△ Mechanism confirmed, baseline not beaten 2026

Differentiable Widest-Path Loss

Implementation & benchmark of arXiv:2607.07123 — Widest-Path Reachability Fields for Connectivity-Preserving Slender Structure Segmentation

Usefulness8/10

Difficulty5/10

Novelty6/10

Regularization Accuracy Stability Generalization Cnn Loss Regularization Graph nn Graph theory Optimization Geometry

Source paper: Widest-Path Reachability Fields for Connectivity-Preserving Slender Structure Segmentation arXiv:2607.07123 ⓘ · analyzed Aug 30, 2026

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 Not peer-reviewed.

Idea description

Add a reachability loss that forces annotated structural endpoints or landmarks to remain connected through high-confidence pixels. Instead of averaging error over all pixels, the loss penalizes the weakest pixel on the best available path, directly targeting bottlenecks responsible for broken vessels, roads, and cracks.

Formulas

$(\Pi_sX)(y,x)=\max_{0\leq i,j<s}X\big(s(y-1)+i+1,\,s(x-1)+j+1\big)$

$$R(a,t)=\max_{\pi:a\leadsto t}\;\min_{u\in\pi}p_u,\qquad c(\pi)=\min_{u\in\pi}p_u.$$

$$\operatorname{smin}_\tau(z_1,\ldots,z_n)=-\tau\log\sum_{i=1}^{n}e^{-z_i},\qquad \operatorname{smax}_\tau(z_1,\ldots,z_n)=\tau\log\sum_{i=1}^{n}e^{z_i}.$$

$$R_{k+1}(v)=\operatorname{smax}_\tau\!\left\{\operatorname{smin}_\tau\big(R_k(u),p_v\big):u\in N(v)\right\},\qquad L_{\mathrm{reach}}=\sum_{(a,t)\in\mathcal P}w_{at}\big(1-R_K(t)\big).$$

Mathematical statement

The paper defines cell-wise max pooling from the pixel grid to a graph grid: for a scalar field X on the original grid and stride s, $(\Pi_sX)(y,x)=\max_{0\le i,j<s}X(s(y-1)+i+1,s(x-1)+j+1)$. Use this to obtain graph-node probabilities p_v in [0,1] from the network foreground probability map. For a source node a, a path $\pi=(v_0=a,v_1,\ldots,v_m=t)$ has capacity $c(\pi)=\min_{u\in\pi}p_u$, and widest-path reachability is $R(a,t)=\max_{\pi:a\leadsto t}c(\pi)$. The outer maximum chooses the strongest route while the inner minimum is controlled by its bottleneck pixel. For differentiable training, replace min by $\operatorname{smin}_\tau(z_1,\ldots,z_n)=-\tau\log\sum_i\exp(-z_i/\tau)$ and max by $\operatorname{smax}_\tau(z_1,\ldots,z_n)=\tau\log\sum_i\exp(z_i/\tau)$. A finite-step Bellman relaxation is $R_0(v)=1$ for source nodes and 0 otherwise, $R_{k+1}(v)=\operatorname{smax}_\tau\{\operatorname{smin}_\tau(R_k(u),p_v):u\in N(v)\}$, followed by $L_{\mathrm{reach}}=\sum_{(a,t)\in\mathcal P}w_{at}(1-R_K(t))$. Here N(v) is the 4- or 8-neighbor set, $\mathcal P$ is a set of source-target pairs sampled from the ground-truth skeleton, w_at is an optional inverse-frequency pair weight, K is the number of propagation iterations, and $\tau$ controls approximation smoothness. The extracted paper formula supplies the occupancy-preserving projection; the Max-Min path expression and soft Bellman recursion provide the implementable training formalization of the stated WPRF objective.

Implementation notes

Integrate the module after the network foreground sigmoid and before the total segmentation loss. First construct a graph on a downsampled probability map: if the output has stride s, apply the paper's cell-wise max pooling to the foreground probability p, yielding one node value p_v per graph cell. Extract a one-pixel-wide skeleton from the ground-truth foreground mask with deterministic thinning, identify endpoints and junction-adjacent landmarks, and sample source-target pairs along the same ground-truth connected component. Reject pairs whose Euclidean distance is below a minimum length so the module tests meaningful connectivity. Use 4-neighbor connectivity for the baseline and optionally 8-neighbor connectivity as an ablation. For each source a, initialize R[0,a]=1 and all other nodes to zero. For k=0,...,K-1, compute each neighbor candidate q(u,v)=smin_tau(R[k,u],p[v]), aggregate candidates with smax_tau, and retain R[k+1,v]. Gather R[K,t] for target nodes and add lambda times L_reach to BCE or Dice loss. Use log-sum-exp for numerical stability, tau=0.05 to 0.2, K equal to the maximum graph distance of sampled pairs or a clipped value such as 32, and detach neither p nor R so gradients reach the logits. Estimate no mathematical quantities: s, K, tau, lambda, and pair weights are hyperparameters; the skeleton and pair set are recomputed per training sample. Start with a small U-Net on DRIVE retinal vessels or CrackForest, comparing BCE+Dice against BCE+Dice+reachability. Measure clDice, number of predicted connected components, endpoint-to-endpoint success rate, pixel IoU, and wall-clock training overhead. Success means fewer broken components and at least a 2-point clDice gain without materially worse pixel IoU or more than 20% training-time overhead.

