symmetry-group-identifier
$
npx mdskill add lyndonkl/claude/symmetry-group-identifierTranslate detected symmetries into formal group theory specifications.
- Converts identified transformations into cyclic, dihedral, or Lie group labels.
- Depends on prior symmetry discovery and Visual Group Theory taxonomy.
- Selects group names based on discrete versus continuous transformation sets.
- Outputs standardized group identifiers for equivariant network design.
SKILL.md
.github/skills/symmetry-group-identifierView on GitHub ↗
---
name: symmetry-group-identifier
description: Maps identified symmetries to mathematical groups (cyclic, dihedral, symmetric, SO(3), SE(3), E(3)) for equivariant neural network architecture design, using taxonomy and foundations from Visual Group Theory. Use when candidate symmetries have been identified and need formalization into group theory language, or when user mentions cyclic groups, dihedral groups, Lie groups, SO(3), SE(3), or permutation groups.
---
# Symmetry Group Identifier
Knowing your symmetry group tells you which equivariant architecture patterns to use. This skill formalizes identified transformations into the language of group theory.
## Workflow
Copy this checklist and track your progress:
```
Group Identification Progress:
- [ ] Step 1: List symmetries from discovery phase
- [ ] Step 2: Classify each as discrete or continuous
- [ ] Step 3: Match to specific groups using taxonomy
- [ ] Step 4: Determine how groups combine
- [ ] Step 5: Verify group properties
- [ ] Step 6: Document final group specification
```
**Step 1: List symmetries from discovery phase**
Gather the identified symmetries from the discovery phase. List each identified transformation and whether it requires invariance or equivariance. Note confidence levels. If symmetries haven't been discovered yet, work with user to identify them through domain analysis first.
**Step 2: Classify each as discrete or continuous**
For each symmetry, determine: Is the transformation set finite (discrete) or infinite (continuous)? Discrete examples: 90° rotations (4 elements), permutations of n items (n! elements). Continuous examples: rotation by any angle, translation by any distance. Use [Group Taxonomy](#group-taxonomy) to guide classification. For mathematical foundations, see [Group Theory Primer](./resources/group-theory-primer.md).
**Step 3: Match to specific groups using taxonomy**
Use the [Discrete Groups](#discrete-groups) and [Continuous Groups](#continuous-groups-lie-groups) reference sections. Identify the specific group name and notation for each symmetry. Common matches: n-fold rotation → Cₙ, rotation+reflection → Dₙ, permutation → Sₙ, 3D rotation → SO(3), rigid motion → SE(3), full Euclidean → E(3). For detailed Lie group information (SO(3), SE(3), E(3)), consult [Lie Groups Reference](./resources/lie-groups.md).
**Step 4: Determine how groups combine**
If multiple symmetries are present, determine how they combine. Direct product (G × H): symmetries act independently. Semidirect product (G ⋊ H): one symmetry "twists" the other (e.g., SE(3) = SO(3) ⋊ ℝ³). Use [Combining Groups](#combining-groups) reference.
**Step 5: Verify group properties**
Check that identified structure satisfies group axioms: closure, associativity, identity, inverses. Verify important properties: Is it compact? (affects representation theory). Is it abelian? (commutative or not). Is it connected? (affects implementation). Use [Group Properties Checklist](#group-properties-checklist). For detailed verification methodology, see [Methodology](./resources/methodology.md).
**Step 6: Document final group specification**
Create specification using [Output Template](#output-template). Include: group name/notation, dimension/size, key properties, invariance vs equivariance requirements, and recommended architecture family. This specification provides the foundation for architecture design. Quality criteria for this output are defined in [Quality Rubric](./resources/evaluators/rubric_group_identification.json).
## Group Taxonomy
### Overview Diagram
```
SYMMETRY GROUPS
│
┌───────────────┴───────────────┐
│ │
DISCRETE CONTINUOUS
│ (Lie Groups)
│ │
┌─────┼─────┐ ┌────────┼────────┐
│ │ │ │ │ │
Cyclic Dihedral Symmetric SO(n) SE(n) E(n)
Cₙ Dₙ Sₙ rotations rigid Euclidean
only motions (w/ reflect)
```
### Quick Reference Table
| Symmetry Type | Group | Notation | Elements | Common Use |
|---------------|-------|----------|----------|------------|
| n-fold rotation | Cyclic | Cₙ | n | Image rotation (90°, 60°) |
| Rotation + reflection | Dihedral | Dₙ | 2n | Regular polygons |
| Permutation | Symmetric | Sₙ | n! | Sets, graphs |
| 2D rotation (continuous) | Special orthogonal | SO(2) | ∞ | Continuous rotation |
