{
  "docId": "019dd923-5e88-73ef-bd5c-f812573a947a",
  "docSlug": "eea7524c557036f4",
  "documentTitle": "2020 Air Street Capital The State of AI Report 2020",
  "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": 61,
  "pageCount": 177,
  "prevPage": 60,
  "nextPage": 62,
  "slideType": "industry_trends",
  "function": "summarize",
  "density": "dense",
  "nDataPoints": 4,
  "notes": "The slide compares CNNs and TICK-GP models on digit recognition and compares SVGP vs VISH on flight delay prediction.",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "other",
    "bar_chart_grouped"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/61",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-61",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "GP training time is reduced from 15 mins to just 40 sec (Fig B) when predicting delay of commercial flights with a dataset of 6 million data points.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c5-73ac-aa65-89f44346e83c",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.25,
        "w": 0.3,
        "x": 0.65,
        "y": 0.65
      },
      "kind": "chart",
      "text": "Fig B: Comparison of SVGP and VISH training time and NLPD error.",
      "attrs": null,
      "subkind": "bar-grouped",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "06f32cf4-5f12-4c06-af04-f0f046cd0471",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.25,
        "w": 0.3,
        "x": 0.65,
        "y": 0.33
      },
      "kind": "chart",
      "text": "Fig A: Comparison of CNN and TICK-GP uncertainty calibration on digit recognition.",
      "attrs": null,
      "subkind": "bar-grouped",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "3fd8a9ff-cd19-48c1-bb28-8053a4130765",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.55,
        "w": 0.55,
        "x": 0.02,
        "y": 0.33
      },
      "kind": "list",
      "text": "GP models benefit from several key features, like depth and convolutions, inspired from neural networks (NNs).\nHowever, GPs have better calibrated uncertainty compared to NNs (Fig A), which here are shown to be confidently wrong more often (attributing no mass to the true label - shown in blue).\nGP training time is reduced from 15 mins to just 40 sec (Fig B) when predicting delay of commercial flights with a dataset of 6 million data points.\nThis GP method circumvents the need to invert large matrices, significantly speeding up training time and enabling a quicker response and adaptation to emerging events.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "c7b61347-d0e6-4af6-b0ed-f6071fc0bd37",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "Training time: 40 sec",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-47c5-73ac-aa65-8eb118a6ddb6",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.08,
        "w": 0.95,
        "x": 0.02,
        "y": 0.22
      },
      "kind": "paragraph",
      "text": "GPs are becoming more accurate and faster to train, whilst retaining their favourable properties, like calibrated uncertainty, making them more relevant for real-world applications today.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "3889c934-ca8c-45d3-a1d2-a2ebdfacce03",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.8,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "Gaussian Processes (GPs) Strike Back: Quantified uncertainty and faster training speed",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "93549c4d-020a-4d8c-96a1-5f8341017157",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Audience Definition",
      "slug": "audience-definition",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "1ecdc9fb-dc1e-4b11-afb3-46083244e06b",
      "evidence": "GP models benefit from several key features, like depth and convolutions, inspired from neural networks (NNs).",
      "confidence": 0.7
    },
    {
      "name": "Concrete Language",
      "slug": "concrete-language",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8ece-f890e57562dd",
      "evidence": "15 mins -> 40 sec specific speedup.",
      "confidence": 75
    },
    {
      "name": "Data Story Arc",
      "slug": "data-story-arc",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "31b8520a-0f4f-498a-8c41-441ec53a5022",
      "evidence": "GP training time is reduced from 15 mins to just 40 sec (Fig B) when predicting delay of commercial flights with a dataset of 6 million data points.",
      "confidence": 0.7
    }
  ],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 129,
      "from": 10,
      "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.",
      "position": 1,
      "confidence": 60,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 129,
      "from": 10,
      "beatId": "019dd95a-0682-776c-8e34-fc384f439bba",
      "arcName": "The Mountain",
      "arcSlug": "mountain",
      "beatName": "Rising Action",
      "beatSlug": "mountain-rising-action",
      "evidence": "Research, Talent and Industry sections build momentum of AI progress.",
      "position": 2,
      "confidence": 40,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [
    {
      "to": 62,
      "from": 58,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3c-fd470c64e823",
      "evidence": "Three short bursts of new research direction case studies.",
      "position": 6,
      "objective": "Map emerging research themes: federated learning, GPs, quantum",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 65,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}