Verification

Mechanism confirmed, baseline not beaten

Stage 1 · Toy mechanism gate: Passed ✓

Stage 2 · Mechanism transferred to benchmark: Confirmed ✓

Stage 2 · Practical benchmark result: No measurable effect

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 a differentiable widest-path MVP with cell-wise max pooling, softmin/softmax operators, finite-step Bellman propagation, and a toy optimization comparison. The mechanism manifested: low-temperature reachability converged to the hard bottleneck (0.350), first meaningful reach appeared at K=5 for graph distance 5, and bottleneck gradients were approximately 1 below the competing capacity and 0 above it. However, the literal recurrence is non-monotonic in K because it permits exact-length backtracking, and the toy optimization gain was negligible relative to its large runtime overhead, so this is not evidence of a practical training win.

Agent confidence: 8/10

Baseline: BCE optimization: final BCE 0.01883, IoU 1.00, soft reach 0.81587, 0.27 s
Idea: BCE + 0.35 reach loss: final BCE 0.01683, IoU 1.00, soft reach 0.81782, 16.20 s; tau sweep reached hard widest-path value 0.350 at tau <= 0.02; K=5 first exceeded 0.1 reach (0.886); bottleneck gradients were [1.0, 1.0, 0.50, 0.0, 0.0] for bottleneck values [0.20, 0.40, 0.60, 0.79, 0.81]

Limitations: Only a synthetic 1-D/2-D toy task was tested; no DRIVE, CrackForest, skeleton extraction, U-Net integration, clDice, connected-component, endpoint-success, or realistic FLOP comparison was performed. The implementation uses explicit Python neighbor loops and is an MVP rather than a production-efficient loss.

How to run: python3 widest_path_experiment.py

Stage 2 — Benchmark vs. tuned baseline (arXiv:2607.07123) Sep 2, 2026

No significant difference vs. the tuned baseline (−-0.0%, p=1, wins 0 / 8 paired seeds).

Benchmark: Custom_tracks/connected_paths
Model: tiny_segmentation
Paired seeds: 8
Baseline mean: 0.0433
Idea mean: 0.0433
Effect (Δ): 0 (−-0.0%; negative = idea better)
Wins: 0 / 8 paired seeds
p-value: 1 (permutation test, 20 000 shuffles)
Smallest detectable effect: n/a
Mechanism: Confirmed ✓
Practical effect: No effect
Baseline tuning: swept over 3 configs

Limitations:

This is a synthetic custom structural-connectivity track rather than DRIVE or CrackForest. The implementation uses a tiny CNN, short 8-epoch training, simplified deterministic endpoint selection rather than skeletonization, and reports MSE rather than clDice, connected components, or endpoint-success metrics. The run fell back to CPU; training overhead was not included in the final comparison.

How to run: python3 bench_experiment.py

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

Stage 2 — Benchmark Aug 30, 2026 latest Failed ✗

Agent confidence: 9/10

Implemented a structurally matched custom connected-path image track and a shared tiny segmentation CNN, with differentiable widest-path loss as the only intervention. The neural-scale mechanism check confirmed that reducing a path bottleneck lowered reach by 0.538 versus a predicted 0.6. However, the eight-seed comparison was exactly tied with paired delta 0 and permutation p=1.0, so there is no observed training win.

Baseline: Best sweep lr=0.003, weight_decay=0.0001; MSE 0.04326914297416806 ± 0.010400941863666063.
Idea: Best of 3 settings; MSE 0.04326914297416806 ± 0.010400941863666063; paired delta 0.0, p=1.0; endpoint reach mean 0.0.
Mechanism signature: Reach changed from 0.7489715814590454 to 0.211371049284935 after lowering a bottleneck, observed drop 0.5376005172729492 versus predicted 0.6; confirmed=true.

Benchmark result