silva_multiscale_graph_implicit Reproduction Dossier
multiscale graph implicit network with nodewise fusion. This dossier connects the source mechanism to its SILVA implementation, compact evidence, replaceable components, and source-scale route. Existing tests and notebooks remain the executable authority.
Evidence boundary
The mechanism is compact-verified in the package suite.
The final source-scale stage remains planned until the cited data, complete
optimization budget, checkpoints, and evaluation protocol have actually run.
Identity and Sources
| Field | Value |
|---|---|
| Domain | graphs and distributed systems |
| Task contract | N,D node features and graph operator -> per-scale equilibria and fused node output |
| Source relation | paper-adaptation |
| References | [79] |
| Repositories | https://github.com/liu-jc/MGNNI |
| Editable scale plan | experiments/reproduction/configs/silva_multiscale_graph_implicit.json |
Governing Equation
The domain-level state contract is
The implementation registry specializes it operationally as
Define the root residual
At a regular equilibrium, differentiating \(R_\theta(z^\star;x)=0\) gives
This identity explains why the forward residual, the conditioning derivative, and the adjoint linear solve must be diagnosed separately from the task metric.
What Is Preserved
- one infinite graph equilibrium for each declared graph-power scale
- independent normalized channel factors across scales
- nodewise softmax attention over converged scale states
What Can Be Replaced
Each item below is an explicit control rather than an undocumented modification:
- replace each scale factor, source, fusion attention, readout, or solver
- add graph powers while retaining separately inspectable equilibria
- graph-power scales
- per-scale widths
- sparse edges
- solver budgets
Constructor and Shape Contract
silva_multiscale_graph_implicit(in_dim: 'int', state_dim: 'int', out_dim: 'int', *, scales: 'Sequence[int]' = (1, 2), gamma: 'float' = 0.8, source: 'nn.Module | None' = None, graph_source: 'nn.Module | None' = None, readout: 'nn.Module | None' = None, fusion: 'ScaleFusion' = 'attention', attention_dim: 'int | None' = None, gram_epsilon: 'float' = 1e-12, config: 'SolverConfig | Sequence[SolverConfig] | None' = None) -> 'None'
The transition must preserve the declared equilibrium-state shape even when the encoder, branch operators, constraints, solver, and readout are replaced. Test the transition by itself before testing the complete root solve.
Progressive Experiment Ladder
1. Equation and tensor contract
Objective: Make the state, conditioning variables, operator, and readout explicit.
Procedure:
- Write and evaluate the family equation:
Z_m_star=gamma S^m Z_m_star g(F_m)^T+X; Z=sum_m beta_m(Z_m_star)Z_m_star - Declare every tensor axis, boundary, mask, graph, or physical unit.
- Check the transition output has exactly the same shape as the equilibrium state.
Acceptance checks:
- finite transition values
- shape-preserving state update
- all conditioning variables affect the intended branch
Evidence target: contract-verified.
2. Primitive mechanism reconstruction
Objective: Build the retained source mechanism from replaceable modules.
Procedure:
- one infinite graph equilibrium for each declared graph-power scale
- independent normalized channel factors across scales
- nodewise softmax attention over converged scale states
Acceptance checks:
- primitive modules expose trainable parameters and gradients
- mechanism-specific invariance or constraint check passes
- direct transition evaluation is deterministic under a fixed seed
Evidence target: compact-verified.
3. Public abstraction equivalence
Objective: Verify that the assembled family evaluates the same transition as its primitives.
Procedure:
- Copy the primitive module parameters into the public family constructor.
- Evaluate one transition and one complete equilibrium with identical inputs.
- Compare outputs, residuals, and parameter gradients with declared tolerances.
Acceptance checks:
- transition outputs agree
- equilibrium residual is finite and decreases
- primitive and assembled gradients agree on the compact case
Evidence target: compact-verified.
4. Compact real or analytic task
Objective: Exercise training, evaluation, diagnostics, and serialization end to end.
Procedure:
- make_mgnni_multiscale_dataset gives per-scale graph equilibria and node-dependent target fusion weights.
Acceptance checks:
- record node accuracy
- record per-scale residual
- record attention entropy and scale usage
- record runtime and memory
- checkpoint reload reproduces the recorded prediction
- result record contains data and configuration fingerprints
Evidence target: compact-verified.
5. Official-data subset
Objective: Validate the complete source data path before spending the full budget.
Procedure:
- Acquire a declared graph benchmark and preserve the official split, graph normalization, and feature preprocessing.
- Freeze preprocessing, split logic, metric code, and checkpoint format.
- Run a deterministic subset large enough to expose batching and memory failures.
Acceptance checks:
- dataset receipt and checksum are stored
- resume and evaluation paths reproduce the same subset metric
- memory and runtime are measured rather than estimated
Evidence target: subset-verified.
6. Source-scale reproduction or declared extension
Objective: Run the cited protocol, or change it explicitly as a SILVA extension.
Procedure:
- Acquire a declared graph benchmark and preserve the official split, graph normalization, and feature preprocessing.
- Match graph-power scales, per-scale channel factors, equilibrium budgets, and nodewise attention fusion.
