# Degree-Calibrated Stable Residual Flow

- ID: 3127
- Canonical URL: https://synthcore.org/idea/3127/degree-calibrated-stable-residual-flow
- API JSON: https://synthcore.org/api/idea/3127.json
- API Markdown: https://synthcore.org/api/idea/3127.md
- Verification status: unverified
- Source: [arXiv:2609.03120](https://arxiv.org/abs/2609.03120)
- Category: architecture
- Solves: stability, accuracy
- ML areas: ssm, rnn, mlp, training-dynamics
- Math tags: dynamical-systems, control-theory, linear-algebra
- Ratings: usefulness 7/10; difficulty 6/10; novelty 8/10

## Idea description

Construct a continuous-depth or recurrent residual block with a prescribed polynomial Lyapunov decay near its equilibrium. The architecture combines a fixed radial stabilizer with a learned component that is constrained to have zero radial projection, allowing slow algebraic transients and long memory while preventing asymptotic hidden-state growth.

## Mathematical statement

Let z(t) be a hidden state, z_star its equilibrium, r = ||z - z_star||^2 divided by 2, a greater than zero a stabilizing gain, and m greater than or equal to zero a chosen degree. Use the vector field dz/dt = -a r^m (z-z_star) + P_perp(z-z_star) h_phi(z,t), where P_perp(u) = I - uu^T/(||u||^2 + epsilon) approximately removes the radial component of the learned field. For the idealized field without projection error, dr/dt = -a r^(m+1). Thus r decays exponentially when m equals zero and as t^(-1/m) when m is positive; the state norm decays as t^(-1/(2m)). A sampled radial penalty can enforce the same inequality for the learned discretized vector field.

## Key formulas

- $$dz/dt = -a r^m (z-z_star) + P_perp(z-z_star) h_phi(z,t),   r = ||z-z_star||^2/2$$
- $$P_perp(u) = I - uu^T/(||u||^2 + epsilon)$$
- $$dr/dt = (z-z_star)^T dz/dt <= -a r^(m+1)$$
- $$L_rad = mean(max(0, (z-z_star)^T f_phi(z) + a r^(m+1))^2)$$

## Implementation notes

Use this in a neural ODE, continuous-depth ResNet, or recurrent state-space block. Set z_star to zero initially and parameterize the hidden dynamics as f_phi(z,t,x) = -a r^m z + P_perp(z) h_phi(z,t,x). For an Euler block, update z_next = z + delta_t f_phi(z,t,x); for longer trajectories use an adaptive ODE solver. At sampled hidden states, compute r and the radial derivative z transpose f_phi(z), then add L_rad to the task loss. A hard alternative is to rescale any proposed update whose empirical radial derivative exceeds the allowed value. Compute the radial quantities exactly on minibatch states; estimate global validity by sampling states, inputs, and small perturbations because the penalty is not a global certificate. The paper supplies the degree-to-decay comparison; the tangent learned field and sampled penalty are the proposed neural adaptation. First test m equal to 0, 1, and 2 against an unconstrained residual MLP and a spectrally normalized residual block on a synthetic long-horizon sequence task and sequential MNIST. After the input is removed, measure hidden-state norms for 100 integration steps and plot r on log-log axes. The prediction is exponential decay for m equal to zero and a log-log slope near -1/m for r when m is positive, with no asymptotic growth provided the radial inequality holds. Increase delta_t to identify the numerical instability boundary and report whether the observed decay law persists below that boundary.

## Disclaimer

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