{
  "patent_number": "US 20250307927",
  "country": "US",
  "title": "How AI Analyzes Financial Data Across Different Timeframes",
  "original_title": "Multi-Scale Temporal Attention Processing System for Multimodal Deep Learning with Vector-Quantized Variational Autoencoder",
  "summary": "This patent describes an AI system that processes various types of financial data, like stock prices and news, by looking at it simultaneously through quarterly, weekly, and daily lenses, and then combining these views, especially adapting to market changes.",
  "what_it_does": "The system takes in different kinds of information, called \"multimodal data,\" such as time-series data (like stock prices), text (news articles), sentiment (public mood), and structured tables (financial reports). It then splits this data into \"exactly three\" separate processing streams: one for quarterly patterns, one for weekly cycles, and one for real-time daily activity, as stated in claim 1. Each stream uses a \"scale-specific attention mechanism\" optimized for its timeframe. Crucially, these streams talk to each other through \"bidirectional cross-temporal gradient flow,\" meaning changes in understanding at one level affect the others. The system also \"dynamically weights\" how much each timeframe contributes based on \"real-time market volatility indicators,\" as described in claim 1, giving more importance to short-term data during choppy markets (claim 3). Finally, it combines these insights into a single, unified data picture. For example, it could analyze a company's quarterly earnings report, weekly stock performance, and intraday trading volume, adjusting its focus if the market suddenly becomes turbulent.",
  "what_it_does_not_cover": [
    "Processing data using fewer or more than exactly three temporal streams.",
    "Systems that do not use a hierarchical organization for their temporal processing streams.",
    "AI models where attention weight adjustments at one temporal scale do not automatically influence computations at the other two scales (no bidirectional cross-temporal gradient flow).",
    "Systems that do not dynamically adjust the weighting of temporal levels based on real-time market volatility indicators.",
    "Analyzing data modalities other than time-series, textual, sentiment, and structured tabular data as the core input."
  ],
  "filed": "2025-06-10",
  "granted": null,
  "expires": "2045-06-10",
  "status": "active",
  "holder": "Atombeam Technologies",
  "holder_url": "https://patentbrief.org/company/atombeam-technologies",
  "inventors": [
    {
      "name": "Brian Galvin",
      "url": "https://patentbrief.org/inventor/brian-galvin"
    }
  ],
  "times_cited": 18,
  "tags": [
    "financial_services",
    "ai_ml",
    "software"
  ],
  "abstract": "A system and method for multi-scale temporal attention processing in multimodal technology deep learning systems. This system processes time-series, textual, sentiment, and structured tabular data across three hierarchically-organized temporal streams—quarterly, weekly, and intraday levels—with bidirectional cross-temporal information flow. Scale-specific attention mechanisms are optimized for respective temporal granularities, while an adaptive controller dynamically weights each temporal level based on real-time market volatility indicators. A multi-scale fusion processor integrates attention-weighted representations to generate temporally unified representations preserving both short-term market dynamics and long-term trends. This approach enables superior forecasting and risk assessment by leveraging temporal correlations across multiple time scales while automatically adapting to changing market conditions. The system facilitates interpretable AI analysis through attention visualization and enables synthetic scenario generation for model testing.",
  "url": "https://patentbrief.org/patent/us/20250307927/multi-scale-temporal-attention-processing-system-for-multimodal-deep-learning-wi",
  "markdown_url": "https://patentbrief.org/patent/us/20250307927/multi-scale-temporal-attention-processing-system-for-multimodal-deep-learning-wi/md",
  "google_patents_url": "https://patents.google.com/patent/US20250307927",
  "relatedPatents": [
    {
      "patentNumber": "12327190",
      "countryCode": "US",
      "title": "AI System for Analyzing Mixed Financial Information",
      "url": "https://patentbrief.org/patent/us/12327190/multimodal-financial-technology-deep-learning-core-with-joint-optimization-of-ve"
    },
    {
      "patentNumber": "12307349",
      "countryCode": "US",
      "title": "How an AI System Answers Financial Questions Using Specialized Bots",
      "url": "https://patentbrief.org/patent/us/12307349/systems-and-methods-of-large-language-model-driven-orchestration-of-task-specifi"
    },
    {
      "patentNumber": "11922469",
      "countryCode": "US",
      "title": "How Computers Rank Financial News for Analysts",
      "url": "https://patentbrief.org/patent/us/11922469/automated-news-ranking-and-recommendation-system"
    },
    {
      "patentNumber": "11170293",
      "countryCode": "US",
      "title": "How AI Systems Learn to Predict and Act Simultaneously",
      "url": "https://patentbrief.org/patent/us/11170293/alphago-policy-and-value-networks"
    },
    {
      "patentNumber": "20240338414",
      "countryCode": "US",
      "title": "How AI Understands Answers Across Multiple Documents",
      "url": "https://patentbrief.org/patent/us/20240338414/inter-document-attention-mechanism"
    }
  ]
}