- Validate per-scale states and normalized attention on the compact case, then report task accuracy, residuals, fusion statistics, runtime, and memory.
- source graph split, features, graph normalization, labels, and scale list
- per-scale widths, gamma, attention dimension, optimizer, early stopping, and seeds
- node accuracy, per-scale residuals, attention statistics, runtime, and memory
Acceptance checks:
- all required artifacts are archived
- reported metrics use the cited evaluation protocol
- every architectural or training deviation is listed
- claims match the achieved evidence status
Evidence target: planned.
Data, Access, and Storage
Candidate datasets:
- Cora
- Citeseer
- Pubmed
- Amazon
- Coauthor
- compact multiscale chain graphs
Authoritative routes:
- https://arxiv.org/abs/2210.08353
- https://github.com/liu-jc/MGNNI
Access obligations:
- Use the benchmark provider's original graph, labels, and declared transductive split.
- Cache graph powers or sparse propagation plans by dataset checksum, normalization, and scale list.
Storage planning:
- State storage scales with nodes times state width times the number of graph scales, plus per-scale solver history.
- Cache sparse graph powers or repeated sparse propagation plans rather than materializing dense matrices.
Preprocessing record:
- record dataset version, split, normalization, shape convention, and seed
- preserve masks, graph indices, boundaries, or physical units required by the domain
Metrics and Current Evidence
Required metrics:
- node accuracy
- per-scale residual
- attention entropy and scale usage
- runtime and memory
The graph compact suite ran this family on the same task and data as
the other compatible families in that suite.
| Measure | Recorded value |
|---|---|
| Initial loss | 0.628229 |
| Final loss | 0.0291975 |
| Fractional loss reduction | 0.954 |
| Residual or final increment norm | 0.00382243 |
| Iterations or tied increments | 56 |
| Parameter count | 111 |
| Final gradient norm | 0.259409 |
These values are compact-verified evidence. They establish finite optimization, gradient flow, and diagnostic reporting; they are not a publication ranking.
Executed notebook paths:
- notebooks/package_api/40_silva_multiscale_graph_implicit.ipynb
Mechanism tests:
- tests/test_structured_equilibria.py
Compact Defaults
| Option | Value |
|---|---|
tier |
'smoke' |
config |
SolverConfig(solver='anderson', max_iter=12, tol=1e-05, alpha=1.0, history=3, ridge=0.0001, beta=1.0, stop_mode='relative', relative_eps=1e-08, anderson_batch_dims=0, track_residuals=True, reengage=True, backward_mode='implicit', backward_solver='gmres', backward_max_iter=20, backward_tol=1e-05, backward_stop_mode='relative', backward_relative_eps=1e-08, phantom_steps=1, phantom_tau=1.0, neumann_terms=5, shine_refine_steps=0, indexing=(), return_best=True) |
Full Defaults
| Option | Value |
|---|---|
tier |
'full' |
config |
SolverConfig(solver='anderson', max_iter=60, tol=1e-05, alpha=1.0, history=6, ridge=0.0001, beta=1.0, stop_mode='relative', relative_eps=1e-08, anderson_batch_dims=0, track_residuals=True, reengage=True, backward_mode='implicit', backward_solver='gmres', backward_max_iter=80, backward_tol=1e-05, backward_stop_mode='relative', backward_relative_eps=1e-08, phantom_steps=1, phantom_tau=1.0, neumann_terms=5, shine_refine_steps=0, indexing=(), return_best=True) |
Defaults establish a starting budget; the cited source protocol takes precedence whenever reproduction is the claim.
Source-Scale Checklist
- Acquire a declared graph benchmark and preserve the official split, graph normalization, and feature preprocessing.
- Match graph-power scales, per-scale channel factors, equilibrium budgets, and nodewise attention fusion.
- Validate per-scale states and normalized attention on the compact case, then report task accuracy, residuals, fusion statistics, runtime, and memory.
Benchmark-specific requirements:
- source graph split, features, graph normalization, labels, and scale list
- per-scale widths, gamma, attention dimension, optimizer, early stopping, and seeds
- node accuracy, per-scale residuals, attention statistics, runtime, and memory
Required archived artifacts:
- machine-readable model and solver configuration
- dataset receipt with source revision, split, license, and checksum
- preprocessing and normalization record
- seeded training and evaluation log
- checkpoint and optimizer-resume state for trained experiments
- task metrics and equilibrium diagnostics in a machine-readable result
- runtime, peak-memory, device, precision, and dependency record
- declared deviations from the cited protocol
Reporting Rule
Report the achieved evidence status, not the intended one. A compact or subset run may validate the implementation and data path, but only a completed cited protocol supports a source-scale reproduction statement. Modified operators are valuable SILVA extensions when every deviation is named and measured.
Where to Go Next
| Question | Page |
|---|---|
| Where are all family dossiers? | Family Dossier Index |
| How is a custom family assembled? | Advanced Extension Handbook |
| How are experiment stages represented in the API? | Research-Depth API |
| Which lab inspects every dossier? | Family Dossier Lab |