File size: 2,646 Bytes
ab567d1
c9850e7
 
 
 
ab567d1
 
 
 
 
c9850e7
 
23e273d
 
 
c9850e7
23e273d
ab567d1
c9850e7
 
 
ab567d1
dc15376
562d09d
c9850e7
 
 
ab567d1
 
562d09d
23e273d
c9850e7
 
ab567d1
 
562d09d
c9850e7
 
 
562d09d
8d62781
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
562d09d
 
c9850e7
 
 
ab567d1
c9850e7
ab567d1
8d62781
23e273d
8d62781
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
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
{
  "schema_version": 1,
  "title": "Repro - FiBeR DP-Noise Attenuation (uMdZHVXdOL)",
  "emoji": "πŸ”‡",
  "space_id": "DineshAI/uMdZHVXdOL",
  "paper": {
    "arxiv_id": "2605.03425",
    "openreview_id": "uMdZHVXdOL",
    "title": "FiBeR: A Differentially Private Optimizer with Filter-Aware Innovation Bias Correction"
  },
  "tags": [
    "icml2026-repro",
    "paper-uMdZHVXdOL",
    "google-colab-g4",
    "gpu-rtx-pro-6000-blackwell"
  ],
  "updated_at": "2026-07-23T03:08:48+00:00",
  "root": {
    "slug": "index",
    "title": "Repro - FiBeR DP-Noise Attenuation (uMdZHVXdOL)",
    "file": "pages/index.md",
    "children": [
      {
        "slug": "claim-attenuation",
        "title": "Claim 1 β€” Attenuation identity",
        "file": "pages/claim-attenuation/page.md",
        "children": []
      },
      {
        "slug": "methods-environment",
        "title": "Methods & hardware β€” CPU C1 + Colab G4 training",
        "file": "pages/methods-environment/page.md",
        "children": []
      },
      {
        "slug": "negative-controls",
        "title": "Negative controls",
        "file": "pages/negative-controls/page.md",
        "children": []
      },
      {
        "slug": "claim-2-dp-training-comparison",
        "title": "Claim 2 β€” prior reduced-scale MNIST comparison",
        "file": "pages/claim-2-dp-training-comparison/page.md",
        "children": []
      },
      {
        "slug": "claim-3-literal-attribution",
        "title": "Claim 3 β€” literal numeric-attribution audit",
        "file": "pages/claim-3-literal-attribution/page.md",
        "children": []
      },
      {
        "slug": "claim-2-cifar10-g4",
        "title": "Claim 2 β€” CNN5/CIFAR-10 G4 campaign",
        "file": "pages/claim-2-cifar10-g4/page.md",
        "children": []
      },
      {
        "slug": "claim-3-cifar100-g4",
        "title": "Claim 3 β€” WideResNet/CIFAR-100 G4 campaign",
        "file": "pages/claim-3-cifar100-g4/page.md",
        "children": []
      },
      {
        "slug": "claim-5-matched-compute-g4",
        "title": "Claim 5 β€” matched compute G4 campaign",
        "file": "pages/claim-5-matched-compute-g4/page.md",
        "children": []
      },
      {
        "slug": "claim-6-filter-ablation-g4",
        "title": "Claim 6 β€” filter ablation and omega sweep",
        "file": "pages/claim-6-filter-ablation-g4/page.md",
        "children": []
      },
      {
        "slug": "conclusion",
        "title": "Conclusion",
        "file": "pages/conclusion/page.md",
        "children": []
      }
    ]
  },
  "agent_view_tokens": 8000,
  "revision": "1784776128000000000"
}