| 3D rotation | Special orthogonal | SO(3) | ∞ | 3D orientation |
| 3D rigid motion | Special Euclidean | SE(3) | ∞ | Robotics, molecules |
| 3D with reflections | Euclidean | E(3) | ∞ | Chemistry, physics |
## Discrete Groups
### Cyclic Groups (Cₙ)
**What they represent**: Rotations by multiples of 360°/n
**Elements**: {e, r, r², ..., rⁿ⁻¹} where rⁿ = e (identity)
| Group | Rotations | Example |
|-------|-----------|---------|
| C₂ | 0°, 180° | Playing cards |
| C₄ | 0°, 90°, 180°, 270° | Square images |
| C₆ | 60° increments | Hexagonal patterns |
**Use when**: Rotation symmetry present but NOT reflection symmetry.
### Dihedral Groups (Dₙ)
**What they represent**: Rotations + reflections of regular n-gon
**Elements**: n rotations + n reflections = 2n total
| Group | Elements | Example |
|-------|----------|---------|
| D₄ | 8 | Square with diagonals (p4m group) |
| D₆ | 12 | Regular hexagon |
**Use when**: Both rotation AND reflection symmetry present.
### Symmetric Groups (Sₙ)
**What they represent**: All permutations of n elements
**Elements**: n! permutations
**Use when**: Element ordering is arbitrary (sets, graphs, point clouds).
## Continuous Groups (Lie Groups)
### SO(2) - 2D Rotations
**Elements**: Rotation by any angle θ ∈ [0, 2π)
**Matrix form**: R(θ) = [[cos θ, -sin θ], [sin θ, cos θ]]
**Use when**: Continuous rotation symmetry in 2D.
### SO(3) - 3D Rotations
**Elements**: All rotations in 3D (3 degrees of freedom)
**Representations**: Rotation matrices, quaternions, Euler angles, axis-angle
**Use when**: 3D orientation doesn't matter, but handedness does.
### SE(3) - 3D Rigid Motions
**Elements**: Rotations + Translations in 3D
**Structure**: SE(3) = SO(3) ⋊ ℝ³ (semidirect product)
**Use when**: Objects can be anywhere and in any orientation, handedness matters.
### E(3) - Full Euclidean Group
**Elements**: SE(3) + Reflections
**Structure**: E(3) = O(3) ⋊ ℝ³
**Use when**: SE(3) symmetry PLUS reflection symmetry (most molecules).
### Group Hierarchy
```
E(3) = O(3) ⋊ ℝ³
│ exclude reflections
▼
SE(3) = SO(3) ⋊ ℝ³
│ exclude translations
▼
SO(3)
│ 2D restriction
▼
SO(2)
```
## Combining Groups
### Direct Product (G × H)
**When to use**: Symmetries act independently (neither affects the other).
**Example**: Image with separate translation and color permutation → SE(2) × S₃
**Property**: (g₁, h₁) · (g₂, h₂) = (g₁g₂, h₁h₂)
### Semidirect Product (G ⋊ H)
**When to use**: One symmetry "twists" the other (don't commute).
**Example**: SE(3) = SO(3) ⋊ ℝ³ (rotating then translating ≠ translating then rotating)
**Common cases**: SE(n) = SO(n) ⋊ ℝⁿ, E(n) = O(n) ⋊ ℝⁿ, Dₙ = Cₙ ⋊ C₂
## Group Properties Checklist
For your identified group, verify:
| Property | Question | Why It Matters |
|----------|----------|----------------|
| Compact | Is the group "bounded"? | Affects representation theory |
| Abelian | Does order matter? (g₁g₂ = g₂g₁?) | Simplifies architecture |
| Connected | Is group in one piece? | Affects irreducible representations |
| Finite | Finite number of elements? | Discrete vs continuous architecture |
## Group Selection by Domain
| Domain | Typical Group | Notes |
|--------|--------------|-------|
| 2D Image Classification | C₄ or D₄ | p4 or p4m groups |
| 3D Molecular Energy | E(3) × Sₙ | Full Euclidean + atom permutation |
| 3D Molecular Chirality | SE(3) × Sₙ | No reflections |
| Point Cloud Classification | SO(3) × Sₙ | Rotation + permutation |
| Graph Classification | Sₙ | Permutation invariant |
| Robotics | SE(3) | Sometimes with gravity constraint |
## Output Template
```
SYMMETRY GROUP SPECIFICATION
============================
Identified Symmetries:
1. [Symmetry] → Group: [name] ([notation])
2. [Symmetry] → Group: [name] ([notation])
Combined Group Structure:
- Full group: [G₁ × G₂] or [G₁ ⋊ G₂]
- Size: [# elements] or [continuous]
Group Properties:
- Compact: [Yes/No]
- Abelian: [Yes/No]
- Connected: [Yes/No]
Symmetry Requirements:
- [Group]: [Invariant/Equivariant] for [task type]
Recommended Architecture Family:
- [Architecture] supporting [group]
NEXT STEPS:
- Empirically validate symmetry hypotheses if not yet confirmed
- Design equivariant architecture based on group specification
```
More from lyndonkl/claude
- abstraction-concrete-examplesBuilds structured abstraction ladders that translate high-level principles into concrete, actionable examples across 3-5 levels. Bridges communication gaps, reveals hidden assumptions, and tests whether abstract ideas work in practice. Use when explaining concepts at different expertise levels, moving between abstract principles and concrete implementation, identifying edge cases by testing ideas against scenarios, designing layered documentation, decomposing complex problems into actionable steps, or bridging strategy-execution gaps.
