Temporal Knowledge Graph Patents: Who Leads, Where the Gaps Are 2026
Filing growth compares 2021 (2 records) with 2024 (5) — a three-year span. 2024 is the most recent year we treat as complete: publication lags filing by roughly 18 months, so 2025 onwards are still filling in and any growth rate that ends there would understate the field. Top-5 share is the combined record count of the five largest assignees divided by all 47 records in scope (CR5), not by the ranked leaders only.
What the temporal knowledge graph patent record shows
Temporal knowledge graphs extend standard knowledge graph structures with timestamps, letting a system represent not just that two entities are related but when that relation held or is predicted to hold. The patent record in scope — 47 published families spanning 2015 through mid-2026 — covers everything from recommendation engines that reason over time-dependent user states to AI agent platforms that use graph structures for decision-making. Filing is concentrated: one assignee accounts for 14 of the 47 records, and the top 5 combined hold 74.5% of the field. That leaves a long tail of single- or double-filing entrants working narrower angles.
Most activity sits inside general-purpose data processing and AI computing classes rather than a single vertical, which suggests the core technique is still being fitted to specific use cases rather than standardized around one. The representative filing from NEC Laboratories Europe — a recommender system that simulates future knowledge graph states from an initial timestamped graph — is a useful anchor for what a mid-complexity claim in this space looks like.
Filing trends and technology composition
The dataset spans 47 published records across 18 ranked assignees, with filings concentrated in general data-processing and AI-computing classes. Because publication lags filing by roughly 18 months, the most recent years in the trend below are undercounted and should not be read as a slowdown.
A rising filing curve through a 2022 peak
Filings ran from 8 records in 2017 to a peak of 14 in 2022, and grew 150% from 2021 (2) to 2024 (5) — the last year in this trend that can be treated as complete. 2025 and 2026 figures will continue to fill in as publications catch up with filing dates.
Data processing and AI computing dominate the class mix
G06F (electric digital data processing) appears in 72.3% of the 47 records and G06N (AI-based computing) in 46.8%, confirming that most temporal knowledge graph work is filed as general computing infrastructure rather than a narrow application. Smaller shares in traffic control (G08G, 2.1%) and healthcare informatics (G16H, 2.1%) mark branches with very little claim density so far.
Shares are the percentage of the 47 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Go deeper on Knowledge Graphs & Graph Computing: Temporal Knowledge Graphs Patent Landscape with Eureka
This page is one run against one query. Ask Eureka your own question about knowledge graphs & graph computing: temporal knowledge graphs patent landscape and every answer comes back with the patent numbers behind it.
Try EurekaThe records shaping this landscape
Recommender system for time-dependent actions based on temporal knowledge graphs
Filed by NEC Laboratories Europe, this record describes a method that organizes domain information into an initial temporal knowledge graph, predicts future entities, links and attributes for later points in time, and simulates the outcomes of candidate actions by generating predicted future graphs. It is a useful anchor for the field because it combines graph construction, temporal prediction and action simulation in a single claim structure.Filed 2024-03-21 · NEC Laboratories Europe GmbH
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20250259041A1 | Ai agent decision platform with deontic reasoning | 63 |
| 2 | US20250259082A1 | Ai agent decision platform with deontic reasoning and quantum-inspired token management | 51 |
| 3 | US12254420B1 | Point-of-interest recommendation method based on temporal knowledge graph | 34 |
| 4 | US20240184829A1 | Methods and systems for controlled modeling and optimization of a natural language database interface | 24 |
| 5 | CN116010722A | 一种基于网格时空知识图谱的动态多目标时空问题的查询方法 | 19 |
| 6 | US20240378247A1 | Analyzing social media data to identify markers of coordinated movements, using stance detection, and using c… | 16 |
| 7 | WO2023034358A2 | Analyzing social media data to identify markers of coordinated movements, using stance detection, and using c… | 14 |
| 8 | US20240095805A1 | Recommender system for time-dependent actions based on temporal knowledge graphs | 12 |
| 9 | WO2022174356A1 | Methods and systems for controlled modeling and optimization of a natural language database interface | 11 |
| 10 | US20180032506A1 | Systems and methods for disambiguating a term based on static and temporal knowledge graphs | 10 |
Citation counts reflect influence inside this searched corpus and favor older records; they are not a measure of current commercial importance.
Patent titles are shown in the language they were filed in, not translated, so that each record stays verifiable against the original filing — a translated title will not match in Eureka or in any national register. Each row carries its publication number; clicking a row searches Eureka by that number.
Put your own technology through the same analysis
Eureka on the web
When you want the answer in the next five minutes.
The agent works the prompt against patents and technical literature, citing every source.
Run your analysis now →MCP server & REST API
When it has to run inside your own pipeline.
Patent search, landscape analysis and assignee resolution as MCP tools. Drop them into any agent framework, or call REST directly.
Browse MCP servers →What the concentration and class data mean for filing decisions
Three patterns stand out once the ranking, trend and IPC composition are read together: a dominant filer, a fast recent filing rise, and application branches that remain almost untouched.
One leader, then a steep drop-off
The leading assignee alone holds 14 of the 47 records in scope, and the top 5 combined reach 74.5%. Anyone filing new claims in the core recommendation or reasoning space should expect to be reviewed against that leader's portfolio first.
