SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.05367 · 2021

Pulling back information geometry

arXiv · PDF · Open in the Atlas

Code that ran

2 of 3 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 0 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2106.05367.json · badge: /api/ran/2106.05367.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
machinelearninglifescience/stochman canonical (classifier label)
classifier basis: classifier_2025
2 ran · 1 no recorded run
FunctionStatusWhere it lives
geodesic_minimizing_energy Ran
ran, violated the contract we drafted
machinelearninglifescience/stochman/stochman/geodesic.py
code served (permissive licence) · get_code("380d519d9cb64725")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class CubicSpline(nn.Module):
    """A simple curve representation with learnable control nodes."""
    def __init__(self, begin, end, num_nodes=8, requires_grad=True):
        super().__init__()
        self.begin = begin
        self.end = end
        self.num_nodes = num_nodes
        self.device = begin.device
        
        # Create learnable interior nodes
        interior = torch.linspace(0, 1, num_nodes).unsqueeze(-1) * (end - begin) + begin
        self.nodes = nn.Parameter(interior, requires_grad=requires_grad)
    
    def forward(self, alpha):
        """Evaluate curve at parameter values alpha."""
        # Simple linear interpolation for demonstration
        # alpha shape: (eval_grid,)
        # Output shape: (eval_grid, dim)
        batch_size = alpha.shape[0]
        dim = self.begin.shape[-1]
        
        # Concatenate begin, nodes, end
        all_nodes = torch.cat([self.begin, self.nodes, self.end], dim=0)
        
        # Simple piecewise linear interpolation
        result = []
        for a in alpha:
            idx = a * (all_nodes.shape[0] - 1)
            lower_idx = int(torch.floor(idx))
            upper_idx = min(lower_idx + 1, all_nodes.shape[0] - 1)
            t = idx - lower_idx
            
            point = (1 - t) * all_nodes[lower_idx] + t * all_nodes[upper_idx]
            result.append(point)
        
        return torch.stack(result, dim=0)

class Sphere(nn.Module):
    """A simple sphere manifold."""
    def __init__(self):
        super().__init__()
    
    def curve_energy(self, curve_points):
        """
        Compute energy of curve points on the sphere.
        curve_points shape: (eval_grid, dim)
        Returns: (eval_grid,) tensor of energies
        """
        # Simple energy: sum of squared differences between consecutive points
        # This encourages geodesic-like behavior
        diffs = torch.diff(curve_points, dim=0)
        energy = torch.sum(diffs ** 2, dim=1)
        
        # Pad to match input size
        energy = torch.cat([energy, energy[-1:]], dim=0)
        return energy

# Create test objects
p0 = torch.tensor([[0.1, 0.1, 0.2, 0.3]], dtype=torch.float32)
p1 = torch.tensor([[0.3, 0.7, 0.5, 0.4]], dtype=torch.float32)

curve = CubicSpline(begin=p0, end=p1, num_nodes=8, requires_grad=True)
manifold = Sphere()

ARGS = {
    "curve": curve,
    "manifold": manifold,
    "optimizer": torch.optim.Adam,
    "max_iter": 150,
    "eval_grid": 20,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
geodesic_minimizing_energy_sgd Ran
ran, violated the contract we drafted
machinelearninglifescience/stochman/stochman/geodesic.py
code served (permissive licence) · get_code("af7b58d25a7f044b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class SimpleCurve(nn.Module):
    """A simple curve object with learnable parameters and a device attribute."""
    def __init__(self):
        super().__init__()
        # Create learnable parameters representing control points
        self.parameters_tensor = nn.Parameter(torch.randn(8, 4))
        self.device = torch.device('cpu')
    
    @property
    def parameters(self):
        """Return the learnable parameters."""
        return self.parameters_tensor
    
    def forward(self, alpha):
        """
        Evaluate the curve at points specified by alpha.
        alpha: tensor of shape (eval_grid, 1) with values in [0, 1]
        Returns: tensor of shape (eval_grid, 4)
        """
        # Simple linear interpolation between start and end
        start = self.parameters_tensor[0:1, :]  # (1, 4)
        end = self.parameters_tensor[-1:, :]    # (1, 4)
        result = start * (1 - alpha) + end * alpha  # (eval_grid, 4)
        return result


class SimpleManifold:
    """A simple manifold object that provides curve_energy method."""
    def curve_energy(self, curve_points):
        """
        Compute energy of curve points.
        curve_points: tensor of shape (eval_grid, 4)
        Returns: tensor of shape (eval_grid,) with energy at each point
        """
        # Simple energy: sum of squared coordinates
        energy = torch.sum(curve_points ** 2, dim=1)
        return energy


# Construct the arguments
curve = SimpleCurve()
manifold = SimpleManifold()

ARGS = {
    'curve': curve,
    'manifold': manifold,
    'optimizer': torch.optim.Adam,
    'max_iter': 150,
    'eval_grid': 10,
    'dt': 0.005
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
shooting_geodesic Not yet run
unverified (harvested, no recorded run)
machinelearninglifescience/stochman/stochman/geodesic.py
code served (permissive licence) · get_code("04e61a6165b7fcf4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 3 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Latent space geometry has shown itself to provide a rich and rigorous framework for interacting with the latent variables of deep generative models. The existing theory, however, relies on the decoder being a Gaussian distribution as its simple reparametrization allows us to interpret the generating process as a random projection of a deterministic manifold. Consequently, this approach breaks down when applied to decoders that are not as easily reparametrized. We here propose to use the Fisher-Rao metric associated with the space of decoder distributions as a reference metric, which we pull back to the latent space. We show that we can achieve meaningful latent geometries for a wide range of decoder distributions for which the previous theory was not applicable, opening the door to `black box' latent geometries.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2106.05367")
get_code_for_paper("2106.05367")
have("2106.05367")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2106.05367.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2106.05367.svg)](https://syntology.ai/paper/2106.05367)

Connect an agent — have() is free.