- academic-letter-architectGuides the creation of evidence-based academic recommendation letters, reference letters, and award nominations that combine concrete examples, meaningful comparisons, and genuine enthusiasm. Use when writing recommendation letters for students, postdocs, or colleagues, or when user mentions recommendation letter, reference, nomination, letter of support, endorsement, or needs help with strong advocacy and comparative statements.
- adr-architectureDocuments significant architectural and technical decisions with full context, alternatives considered, trade-offs analyzed, and consequences understood. Creates a decision trail that helps teams understand why decisions were made. Use when choosing between technology options, making infrastructure decisions, establishing standards, migrating systems, or when user mentions ADR, architecture decision, technical decision record, or decision documentation.
- adverse-selection-priorProduces a Bayesian prior probability that an offered transaction is +EV for the recipient, given that the counterparty chose to propose it. Applies Akerlof market-for-lemons logic -- if they offered it, they believe it is +EV for them, so the prior that it is +EV for us is materially below 50%. Reusable across trade evaluation, waiver drops (another team dropping a player is also adverse selection), job-offer analysis, M&A, and any "someone offered me this" situation. Use when you receive an unsolicited trade/offer/proposal, analyzing incoming trade prior, evaluating why a counterparty proposed a deal, or when user mentions adverse selection, market for lemons, why did they offer this, incoming trade prior, they proposed it, Bayesian adjustment on received offer.
- alignment-values-north-starCreates actionable alignment frameworks that give teams a shared North Star (direction), values (guardrails), and decision tenets (behavioral standards). Enables autonomous decision-making while maintaining organizational coherence. Use when starting new teams, scaling organizations, defining culture, establishing product vision, resolving misalignment, creating strategic clarity, or when user mentions North Star, team values, mission, principles, guardrails, decision framework, or cultural alignment.
- analogy-weight-checkFor every analogy in a substacker draft, verifies it carries mechanical weight — the analogy does real work explaining the mechanism, not merely decorates it. Cross-references analogy-catalog.md for novelty (is this analogy reused from a prior post?) and domain fit (biology > organizational > sports preferred; physics/military disfavored). Use whenever an analogy appears in the draft. Trigger keywords: analogy weight, decorative, mechanical weight, reused analogy, catalog check, metaphor check.
- answer-uncomfortable-questionTakes one strategic question about substacker ("should we launch paid?", "is this section dead?", "are we writing for the wrong audience?") and produces the mandatory evidence + reasoning + downside triad plus a recommendation. Used 3 times per Growth Strategist review. Trigger keywords: uncomfortable question, strategic question, evidence reasoning downside, triad.
- attribute-performanceFor each substacker post that materially over- or under-performs the rolling baseline (|z| ≥ 1.0), produces a plain-English attribution paragraph with calibrated confidence (high / medium / low / unexplained). Considers subject-line effect, topic zeitgeist, external share, day-of-week, length effect, and audience-notes signals. Labels unexplained outliers explicitly rather than fabricating a story. Use after compute-baseline when outlier posts exist. Trigger keywords: attribution, why did this post work, outlier explanation, performance analysis.
- auction-first-price-shadingComputes the optimal shaded bid for a first-price sealed-bid auction given a true private value, an estimate of the number of competing bidders N, and a value-distribution assumption. Implements the `(N-1)/N` equilibrium shading rule for uniform private values, adjusts for log-normal or empirical value distributions, layers a risk-aversion adjustment, and caps output against the bidder's remaining budget. Domain-neutral auction theory reusable across fantasy sports (baseball FAAB, NBA/NHL waiver auctions), prediction-market limit sizing, sealed procurement bids, and any blind-bid context. Use when user mentions "first-price auction bid", "sealed bid shading", "(N-1)/N", "FAAB bid amount", "auction shading", "optimal bid first-price", "bid for sealed-bid", "blind bid sizing", or when downstream logic needs a principled shade factor rather than an ad-hoc heuristic.
- auction-winners-curse-haircutApplies a Bayesian haircut to a bid valuation for common-value auctions where winning is itself evidence the bidder over-estimated. Takes a raw valuation, a value-type classification (common_value / private_value / mixed), the number of informed bidders N, and a signal-dispersion estimate, and returns an adjusted valuation. Domain-neutral and reusable across fantasy FAAB, prediction markets, M&A bids, ad-auction budgets, and any generic bidding context. Use when user mentions "winner's curse", "common value auction", "valuation haircut", "adverse valuation", "Bayesian bid adjustment", or "over-paying in auction".