{
  "docId": "019dd923-5e88-73ef-bd5c-f341d4394195",
  "docSlug": "46f66c49fd159048",
  "documentTitle": "2018 Air Street Capital The State of AI Report 2018",
  "authorId": "AirStreetCapital",
  "authorName": "Air Street Capital",
  "documentKindSlug": "consulting-deck",
  "documentKindLabel": "Consulting deck",
  "sourceTypeSlug": "vc_research",
  "sourceTypeLabel": "VC research",
  "presentationDate": null,
  "orientation": "landscape",
  "aspectRatio": 1.777,
  "pageNumber": 53,
  "pageCount": 156,
  "prevPage": 52,
  "nextPage": 54,
  "slideType": "case_study",
  "function": "illustrate_case",
  "density": "dense",
  "nDataPoints": 15,
  "notes": "The slide shows two neural network cell architectures and a scatter plot comparing accuracy vs. computational cost.",
  "elementsJson": [
    "headline_text",
    "action_title",
    "paragraph",
    "process_diagram",
    "scatter_plot",
    "logo_grid"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5c-f341d4394195/53",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195#slide-53",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "AutoML traversed the architecture search space to find two new cell designs that could be integrated into a final model that outperformed all existing human-crafted models.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a80e-c8813b797232",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.48,
        "w": 0.4,
        "x": 0.55,
        "y": 0.42
      },
      "kind": "chart",
      "text": "Accuracy vs. # Mult-Add operations",
      "attrs": null,
      "subkind": "scatter",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "ac35b12d-d85d-4923-bd3a-1a5ee8fe20a6",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.45,
        "x": 0.05,
        "y": 0.45
      },
      "kind": "image",
      "text": "Normal and Reduction Cell diagrams",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "2ab3e729-f3d3-41f4-9e5e-79c4a781943d",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "accuracy (precision @1)",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a80e-cd3c22173e6f",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.95,
        "x": 0.02,
        "y": 0.22
      },
      "kind": "paragraph",
      "text": "Google's AutoML automatically discovers the best model architecture for a computer vision task. AutoML traversed the architecture search space to find two new cell designs (Normal and Reduction, left figueres) that could be integrated into a final model (NASNet, right graph) that outperformed all existing human-crafted models.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "fc803c48-e9f7-4ea0-9a97-d2c78e1edb89",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.35,
        "x": 0.02,
        "y": 0.15
      },
      "kind": "title",
      "text": "AI to automate away AI engineers",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "5e55633b-5922-46c2-9476-f77769d125ad",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Action Titles",
      "slug": "action-titles",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8ecb-16354f7fb19c",
      "evidence": "Title 'AI to automate away AI engineers'.",
      "confidence": 88
    },
    {
      "name": "Unexpected Pattern",
      "slug": "unexpected-pattern",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8ecb-1bf7068e6cbc",
      "evidence": "Recursive twist — AI replacing AI engineers — surprises.",
      "confidence": 70
    },
    {
      "name": "Waterfall chart",
      "slug": "waterfall-chart",
      "agent": null,
      "layer": "slide",
      "matchId": "38105218-80ce-4f5c-9713-9171e92b6162",
      "evidence": "chart/scatter: Accuracy vs. # Mult-Add operations",
      "confidence": 0.7
    }
  ],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 69,
      "from": 4,
      "beatId": "019dd95a-0682-776c-8e34-ad4df4fe3ce7",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Definitions then research breakthroughs (transfer learning, hardware, RL) and talent supply data.",
      "position": 1,
      "confidence": 78,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 55,
      "from": 10,
      "beatId": "019dd95a-0682-776c-8e34-be65c7627fe7",
      "arcName": "The Onion",
      "arcSlug": "onion",
      "beatName": "First Layer",
      "beatSlug": "onion-first-layer",
      "evidence": "Hardware, vision, RL, bias - technical research layer.",
      "position": 2,
      "confidence": 55,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [
    {
      "to": 55,
      "from": 51,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3c-619c71c15583",
      "evidence": "Architecture iteration -> AutoML automating engineers -> federated learning -> consensus mechanism.",
      "position": 8,
      "objective": "Catalogue frontier research trends (AutoML, federated)",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 70,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}