{
  "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": 33,
  "pageCount": 136,
  "prevPage": 32,
  "nextPage": 34,
  "slideType": "case_study",
  "function": "illustrate_case",
  "density": "overcrowded",
  "nDataPoints": 18,
  "notes": "The slide highlights the application of neural networks and Monte Carlo tree search in chemistry.",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "paragraph",
    "line_chart",
    "infographic"
  ],
  "metadataConfidence": 0.95,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5c-f6fdd8f57895/33",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5c-f6fdd8f57895",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5c-f6fdd8f57895.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5c-f6fdd8f57895#slide-33",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "This method is far faster that the state-of-the-art computer-assisted synthesis planning.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ab-72d9-bbc4-4ef5bd9b685d",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.4,
        "w": 0.3,
        "x": 0.65,
        "y": 0.25
      },
      "kind": "chart",
      "text": "Solved ratio vs Time limit (s) for 3N-MCTS, Neural BFS, and Heuristic BFS",
      "attrs": null,
      "subkind": "line",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "36275c4b-b261-4c6e-b500-a0bdb5fbb0bb",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.2,
        "x": 0.45,
        "y": 0.28
      },
      "kind": "image",
      "text": "Chemical reaction pathway diagrams",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "cc67ab4d-100f-41a4-bb36-1bb11d1655bf",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.3,
        "w": 0.45,
        "x": 0.03,
        "y": 0.28
      },
      "kind": "list",
      "text": "A system built from three NNs (3N-MCTS): 1) Guide the search in promising directions; 2) Predict feasibility; 3) Estimate position value and iterate.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "edb993b6-5b9d-44e9-8a78-3b9e2701674e",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "Solved ratio: 80%",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-41ab-72d9-bbc4-5111bc9ae62d",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.95,
        "x": 0.03,
        "y": 0.68
      },
      "kind": "paragraph",
      "text": "This method is far faster that the state-of-the-art computer-assisted synthesis planning. In fact, 3N-MCTS solves more than 80% of a molecule test set with a time limit of 5 seconds per target molecule.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "8bbd2fcf-6009-4ff4-a4d5-d7124e243ad7",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.95,
        "x": 0.03,
        "y": 0.22
      },
      "kind": "subtitle",
      "text": "Using neural networks with Monte Carlo tree search to solve retrosynthesis by training on 12.4 million reactions.",
      "attrs": null,
      "subkind": null,
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "b195ee5d-7395-4a12-ba3a-e8333f7d0986",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.5,
        "x": 0.03,
        "y": 0.14
      },
      "kind": "title",
      "text": "Machines learn how to synthesise chemical molecules",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "de0fd9f2-b224-4155-b259-b0c73691ed1d",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Action Titles",
      "slug": "action-titles",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8ecc-f15924ec6716",
      "evidence": "Title positions ML as faster-than-SOTA chemistry planner.",
      "confidence": 75
    }
  ],
  "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": 33,
      "from": 28,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3c-a49628ab8218",
      "evidence": "Six contiguous medical AI case studies with quantitative outcomes (0.97 AUC, 73%, 80%).",
      "position": 5,
      "objective": "Compile clinical-grade ML wins across imaging, cardiology, neuro and chemistry",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 75,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}