{
  "article_slug": "mixup-vicinal-risk",
  "figures": [
    {
      "id": "mixup-vicinal-risk-1",
      "kind": "line_chart",
      "title": "Endpoint agreement leaves the interior target unspecified",
      "x_label": "Input x",
      "y_label": "Target probability / prediction",
      "series": [
        {
          "name": "Linear mixed target",
          "x": [
            0.0,
            0.01,
            0.02,
            0.03,
            0.04,
            0.05,
            0.06,
            0.07,
            0.08,
            0.09,
            0.1,
            0.11,
            0.12,
            0.13,
            0.14,
            0.15,
            0.16,
            0.17,
            0.18,
            0.19,
            0.2,
            0.21,
            0.22,
            0.23,
            0.24,
            0.25,
            0.26,
            0.27,
            0.28,
            0.29,
            0.3,
            0.31,
            0.32,
            0.33,
            0.34,
            0.35,
            0.36,
            0.37,
            0.38,
            0.39,
            0.4,
            0.41,
            0.42,
            0.43,
            0.44,
            0.45,
            0.46,
            0.47,
            0.48,
            0.49,
            0.5,
            0.51,
            0.52,
            0.53,
            0.54,
            0.55,
            0.56,
            0.57,
            0.58,
            0.59,
            0.6,
            0.61,
            0.62,
            0.63,
            0.64,
            0.65,
            0.66,
            0.67,
            0.68,
            0.69,
            0.7,
            0.71,
            0.72,
            0.73,
            0.74,
            0.75,
            0.76,
            0.77,
            0.78,
            0.79,
            0.8,
            0.81,
            0.82,
            0.83,
            0.84,
            0.85,
            0.86,
            0.87,
            0.88,
            0.89,
            0.9,
            0.91,
            0.92,
            0.93,
            0.94,
            0.95,
            0.96,
            0.97,
            0.98,
            0.99,
            1.0
          ],
          "y": [
            0.0,
            0.01,
            0.02,
            0.03,
            0.04,
            0.05,
            0.06,
            0.07,
            0.08,
            0.09,
            0.1,
            0.11,
            0.12,
            0.13,
            0.14,
            0.15,
            0.16,
            0.17,
            0.18,
            0.19,
            0.2,
            0.21,
            0.22,
            0.23,
            0.24,
            0.25,
            0.26,
            0.27,
            0.28,
            0.29,
            0.3,
            0.31,
            0.32,
            0.33,
            0.34,
            0.35,
            0.36,
            0.37,
            0.38,
            0.39,
            0.4,
            0.41,
            0.42,
            0.43,
            0.44,
            0.45,
            0.46,
            0.47,
            0.48,
            0.49,
            0.5,
            0.51,
            0.52,
            0.53,
            0.54,
            0.55,
            0.56,
            0.57,
            0.58,
            0.59,
            0.6,
            0.61,
            0.62,
            0.63,
            0.64,
            0.65,
            0.66,
            0.67,
            0.68,
            0.69,
            0.7,
            0.71,
            0.72,
            0.73,
            0.74,
            0.75,
            0.76,
            0.77,
            0.78,
            0.79,
            0.8,
            0.81,
            0.82,
            0.83,
            0.84,
            0.85,
            0.86,
            0.87,
            0.88,
            0.89,
            0.9,
            0.91,
            0.92,
            0.93,
            0.94,
            0.95,
            0.96,
            0.97,
            0.98,
            0.99,
            1.0
          ]
        },
        {
          "name": "Hard threshold target",
          "x": [
            0.0,
            0.01,
            0.02,
            0.03,
            0.04,
            0.05,
            0.06,
            0.07,
            0.08,
            0.09,
            0.1,
            0.11,
            0.12,
            0.13,
            0.14,
            0.15,
            0.16,
            0.17,
            0.18,
            0.19,
            0.2,
            0.21,
            0.22,
            0.23,
            0.24,
            0.25,
            0.26,
            0.27,
            0.28,
            0.29,
            0.3,
            0.31,
            0.32,
            0.33,
            0.34,
            0.35,
            0.36,
            0.37,
            0.38,
            0.39,
            0.4,
            0.41,
            0.42,
            0.43,
            0.44,
            0.45,
            0.46,
            0.47,
            0.48,
            0.49,
            0.5,
            0.51,
            0.52,
            0.53,
            0.54,
            0.55,
            0.56,
            0.57,
            0.58,
            0.59,
            0.6,
            0.61,
            0.62,
            0.63,
            0.64,
            0.65,
            0.66,
            0.67,
            0.68,
            0.69,
            0.7,
            0.71,
            0.72,
            0.73,
            0.74,
            0.75,
            0.76,
            0.77,
            0.78,
            0.79,
            0.8,
            0.81,
            0.82,
            0.83,
            0.84,
            0.85,
            0.86,
            0.87,
            0.88,
            0.89,
            0.9,
            0.91,
            0.92,
            0.93,
            0.94,
            0.95,
            0.96,
            0.97,
            0.98,
            0.99,
            1.0
          ],
          "y": [
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            0,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1,
            1
          ]
        }
      ],
      "caption": "Constructed endpoints (0,0),(1,1). Uniform squared probability discrepancy is 1/12, although thresholded class decisions agree except for boundary convention."
    },
    {
      "id": "mixup-vicinal-risk-2",
      "kind": "line_chart",
      "title": "The same mean mixing weight can imply different constraints",
      "x_label": "Symmetric beta shape alpha",
      "y_label": "Moment (dimensionless)",
      "series": [
        {
          "name": "E[lambda*(1-lambda)]",
          "x": [
            0.1,
            0.2,
            0.5,
            1,
            2,
            4,
            8,
            16
          ],
          "y": [
            0.04166666666666667,
            0.07142857142857144,
            0.125,
            0.16666666666666666,
            0.2,
            0.2222222222222222,
            0.23529411764705882,
            0.24242424242424243
          ]
        },
        {
          "name": "Var(lambda)",
          "x": [
            0.1,
            0.2,
            0.5,
            1,
            2,
            4,
            8,
            16
          ],
          "y": [
            0.20833333333333334,
            0.17857142857142858,
            0.125,
            0.08333333333333333,
            0.05,
            0.027777777777777776,
            0.014705882352941176,
            0.007575757575757576
          ]
        }
      ],
      "caption": "Exact Beta(alpha,alpha) moments; mean lambda is 0.5 for every point. Not a universal neural regularization coefficient."
    }
  ]
}