Growth is real but recent years are undercounted
Filings rose from 2 records in 2021 to 5 in 2024, a 150% increase over that span, with a peak of 14 in 2022. Because publication lags filing by about 18 months, 2025 and 2026 counts will rise as more applications publish.
Application verticals are barely claimed
G06F and G06N cover the technical core, but downstream applications like traffic systems and healthcare informatics each show only a single record. That is a narrow but real opening for domain-specific claims that combine temporal graph reasoning with a regulated vertical.
US and PCT routes dominate filing strategy
The United States receives the largest single share of filings at 17, with WIPO (PCT) filings at 8 and China and Europe each at 6. That distribution points to applicants prioritizing US enforceability first and using PCT to keep international options open.
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to knowledge graphs & graph computing: temporal knowledge graphs patent landscape, with the prior art for and against each one.
Where to take this analysis
The dataset points to a field with one dominant filer and several open branches. The next step is deciding whether to file adjacent to the leader or into the white space.
Map claims against the leading assignee
Before drafting in the core reasoning or recommendation space, run a freedom-to-operate check against the leading assignee's 14 records to see exactly where their claim boundaries sit.
Explore claim scope in EurekaTest a vertical-specific claim
Traffic control and healthcare informatics each carry only one record. Drafting a claim that ties temporal graph reasoning to a specific regulated workflow in either vertical starts from an almost-empty field.
Draft a white-space claim in EurekaTrack the 2025–2026 filing catch-up
Publication lag means recent years look artificially low. Set up ongoing monitoring so new publications in this space surface as they clear the 18-month lag rather than being discovered late.
Set up monitoring in EurekaCommon questions about temporal knowledge graph patents
The assignee ranking covers 18 companies across 47 records, and it is concentrated rather than evenly spread: one assignee holds 14 records, and the top 5 combined account for 74.5% of all filings in scope. That leaves a long tail of assignees with one or two records each, typically working narrower application angles such as recommendation or specific industry use cases. Because this ranking reflects only the records returned by this search, it should be read as a snapshot of the searched corpus rather than a claim about every filer in the field.
Filings rose from 2 records in 2021 to 5 in 2024, a 150% increase over that three-year span, with a peak of 14 records in the 2022 filing year. Publication typically lags filing by around 18 months, so the 2025 and 2026 figures in any trend chart are still incomplete and will rise as more applications publish. Treat the recent uptick as evidence of sustained activity rather than a peak that has already passed.
The bulk of activity sits in G06F (electric digital data processing), which appears in 72.3% of the 47 records in scope, and G06N (AI-based computing), at 46.8%. Business and administrative data processing (G06Q) appears in 21.3% of records, while digital transmission (H04L), robotics (B25J), traffic control (G08G) and healthcare informatics (G16H) each cover a much smaller share. Because a single record can carry multiple IPC classes, these shares add up to more than 100% of the record total and should not be summed.
Traffic control systems and healthcare informatics each show only one record among the 47 in scope, a 2.1% share, despite both being natural fits for temporal graph reasoning about evolving states over time. Robotics and manipulator applications (B25J) are similarly thin at 4.3%. These branches are not proven markets, but they are largely unclaimed in this dataset, which makes them worth checking against a fuller prior-art search before drafting.
The United States receives the largest share of filings among the offices tracked, with 17 records, followed by WIPO PCT filings at 8, and China and Europe each at 6. This pattern suggests most applicants in this dataset are anchoring protection in the US market and using PCT filings to preserve the option of expanding internationally later. Applicants targeting China or Europe specifically should expect a smaller but still meaningful set of prior filings to search against.
Research Knowledge Graphs & Graph Computing: Temporal Knowledge Graphs Patent Landscape in depth with Eureka
Go past this page: query the whole knowledge graphs & graph computing: temporal knowledge graphs patent landscape corpus yourself, in your own scope.
Every answer comes back with patent numbers you can open.
Disclaimer. This page is generated from Patsnap Eureka data drawn from a limited snapshot of global patent and scientific-literature records, and is provided for general information and reference only.
Patent data carries inherent limitations: recent filings (typically the most recent 18–24 months) are under-counted due to standard publication lag; counts may be reported at either a patent-family or a patent-record basis and are not always directly comparable; classification, applicant-name, and citation data may contain errors, duplicates, or omissions; and the underlying search query defines and constrains the scope shown. As a result, the analysis may be incomplete or inaccurate and may not reflect the full technology landscape.
Nothing on this page constitutes an exhaustive prior-art, novelty, freedom-to-operate, or validity search, nor does it constitute legal, financial, investment, or professional advice, and it should not be relied upon as such. Any patent, commercial, or strategic decision should be verified independently and reviewed with qualified patent, legal, and domain professionals. Patsnap makes no warranties, express or implied, as to the accuracy, completeness, or fitness for any particular purpose of the information presented.
Machine translation. Assignee and organisation names originally recorded in Chinese, Japanese or Korean have been rendered into English by an AI translation step so that the tables stay readable. These renderings are best-effort and may not match a company’s registered English name; the original name is what the underlying patent record carries, and it is what any Eureka query launched from this page uses.