{
  "docId": "019dd923-5e88-73ef-bd5c-f6fdd8f57895",
  "docSlug": "3a892393f72c1ff5",
  "documentTitle": "2019 Air Street Capital The State of AI Report 2019",
  "authorId": "AirStreetCapital",
  "authorName": "Nathan Benaich and Ian Hogarth",
  "documentKindSlug": "consulting-deck",
  "documentKindLabel": "Consulting deck",
  "sourceTypeSlug": "vc_research",
  "sourceTypeLabel": "VC research",
  "presentationDate": null,
  "orientation": "landscape",
  "aspectRatio": 1.777,
  "pageNumber": 19,
  "pageCount": 136,
  "prevPage": 18,
  "nextPage": 20,
  "slideType": "client_example",
  "function": "illustrate_case",
  "density": "balanced",
  "nDataPoints": 0,
  "notes": "The slide describes the technical architecture and workflow of the Horizon platform.",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "process_diagram"
  ],
  "metadataConfidence": 0.95,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5c-f6fdd8f57895/19",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5c-f6fdd8f57895",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5c-f6fdd8f57895.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5c-f6fdd8f57895#slide-19",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "Facebook release Horizon, the first open source end-to-end platform that uses applied RL to optimize systems in large-scale production environments, such as Messenger suggestions, video stream quality and notifications.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ab-72d9-bbc2-426400d1338e",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.22,
        "x": 0.74,
        "y": 0.35
      },
      "kind": "diagram",
      "text": "Workflow diagram showing Spark, Scipy, PyTorch 1.0, OpenAI Gym, and ONNX components.",
      "attrs": null,
      "subkind": "process",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "53b4fbd4-8802-4599-8ce9-b5551deba44e",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.65,
        "w": 0.68,
        "x": 0.02,
        "y": 0.22
      },
      "kind": "list",
      "text": "Facebook release Horizon, the first open source end-to-end platform that uses applied RL to optimize systems in large-scale production environments, such as Messenger suggestions, video stream quality and notifications.\nHorizon is built on PyTorch 1.0, Caffe2 and Spark, popular tools for ML work.\nIn particular, the system includes workflows for simulated environments as well as a distributed platform for preprocessing, training, and exporting models into production.\nIt focuses on ML-based systems that optimise a set of actions given the state of an agent and its environment (\"policy optimisation\"). The optimisation relies on data that's inherently noisy, sparse, and arbitrarily distributed.\nInstead of online training as in games, Horizon models are trained offline using a policy that a product engineer has designed. Counterfactual policy evaluation (CPE) is used to estimate what the RL model would have done if it were making those past decisions. Once the CPE results are admissible, the RL model is deployed in a small experiment to collect live results.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "509cc9dd-8e57-4a20-80fa-1b07c85cc4cb",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.6,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "What's next in RL: Moving research into production environments",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "14b3cdb1-2be9-433c-afe5-41e524e4b775",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 92,
      "from": 5,
      "beatId": "019dd95a-0682-776c-8e34-ce57e5bbe574",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Definitions through Research/Talent/Industry — fact-dense case studies and benchmarks.",
      "position": 1,
      "confidence": 70,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 92,
      "from": 9,
      "beatId": "019dd95a-0682-776c-8e34-ddbccb391af1",
      "arcName": "Voyage and Return",
      "arcSlug": "voyage-return",
      "beatName": "The Unknown",
      "beatSlug": "voyage-return-the-unknown",
      "evidence": "Research/Talent/Industry frontiers explored.",
      "position": 2,
      "confidence": 45,
      "parentBeatName": "Complication",
      "parentBeatSlug": "complication"
    }
  ],
  "loops": [
    {
      "to": 19,
      "from": 16,
      "name": "Zoom In",
      "slug": "06-zoom-in",
      "bestFor": "Technical deep-dives, case studies, detailed analysis",
      "matchId": "019dd95a-07fe-70ce-8d3c-9eda9346282a",
      "evidence": "Series titled 'What's next in RL: ...' each zooming on a sub-area.",
      "position": 3,
      "objective": "Drill into specific RL research frontiers (sim-to-real, planning, production)",
      "structure": "The Big Picture -> Key Area of Focus -> Specific Detail -> Implication",
      "confidence": 65,
      "description": "Start broad, then progressively focus on specific details that prove your point"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}