---
title: 'TRA-s: Topological Residual Asymmetry with Smoothing'
url: https://www.emergentmind.com/topics/tra-s
type: topic
---

# TRA-s: Topological Residual Asymmetry with Smoothing

Topological Residual Asymmetry with Smoothing (TRA-s) defines a robust criterion for inferring bivariate causal direction in additive-noise models (ANMs) under fixed noise, leveraging geometric signatures obtained from the persistent homology of regressor-residual clouds. It addresses limitations of the original TRA method, which loses discriminatory power when the noise level is fixed and does not vanish. TRA-s employs a binning strategy that induces a one-dimensional geometric signal in the reverse residual direction, restoring identifiability. It is theoretically consistent, computationally practical, and empirically validated across diverse synthetic and real-world benchmarks [2602.00427].

## 1. Motivation and Formal Definition

In the classical Topological Residual Asymmetry (TRA) method, causal direction is inferred by comparing the geometric structure of two residual clouds after copula standardization: the forward direction $X \rightarrow Y$ produces residuals that are approximately independent, yielding a 2D bulk, while the reverse direction $Y \rightarrow X$ under small noise collapses to a one-dimensional tube. The separation is quantified by a 0D persistent-homology functional computed from the Euclidean minimum spanning tree (MST) on the residual cloud, with the raw TRA score $\Delta_n$ providing directionality.

However, under fixed additive noise, the reverse residuals do not collapse; their distribution remains thick, and the original TRA statistic loses its discrimination. TRA-s overcomes this by binning reverse residuals along the $Y$ copula axis and averaging, converting the "thick" cloud into a binned average that again collapses to a 1D signature at mesoscopic scale, while forward residuals retain a 2D structure.

Formally, let $\{(X_i, Y_i)\}_{i=1}^n$ be sampled from an additive-noise model $Y_i = f(X_i) + \varepsilon_i$ with $\varepsilon_i \indep X_i$, $\operatorname{Var}(\varepsilon) = \sigma^2 > 0$. TRA-s proceeds by cross-fitting regressors in both directions, computing residuals, copula-standardizing, binning the reverse direction, and calculating the MST-based persistence profiles over fixed mesoscopic windows. The TRA-s score $\widetilde{\Delta}_n$ is then the difference in windowed TP-profiles between the forward and binned-reverse clouds. If $\widetilde{\Delta}_n > 0$, the method infers $X \to Y$; if $\widetilde{\Delta}_n < 0$, $Y \to X$ [2602.00427].

## 2. Binning and Copula Standardization Procedure

TRA-s modifies the reverse residual cloud via a binning strategy following rank-copula standardization:

- For the reverse direction ($X | Y$), only the $Y$ coordinate is rank-transformed: $U_i = \frac{\mathrm{rank}(Y_i)}{n+1}$.
- $[0,1]$ is partitioned into $B_n$ equal-width bins: $I_{n, b} = (\frac{b-1}{B_n}, \frac{b}{B_n}]$, $b = 1, \dotsc, B_n$.
- For each bin, empirical bin centers $\bar{u}_b$ and mean residuals $\bar{r}_b$ are computed over the points within the bin.
- The binned cloud is $\widehat{\mathcal R}^{(n)}_{X\mid Y} = \{(\bar{u}_b, \bar{r}_b): b = 1, \dots, B_n\} \subset [0,1] \times \mathbb{R}$.

The bin number $B_n$ is chosen such that $B_n \to \infty$, $B_n = o(n)$, and $B_n^{7/3} \log B_n / n \to 0$, ensuring asymptotic validity without over-smoothing or excessive bias [2602.00427].

## 3. Persistent Homology Functional and TRA-s Statistic

The core geometric statistic is a normalized windowed profile of edge-lengths in the MST of the residual cloud, specialized to a "soft window" $[\alpha, \beta]$ that captures mesoscopic geometry:

\[
\overline{\TP_0^{[\alpha, \beta]}(\mathcal{R})} = \frac{1}{(M-1)(\beta-\alpha)} \sum_{e \in \operatorname{MST}(\mathcal{R})} \Psi_{\alpha, \beta}(\|e\|),
\]
with
\[
\Psi_{\alpha, \beta}(t) = (\min\{t, \beta\} - \alpha)_+ = \begin{cases} 0, & t \le \alpha, \\ t - \alpha, & \alpha < t < \beta, \\ \beta - \alpha, & t \ge \beta. \end{cases}
\]

- For the forward ($Y \mid X$) copula cloud, set $M = n$, $\alpha_n = \kappa n^{-2/3}$, and $\beta_n = c_\beta \alpha_n$.
- For the binned-reverse cloud ($X \mid Y$), $M = B_n$, $\widetilde\alpha_n = \kappa B_n^{-2/3}$, $\widetilde\beta_n = c_\beta \widetilde\alpha_n$.

