{
  "docId": "019dd923-5e88-73ef-bd5c-fd5012384ae3",
  "docSlug": "bf350dd574c19997",
  "documentTitle": "2021 Air Street Capital The State of AI Report 2021",
  "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": 12,
  "pageCount": 188,
  "prevPage": 11,
  "nextPage": 13,
  "slideType": "industry_trends",
  "function": "summarize",
  "density": "overcrowded",
  "nDataPoints": 8,
  "notes": "Includes a line chart comparing model performance, a table of COCO detection results, and conceptual diagrams for RegNet and SwAV.",
  "elementsJson": [
    "headline_text",
    "action_title",
    "bullet_list",
    "line_chart",
    "data_table",
    "infographic"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5c-fd5012384ae3/12",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5c-fd5012384ae3",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5c-fd5012384ae3.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5c-fd5012384ae3#slide-12",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "Facebook AI introduces SEER, a 1.3B parameter self-supervised model pre-trained on 1B Instagram images that achieves 84.2% top-1 accuracy on ImageNet, comfortably surpassing all existing self-supervised models.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c3-7407-8d24-f597741be63e",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.25,
        "w": 0.15,
        "x": 0.58,
        "y": 0.345
      },
      "kind": "chart",
      "text": "ImageNet Top-1 Acc vs Number of Parameters",
      "attrs": null,
      "subkind": "line",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "3a4d27b1-a3e9-46a3-9f3d-7302cfe2a310",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.2,
        "w": 0.37,
        "x": 0.58,
        "y": 0.67
      },
      "kind": "image",
      "text": "RegNet and SwAV conceptual diagrams",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "e1bad45c-24c6-409a-8b72-e8c5378a2ae9",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.12,
        "w": 0.45,
        "x": 0.04,
        "y": 0.41
      },
      "kind": "list",
      "text": "SEER combines SwAV, a method to learn image embeddings that yields consistent clustering of images with similar visual concepts, with RegNets, a scalable CNN architecture. It uses uncurated and unlabeled (non-EU) Instagram images.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "0b4f0afe-907d-49ff-899b-b10314022b17",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.08,
        "w": 0.45,
        "x": 0.04,
        "y": 0.21
      },
      "kind": "list",
      "text": "Facebook AI introduces SEER, a 1.3B parameter self-supervised model pre-trained on 1B Instagram images that achieves 84.2% top-1 accuracy on ImageNet, comfortably surpassing all existing self-supervised models.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "104fca4f-5960-4740-bcbc-421bf06a893d",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.08,
        "w": 0.45,
        "x": 0.04,
        "y": 0.31
      },
      "kind": "list",
      "text": "Self-supervision has driven NLP research to new heights. Extending this success to computer vision is hard because models need much more data to capture the semantics of a particular visual concept.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "a5d81eff-0040-44d6-8278-97cea4afea65",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.06,
        "w": 0.45,
        "x": 0.04,
        "y": 0.55
      },
      "kind": "list",
      "text": "SEER is a good few-shot learner: it still achieves 77.9% top-1 accuracy on ImageNet when trained with 10% of the dataset.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "ba4c426f-136d-4ff3-95b5-5cae4b94aa10",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.45,
        "x": 0.04,
        "y": 0.63
      },
      "kind": "list",
      "text": "It also outperforms supervised methods on other tasks like object detection and segmentation.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "cbb15dbb-680f-4217-a280-c219d3162049",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "Top-1 accuracy: 84.2%",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-47c3-7407-8d24-fa10f4d44d8c",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.2,
        "x": 0.75,
        "y": 0.345
      },
      "kind": "table",
      "text": "Detection and Segmentation on COCO table",
      "attrs": null,
      "subkind": "data",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "15dea97c-ec14-4946-97a0-b5f7462cbf62",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.45,
        "x": 0.04,
        "y": 0.14
      },
      "kind": "title",
      "text": "Self-supervision is taking over computer vision",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "32eba72c-ec74-49cb-ba55-9c5cf5351024",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Action Titles",
      "slug": "action-titles",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8ed0-bcda6cd19bd6",
      "evidence": "Title 'Self-supervision is taking over computer vision' states the so-what.",
      "confidence": 88
    },
    {
      "name": "Concrete Language",
      "slug": "concrete-language",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8ed0-c3238889a673",
      "evidence": "1.3B params, 1B Instagram images, 84.2% top-1 accuracy.",
      "confidence": 80
    }
  ],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 152,
      "from": 5,
      "beatId": "019dd95a-0682-776c-8e35-0affbadec38e",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Definitions p.5-6, Exec Summary p.7, then Sections 1-3 catalog evidence.",
      "position": 1,
      "confidence": 78,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 152,
      "from": 10,
      "beatId": "019dd95a-0682-776c-8e35-1b3ee3a1c012",
      "arcName": "The Mountain",
      "arcSlug": "mountain",
      "beatName": "Rising Action",
      "beatSlug": "mountain-rising-action",
      "evidence": "Three sections accumulating evidence of accelerating progress.",
      "position": 2,
      "confidence": 55,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [
    {
      "to": 18,
      "from": 11,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3d-45052222b756",
      "evidence": "Multiple evidence slides (vision, audio, 3D, MLPs comeback) all converging on transformer dominance.",
      "position": 2,
      "objective": "Show transformers spreading across AI domains",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 82,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}