Spaces:
Running
Running
File size: 2,468 Bytes
ed009a5 46a9ed9 ed009a5 46a9ed9 ed009a5 46a9ed9 ed009a5 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 | {
"schema_version": 1,
"title": "Rotary Position Encodings for Graphs (WIRE)",
"emoji": "🌊",
"space_id": "snaykey/repro-wire",
"paper": {
"arxiv_id": "2509.22259",
"openreview_id": "trn64znfNx"
},
"tags": [
"icml2026-repro",
"paper-trn64znfNx"
],
"updated_at": "2026-07-21T17:22:51+00:00",
"root": {
"slug": "index",
"title": "Rotary Position Encodings for Graphs (WIRE)",
"file": "pages/index.md",
"children": [
{
"slug": "executive-summary",
"title": "Executive summary",
"file": "pages/executive-summary/page.md",
"children": []
},
{
"slug": "claim-1-wire-algorithm",
"title": "WIRE applies rotary position encodings to graphs by using Laplacian spectral coordinates to define graph-dependent rotation angles (Figure 1, Section 3).",
"file": "pages/claim-1-wire-algorithm/page.md",
"children": []
},
{
"slug": "claim-2-equivariance",
"title": "The WIRE transformation is equivariant to node-order permutations up to sign flips and rotations in degenerate eigenspaces (Lemma 1).",
"file": "pages/claim-2-equivariance/page.md",
"children": []
},
{
"slug": "claim-3-rope-special-case",
"title": "Regular RoPE is recovered as a special case of WIRE on grid graphs with appropriate learnable frequencies (Theorem 2, Figure 2).",
"file": "pages/claim-3-rope-special-case/page.md",
"children": []
},
{
"slug": "claim-4-effective-resistance",
"title": "WIRE asymptotically depends on graph effective resistance under the paper's spectral-feature assumptions (Theorem 3).",
"file": "pages/claim-4-effective-resistance/page.md",
"children": []
},
{
"slug": "conclusion",
"title": "Conclusion",
"file": "pages/conclusion/page.md",
"children": []
},
{
"slug": "claim-5-empirical-tables",
"title": "Experiments report WIRE improvements on synthetic monochromatic-subgraph and shortest-path tasks, point-cloud transformers, and efficient graph transformer benchmarks (Table 1, Table 2, Table 3, Table 4).",
"file": "pages/claim-5-empirical-tables/page.md",
"children": []
}
]
},
"agent_view_tokens": 2954,
"revision": "1784654571781165900"
} |