correction to output residual, apply delta-u_k = -gamma G_d^dagger ehat_d,k, with 0 < gamma <= 1. For a locally linearized compensated closed loop z_{k+1} = A_cl z_k + B_e(e_k^abs - ehat_d,k), practical stability requires spectral radius rho(A_cl) < 1, and the residual bound scales with the unprojected component (I - Pi_d) P_d d_k.

Implementation notes

1. Integration point: wrap a GRU, S4-style state-space layer, or neural state-space controller that emits a nominal command u_k for a simulated actuator or environment. Maintain buffers of the last T commands, observations, and neural features. Add two small linear heads on these features: an absolute head predicting y_k and an incremental head predicting Delta-y_k. Apply the projected residual correction before the command reaches the actuator or simulator. 2. Pseudocode: collect nominal trajectories; fit P_u, P_y, and P_c by ridge regression; fit Q_u, Q_y, and Q_c on differenced data; collect residual vectors while injecting several constant biases and dead-zone widths; form H_d from the leading principal components of those residuals; compute Pi_d; at every step calculate e_abs, calculate ehat_d = Pi_d e_abs, estimate delta-u = -gamma times the pseudoinverse of G_d times ehat_d, clip the corrected command, and execute it. 3. The mechanism taken from the paper is the absolute-versus-incremental decomposition and fixed projection. Predictor matrices, H_d, G_d, and local spectral radius are estimated empirically. 4. First experiment: train a small GRU to control a damped linear plant with a known plus-or-minus 0.18 command dead-zone. Compare the baseline, projected correction, and unprojected correction on constant-reference tracking. The predicted signature is that incremental residual norm decays toward zero under constant mismatch while absolute residual norm remains proportional to bias. After correction, steady-state error should scale with the norm of (I - Pi_d) P_d d. The measured divergence boundary should occur near rho(A_cl) = 1, within 20 percent.

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 8/10

Built and numerically verified a deterministic projected absolute-residual compensation toy in verify.py. Absolute residual scaling, incremental cancellation, projection residual scaling, and the stability boundary all matched their predictions; tracking error improved from 0.272727 to 0.054545 with compensation. Projected and unprojected compensation were identical here because the toy was one-dimensional and fully covered by the learned mismatch subspace.

Agent confidence: 8/10

Baseline: Mean absolute steady-state tracking error 0.272727; final error 0.272727
Idea: Mean absolute steady-state tracking error 0.054545; final error 0.054545; absolute residual slope 1.2500000000000002 vs predicted 1.25; incremental residual max 0; projection max error 1.0e-10; stability boundary observed 2.0 vs predicted 2.0

Limitations: This is an exact linear toy, not a trained GRU or S4 model. It does not test learned ridge heads, finite-sample estimation, measurement noise, nonlinear dead-zone or quantization effects, online adaptation, multidimensional G_d pseudoinverse behavior, or a meaningful projected-versus-unprojected controller distinction.

How to run: python3 verify.py

Stage 2 — Benchmark vs. tuned baseline (arXiv:2607.28142) Sep 2, 2026

Beats the tuned baseline by -0.0015 metric points (p=0.0081, wins 8 / 8 paired seeds; metric: lower is better).

Benchmark: Dynamics forecasting
Model: rnn_small
Paired seeds: 8
Baseline mean: 0
Idea mean: 0
Effect (Δ): -0.0015 (n/a; negative = idea better)
Wins: 8 / 8 paired seeds
p-value: 0.0081 (permutation test, 20 000 shuffles)
Smallest detectable effect: n/a
Mechanism: Confirmed ✓
Practical effect: Helps
Baseline tuning: swept over 0 configs

Limitations:

Only the built-in actuated-pendulum dynamics track was tested. It does not explicitly inject persistent actuator bias, dead-zone, quantization, measurement noise, or test the stability boundary under controlled mismatch sweeps.

How to run: python3 bench_stage2.py

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

Stage 2 — Benchmark Aug 31, 2026 latest Worked ✓

Agent confidence: 7/10

Implemented projected absolute-residual compensation on the structurally matched actuated-pendulum dynamics track using paired eight-seed training and the benchmark sweep/report protocol. The idea achieved lower test MSE on all 8 seeds: 4.956e-05 versus 1.530e-03 for the baseline, with paired delta -1.481e-03 and permutation p=0.0081. The incremental residual was substantially smaller than the absolute residual in trained-model behavior, although the built-in track does not explicitly inject persistent actuator bias.

Baseline: mean test MSE 0.0015303929612855427; std 0.0006243748661918005; best lr 0.01; 8 paired seeds
Idea: mean test MSE 0.00004955958911523339; std 0.000022175200246237042; best lr 0.01, gamma 0.8; 8 paired seeds
Idea wins: 8/8
Mechanism signature: absolute residual norm 0.7707589045166969; incremental residual norm 0.0819682078436017; disagreement norm 0.8208163678646088; confirmed=true
Paired delta: -0.0014808333721703093
Permutation p value: 0.0081

Benchmark result