File size: 2,874 Bytes
b4a2ae8 | 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 | {
"schema_version": 1,
"title": "Reproduction: High-Dimensional Learning Dynamics of Quantized Models with Straight-Through Estimator",
"emoji": "📉",
"space_id": "snaykey/repro-quantized-ste-dynamics",
"paper": {
"openreview_id": "bI9moH3UZw"
},
"tags": [
"icml2026-repro",
"paper-bI9moH3UZw"
],
"updated_at": "2026-08-02T00:00:00+00:00",
"root": {
"slug": "index",
"title": "Reproduction: High-Dimensional Learning Dynamics of Quantized Models with Straight-Through Estimator",
"file": "pages/index.md",
"children": [
{
"slug": "claim-1-thm43-ste-sde",
"title": "In the high-dimensional limit, the microscopic dynamics of straight-through estimator (STE) training converge to a stochastic differential equation governed by Eq. (15) (Section IV, Theorem IV.3).",
"file": "pages/claim-1-thm43-ste-sde/page.md",
"children": []
},
{
"slug": "claim-2-thm53-ode-concentration",
"title": "The macroscopic summary statistics of STE training concentrate onto a deterministic ODE trajectory (Eqs. (24)-(25)) with concentration rate O_d(d^{-1/2}) as dimension d grows (Section V-B, Theorem V.3).",
"file": "pages/claim-2-thm53-ode-concentration/page.md",
"children": []
},
{
"slug": "claim-3-plateau-drop-bitwidth",
"title": "STE training exhibits an extended plateau followed by a sharp drop in generalization error across bit-widths b in {2,3,4,5}, with the plateau length depending on the quantization range ω (Section VI-A1/VI-A2, Figures 2-3).",
"file": "pages/claim-3-plateau-drop-bitwidth/page.md",
"children": []
},
{
"slug": "claim-4-propvi1-input-fixedpoint",
"title": "For input-only quantization, a closed-form stable fixed point exists under the learning-rate condition 0 < η < 2(σ_ψ^2 + λ)/σ_ψ^4 (Section VI-C1, Proposition VI.1).",
"file": "pages/claim-4-propvi1-input-fixedpoint/page.md",
"children": []
},
{
"slug": "claim-5-thmvi3-three-regime",
"title": "The asymptotic deviation of the STE-trained quantized model from the unquantized linear model is characterized across three regimes, with a leading-order correction term of size O_Δ(Δ^2) depending on the fractional quantization level position (Section VI-C2, Theorem VI.3).",
"file": "pages/claim-5-thmvi3-three-regime/page.md",
"children": []
},
{
"slug": "executive-summary",
"title": "executive-summary",
"file": "pages/executive-summary/page.md",
"children": []
},
{
"slug": "conclusion",
"title": "conclusion",
"file": "pages/conclusion/page.md",
"children": []
}
]
}
} |