The TRA-s test statistic is then:
\[
\widetilde{\Delta}_n = \overline{\TP_0^{[\alpha_n, \beta_n]}(\widetilde{\mathcal R}^{(n)}_{Y\mid X})} - \overline{\TP_0^{[\widetilde\alpha_n, \widetilde\beta_n]}(\widehat{\mathcal R}^{(n)}_{X\mid Y})}.
\]
The direction is assigned according to the sign of $\widetilde{\Delta}_n$ [2602.00427].

## 4. Theoretical Guarantees and Assumptions

Under the following conditions (cf. [2602.00427], Assumptions 3.5–3.7):

- Forward model: $f \in C^1$, branchwise invertible, $|f'| \in [c_f, C_f]$, noise sub-Gaussian.
- Reverse conditional mean $m(y) = \mathbb{E}[X \mid Y = y]$ continuous.
- Regression errors: cross-fitted regression risk $o(1)$ and $\max_i |\widehat{g}(Y_i) - m(Y_i)| = o(\alpha_n)$.
- Binning: $B_n \to \infty$, $B_n = o(n)$, $B_n^{7/3} \log B_n / n \to 0$.
- Reverse fluctuations: $\|X - m(Y)\|_{\psi_2 | Y = y} \le K_0 \sigma$.

Theorem 3.2 ([2602.00427]) establishes that, as $n \to \infty$,
\[
\overline{\TP_0^{[\alpha_n, \beta_n]}(\widetilde R_{Y|X})} \to 1, \qquad
\overline{\TP_0^{[\widetilde\alpha_n, \widetilde\beta_n]}(\widehat R_{X|Y})} \to 0,\qquad
\widetilde{\Delta}_n \to 1,
\]
in probability. Therefore, a rule that declares $X \to Y$ when $\widetilde\Delta_n > \tau_n \downarrow 0$ is consistent with vanishing abstention probability [2602.00427].

## 5. Algorithmic Procedure and Computational Considerations

The TRA-s method is implemented as follows:

1. Cross-fit forward and reverse regressors using $K$-fold splits, producing residuals $r_i^{(Y|X)}$ and $r_i^{(X|Y)}$.
2. Copula-standardize the forward cloud using both $(X, r^{(Y|X)})$ coordinates.
3. Bin the reverse residuals along $Y$ copula with $B_n$ bins, compute bin mean coordinates $(\bar{u}_b, \bar{r}_b)$.
4. Compute MSTs and mesoscopic persistence profiles in both directions.
5. Form the TRA-s statistic and make a directional decision.

Computational complexity per $n$ samples is $O(n \log n)$ for ranking and MST computation (using Delaunay-based MST algorithms), $O(B_n \log B_n)$ for the binned reverse cloud, and regression complexity depends on the choice of estimator (e.g., smoothing splines).

Parameter choices for $B_n$ (recommended $n^{0.4}$) and mesoscopic scaling $(\kappa, c_\beta)$ are robust; $B_n$ must neither be too small (under-smoothing) nor too large (over-smoothing), ensuring resolution of the 1D collapse in the reverse cloud.

## 6. Empirical Validation and Performance

Extensive experiments illustrate the efficacy and stability of TRA-s:

- On synthetic ANMs with various forms (cubic, near-linear, heteroscedastic, non-monotone), TRA-s achieves low directed risk and matches theoretical predictions, while most established baselines (RESIT, IGCI, RECI, CDCI, COMIC, RCC, NCC) exhibit regular failure modes under stress [2602.00427].
- Under confounding, TRA-s combined with the abstention/correction procedure (TRA-C) abstains appropriately, whereas alternatives tend to commit to a potentially incorrect direction.
- On the Tübingen real-world benchmark, TRA-s delivers optimal coverage and second-lowest risk, with the abstaining TRA-C attaining highest decided accuracy and minimal overall risk.
- Ablation studies confirm that smoothing of the reverse direction is necessary for robust separation under fixed noise, and performance is stable to tuning in the recommended $B_n$ and scaling parameter ranges.

## 7. Practical Usage and Limitations

TRA-s is applicable to bivariate causal inference under additive-noise settings, robust to moderate regression error, and computationally tractable for moderate to large $n$ given efficient MST computation. Binning is central: it restores the geometric signature required for correct causal inference when raw residual dispersion precludes a 1D collapse. The approach is grounded in persistent-homology theory, but performance may degrade for non-ANM or heavily confounded distributions (necessitating abstention via TRA-C).

A plausible implication is that TRA-s, by explicit geometric regularization of the residual cloud, defines a broadly applicable, nonparametric, and theoretically rigorous template for distribution-based causal inference that is robust to non-vanishing noise scales, provided the core bivariate ANM structure holds [2602.00427].

Source: https://www.emergentmind.com/topics/tra-s