{
  "wireMagic": "pb3",
  "pens": [ {
  
  } ],
  "wsWinStyles": [ {
  
  } ],
  "wpWinPositions": [ {
  
  } ],
  "events": [ {
    "tStartMs": 7360,
    "dDurationMs": 4400,
    "segs": [ {
      "utf8": "Machine learning is only as good as the\ntraining data you put into it."
    } ]
  }, {
    "tStartMs": 11800,
    "dDurationMs": 4020,
    "segs": [ {
      "utf8": "So, it's super important to use high quality data, and lots of it."
    } ]
  }, {
    "tStartMs": 16760,
    "dDurationMs": 5200,
    "segs": [ {
      "utf8": "But if data is important, it's worth asking where does training data come from?"
    } ]
  }, {
    "tStartMs": 22280,
    "dDurationMs": 3980,
    "segs": [ {
      "utf8": "Often, computers are collecting training data from people like you and me,"
    } ]
  }, {
    "tStartMs": 26260,
    "dDurationMs": 1600,
    "segs": [ {
      "utf8": "without any effort on our part."
    } ]
  }, {
    "tStartMs": 28440,
    "dDurationMs": 3040,
    "segs": [ {
      "utf8": "A video streaming service might keep track of what you watch, then it can recognize patterns"
    } ]
  }, {
    "tStartMs": 31660,
    "dDurationMs": 4340,
    "segs": [ {
      "utf8": "in that data to recommend what you might want to watch next."
    } ]
  }, {
    "tStartMs": 37420,
    "dDurationMs": 5780,
    "segs": [ {
      "utf8": "Other times, you're directly asked to help, like when a website asks you to spot street signs and photos,"
    } ]
  }, {
    "tStartMs": 43780,
    "dDurationMs": 5500,
    "segs": [ {
      "utf8": "You're providing training data to help a\nmachine learn to see, and maybe even one day drive."
    } ]
  }, {
    "tStartMs": 52320,
    "dDurationMs": 4120,
    "segs": [ {
      "utf8": "Medical researchers can use\nmedical images as training data to teach"
    } ]
  }, {
    "tStartMs": 56520,
    "dDurationMs": 3380,
    "segs": [ {
      "utf8": "computers how to recognize and diagnose diseases."
    } ]
  }, {
    "tStartMs": 60300,
    "dDurationMs": 5260,
    "segs": [ {
      "utf8": "Machine Learning needs hundreds and thousands of images, and training direction from a doctor"
    } ]
  }, {
    "tStartMs": 65640,
    "dDurationMs": 4280,
    "segs": [ {
      "utf8": "who knows what to look for, before it can correctly identify disease."
    } ]
  }, {
    "tStartMs": 70520,
    "dDurationMs": 5020,
    "segs": [ {
      "utf8": "Even with thousands of examples, there can be problems with the computer's predictions."
    } ]
  }, {
    "tStartMs": 75880,
    "dDurationMs": 4780,
    "segs": [ {
      "utf8": "If X-ray data is only collected from men, then the computer's predictions may only work for men."
    } ]
  }, {
    "tStartMs": 81880,
    "dDurationMs": 4420,
    "segs": [ {
      "utf8": "It may not recognize diseases when\nasked to diagnose the X-ray of a woman."
    } ]
  }, {
    "tStartMs": 86620,
    "dDurationMs": 4200,
    "segs": [ {
      "utf8": "This blind spot in the training data\ncreates something called bias."
    } ]
  }, {
    "tStartMs": 91260,
    "dDurationMs": 5160,
    "segs": [ {
      "utf8": "Biased data favors some things, and de-prioritizes or excludes others."
    } ]
  }, {
    "tStartMs": 96780,
    "dDurationMs": 5020,
    "segs": [ {
      "utf8": "Depending on how training data is collected, who is doing the collecting, and how the data is fed,"
    } ]
  }, {
    "tStartMs": 101800,
    "dDurationMs": 3540,
    "segs": [ {
      "utf8": "there is a chance that\nhuman bias is included in the data."
    } ]
  }, {
    "tStartMs": 105880,
    "dDurationMs": 4820,
    "segs": [ {
      "utf8": "By learning from bias data, the computer may make biased predictions,"
    } ]
  }, {
    "tStartMs": 110780,
    "dDurationMs": 3540,
    "segs": [ {
      "utf8": "whether the people training the computer\nare aware of it or not."
    } ]
  }, {
    "tStartMs": 114760,
    "dDurationMs": 3640,
    "segs": [ {
      "utf8": "When you are looking at training data, ask yourself two questions:"
    } ]
  }, {
    "tStartMs": 118640,
    "dDurationMs": 2960,
    "segs": [ {
      "utf8": "Is this enough data to accurately train a computer?"
    } ]
  }, {
    "tStartMs": 122320,
    "dDurationMs": 4540,
    "segs": [ {
      "utf8": "And, does this data represent all possible scenarios and users without bias?"
    } ]
  }, {
    "tStartMs": 127460,
    "dDurationMs": 3580,
    "segs": [ {
      "utf8": "This is where you, as the human training, play a crucial role."
    } ]
  }, {
    "tStartMs": 131160,
    "dDurationMs": 3340,
    "segs": [ {
      "utf8": "It's up to you to give your machine unbiased data."
    } ]
  }, {
    "tStartMs": 134500,
    "dDurationMs": 3660,
    "segs": [ {
      "utf8": "That means collecting tons of examples, from lots of sources."
    } ]
  }, {
    "tStartMs": 139300,
    "dDurationMs": 3280,
    "segs": [ {
      "utf8": "Remember, when you pick and choose data for machine learning,"
    } ]
  }, {
    "tStartMs": 142580,
    "dDurationMs": 4080,
    "segs": [ {
      "utf8": "you're actually programming the algorithm, using training data instead of code."
    } ]
  }, {
    "tStartMs": 147100,
    "dDurationMs": 2680,
    "segs": [ {
      "utf8": "The data IS the code."
    } ]
  }, {
    "tStartMs": 150180,
    "dDurationMs": 4500,
    "segs": [ {
      "utf8": "The better the data you provide, the better the computer will learn."
    } ]
  } ]
}