Persistent Homology

Beyond Statistics

Statistics describes the distribution of data: mean, variance, correlations. But distributions are blind to shape.

Consider two point clouds:

Cloud A:          Cloud B:
  ● ●               ●   ●
 ●   ●             ●     ●
●     ●            ●     ●
 ●   ●             ●     ●
  ● ●               ●   ●

Same mean. Same variance. Same point count. But Cloud A is a filled disk; Cloud B is a ring with a hole. The hole is topologically significant—it represents something absent, something that might matter.

Persistent homology is the mathematics of detecting such shapes.

The Vietoris-Rips Complex

Given a point cloud, we construct a simplicial complex by connecting points within a distance threshold ε:

ε = small:     ε = medium:     ε = large:
  ●   ●         ●───●           ●───●
                    │           │╲ ╱│
  ●   ●         ●   ●           ●─╳─●
                                │╱ ╲│
  ●   ●         ●───●           ●───●

As ε increases:

• H0 features (connected components): Merge as clusters connect
• H1 features (loops): Appear when edges close cycles, disappear when interiors fill
• H2 features (voids): Appear when surfaces enclose volumes

Birth and Death

Each topological feature has a birth time (the ε at which it appears) and a death time (the ε at which it vanishes).

Features that persist across a wide range of ε are considered significant—they reflect genuine structure rather than noise.

Persistence Diagram:
        death
          │
          │    ● (noise: short-lived)
          │
          │          ● (signal: long-lived)
          │        ●
          │      ●
          └──────────── birth

Points far from the diagonal represent persistent features.

Death Loops (H1)

In Gaius, H1 features receive special attention as “death loops.” These represent:

• Cycles in data flow: Feedback loops, circular dependencies
• Systemic risks: Self-reinforcing failure modes
• Market structures: Liquidity cycles, regulatory arbitrage loops

When projected onto the grid, death loops appear as ⚠ markers in regions where the underlying embedding space exhibits persistent 1-dimensional holes.

Practical Application

from gtda.homology import VietorisRipsPersistence
from gtda.diagrams import PersistenceEntropy

# Compute persistence diagrams
vr = VietorisRipsPersistence(homology_dimensions=[0, 1, 2])
diagrams = vr.fit_transform([point_cloud])

# Quantify topological complexity
entropy = PersistenceEntropy()
ent = entropy.fit_transform(diagrams)

# Extract significant H1 features
h1_features = diagrams[0][diagrams[0][:, 0] == 1]
persistent_loops = h1_features[h1_features[:, 2] - h1_features[:, 1] > threshold]

Entropy as Summary

Persistence entropy provides a scalar summary of topological complexity:

• Low entropy: Few dominant features (simple structure)
• High entropy: Many features of similar persistence (complex, fractal-like)

Gaius tracks entropy over time. Sudden entropy spikes may indicate regime changes in your domain.

Interpreting Grid Overlays

When viewing the H1 overlay:

Limitations

Persistent homology reveals shape but not causation. A detected loop could represent:

• A real feedback cycle in your domain
• An artifact of the embedding model
• Noise in the underlying data

Domain expertise is required to interpret topological features. Gaius surfaces the structure; you provide the meaning.

Further Reading

• Computational Topology by Edelsbrunner and Harer
• Topological Data Analysis by Carlsson
• giotto-tda documentation: giotto-ai.github.io