{
  "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": 31,
  "pageCount": 156,
  "prevPage": 30,
  "nextPage": 32,
  "slideType": "problem_statement",
  "function": "frame_problem",
  "density": "overcrowded",
  "nDataPoints": 0,
  "notes": "The slide uses four examples of AI mislabeling images to demonstrate the lack of 'common sense world models' in AI.",
  "elementsJson": [
    "headline_text",
    "paragraph",
    "screenshot"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5c-f341d4394195/31",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195#slide-31",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "This suggests that having a common sense world model of objects and people is required for an AI system to truly understand what's happening in a visual scene.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a80e-7955bad9077f",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.2,
        "x": 0.51,
        "y": 0.5
      },
      "kind": "image",
      "text": "People with laptop labeled as woman holding teddy bear",
      "attrs": null,
      "subkind": "photo",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "35fe94cd-df8d-4623-9380-d5107103c843",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.2,
        "x": 0.28,
        "y": 0.5
      },
      "kind": "image",
      "text": "Ferret on remote labeled as cat on couch",
      "attrs": null,
      "subkind": "photo",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "5d7306fb-c18c-41bb-9e7b-f2b517de4ef8",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.2,
        "x": 0.06,
        "y": 0.5
      },
      "kind": "image",
      "text": "Baby holding toothbrush labeled as baseball bat",
      "attrs": null,
      "subkind": "photo",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "5d99add9-be2d-463e-9034-e282862b6722",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.2,
        "x": 0.74,
        "y": 0.5
      },
      "kind": "image",
      "text": "Empty road labeled as horse in middle of road",
      "attrs": null,
      "subkind": "photo",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "9935d279-143e-43c1-b6d7-451eb8e1247a",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.2,
        "x": 0.06,
        "y": 0.86
      },
      "kind": "paragraph",
      "text": "a young boy is holding a baseball bat.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "23fdd42e-1e6f-4181-96c6-5ce596feb19d",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.2,
        "x": 0.28,
        "y": 0.86
      },
      "kind": "paragraph",
      "text": "a cat is sitting on a couch with a remote control.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "3517a98f-1a59-4288-88e0-ca73780ffc30",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.2,
        "x": 0.74,
        "y": 0.86
      },
      "kind": "paragraph",
      "text": "a horse is standing in the middle of a road.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "bc6bbfdc-80ed-4dde-bae1-f1f12fa64630",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.2,
        "x": 0.51,
        "y": 0.86
      },
      "kind": "paragraph",
      "text": "a woman holding a teddy bear in front of a mirror.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "bf80e007-a2df-4e48-a9be-3a492053969d",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.2,
        "w": 0.9,
        "x": 0.05,
        "y": 0.22
      },
      "kind": "paragraph",
      "text": "AI models make obvious mistakes when asked to describe a visual scene based on their understanding of objects. Image captioning helps expose the knowledge that computer vision systems learn by training on images labeled with the objects they contain. Such computer vision models make seemingly obvious mistakes when attempting to describe visual scenes. This suggests that having a common sense world model of objects and people is required for an AI system to truly understand what's happening in a visual scene.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "dd1a7963-8cdb-473d-ab42-613010bb25ab",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.9,
        "x": 0.05,
        "y": 0.15
      },
      "kind": "title",
      "text": "However, detecting objects in images is not enough to produce real scene understanding",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "5f8a83d6-b540-44fd-8d3d-84c316439c2a",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Action Titles",
      "slug": "action-titles",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8eca-a05ddd60664e",
      "evidence": "Title 'detecting objects in images is not enough to produce real scene understanding'.",
      "confidence": 88
    },
    {
      "name": "Problem Statement Canvas",
      "slug": "problem-statement-canvas",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "2228f7da-f69c-43ec-8ec7-801aa19269cb",
      "evidence": "The slide clearly articulates a problem, providing context and highlighting the limitations of current solutions, which is a key component of the problem statement canvas.",
      "confidence": 0.8
    }
  ],
  "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": 35,
      "from": 30,
      "name": "Zoom In",
      "slug": "06-zoom-in",
      "bestFor": "Technical deep-dives, case studies, detailed analysis",
      "matchId": "019dd95a-07fe-70ce-8d3c-514142d7981d",
      "evidence": "Big picture (CV detects nouns) -> focus on verbs/common sense -> specific case (multi-viewpoint scenes).",
      "position": 4,
      "objective": "Zoom from object detection to full scene understanding",
      "structure": "The Big Picture -> Key Area of Focus -> Specific Detail -> Implication",
      "confidence": 75,
      "description": "Start broad, then progressively focus on specific details that prove your point"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}