{
  "docId": "019dd923-5e88-73ef-bd5d-01cf0d8a8fbc",
  "docSlug": "cc8183ec02431b7a",
  "documentTitle": "2022 Air Street Capital The State of AI Report 2022",
  "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": 40,
  "pageCount": 114,
  "prevPage": 39,
  "nextPage": 41,
  "slideType": "industry_trends",
  "function": "summarize",
  "density": "dense",
  "nDataPoints": 1,
  "notes": "The slide discusses the evolution of transformer models from task-specific to generalist architectures.",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "paragraph",
    "process_diagram"
  ],
  "metadataConfidence": 0.95,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc/40",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc#slide-40",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "DeepMind's Gato brings this effort to another level: researchers train a 1.2B parameter transformer to perform hundreds of tasks in robotics, simulated environments, and vision and language.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a217-aa1d31ab04f7",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.3,
        "x": 0.65,
        "y": 0.35
      },
      "kind": "image",
      "text": "Diagram of the Gato model architecture showing inputs (images, text, proprioception) and outputs (actions, text).",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "7734fa0e-bfb1-4bbd-8273-1f61d432972b",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.5,
        "w": 0.6,
        "x": 0.05,
        "y": 0.38
      },
      "kind": "list",
      "text": "Attempts at generalist multitask, multimodal models date back to at least Google’s “One model to learn them all” (2017). DeepMind’s Gato brings this effort to another level: researchers train a 1.2B parameter transformer to perform hundreds of tasks. This partially proves our 2021 Prediction.\nThey showed that scaling consistently improved the model, but it was kept “small” for live low-latency robotics tasks.\nTo train their model on different modalities, all data was serialized into a sequence of tokens which are embedded in a learned vector space. The model is trained in a fully supervised fashion.\nSeparately: With data2vec, on a narrower set of tasks, Meta devised a unified self-supervision strategy across modalities. But for now, different transformers are used for each modality.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "9326d1f1-2039-4dda-b060-beb581af13f3",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.6,
        "x": 0.05,
        "y": 0.2
      },
      "kind": "paragraph",
      "text": "Transformers trained on a specific task (via supervised or self-supervised learning) can be used for a broader set of tasks via fine-tuning. Recent works show that a single transformer can be directly and efficiently trained on various tasks across different modalities (multi-task multimodal learning).",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "46aff0af-8c44-4b13-9fd5-5caa57bf3697",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.7,
        "x": 0.05,
        "y": 0.12
      },
      "kind": "title",
      "text": "2021 Prediction:....culminating in a single transformer to rule them all?",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "c0f9a0f4-b90a-4200-b134-63cac7dccfc0",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Action Titles",
      "slug": "action-titles",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "019dd95a-1055-74e0-a956-1afa30b1521b",
      "evidence": "Title 'a single transformer to rule them all?' states unifying thesis.",
      "confidence": 80
    }
  ],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 80,
      "from": 5,
      "beatId": "019dd95a-0682-776c-8e35-2557e3799e96",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Definitions, exec summary, Research and Industry sections inventory the state of AI.",
      "position": 1,
      "confidence": 78,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 93,
      "from": 11,
      "beatId": "019dd95a-0682-776c-8e35-35abf646aa8b",
      "arcName": "The Mountain",
      "arcSlug": "mountain",
      "beatName": "Rising Action",
      "beatSlug": "mountain-rising-action",
      "evidence": "Research, Industry, Politics sections accumulate signals.",
      "position": 2,
      "confidence": 55,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [
    {
      "to": 43,
      "from": 37,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3d-afe811e71ad6",
      "evidence": "SayCan, ConvNets vs ViT, vision-language unification, Gato, IRIS, NeRFs as parallel evidence.",
      "position": 6,
      "objective": "Show transformers becoming the universal cross-modality architecture",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 78,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}