Book a demo

Cut patent&paper research from weeks to hours with PatSnap Eureka AI!

Try now

Andrew Ng Patents & Innovation Profile — PatSnap Eureka

Andrew Ng Patents & Innovation Profile — PatSnap Eureka
Inventor Profile · PatSnap Eureka

Andrew Y. Ng: Patent Portfolio & Innovation Analysis

Andrew Y. Ng is one of the most influential figures in artificial intelligence, holding 77 curated patents across deep learning, speech recognition, and computer vision, with filings spanning 1995 to 2025. His portfolio is primarily assigned to LandingAI Inc. and covers foundational methods in neural network training, end-to-end speech recognition, and industrial visual inspection — with a broader dataset of 1,714 records attributed to variants of his name across global patent databases.

77
Curated Patents
1995–2025
Years Active
8
Jurisdictions

Patent Filing Activity

Peak filing year was 2023 with 10 filings; the 2019–2025 LandingAI period is the most prolific phase of the portfolio.

Annual Patent Filings by Andrew Y. Ng: 1995=1, 1996=4, 2003=2, 2004=5, 2006=1, 2007=2, 2009=2, 2012=4, 2014=4, 2015=8, 2018=3, 2019=9, 2020=5, 2021=5, 2022=5, 2023=10, 2024=3, 2025=7 Line chart showing Andrew Y. Ng's patent filing activity by year, derived from PatSnap Eureka patent database. Peak year was 2023 with 10 filings. 10 7 5 2 0 1995 2004 2012 2015 2019 2023 2025
📋
77
Curated Patents
40 active · 13 pending · 18 inactive
📅
1995–2025
Filing Period
30 years of innovation activity; peak in 2023
🌐
8
Jurisdictions
US, WO, EP, KR, AU, CN, IN, CA
🏢
LandingAI
Primary Assignee
17 patents assigned
🔬
G06N3
Top Technology
8 patents in neural networks & deep learning
Patent Analytics

Andrew Y. Ng's Patent Filing Patterns

Three distinct career phases are visible in the filing data: an early sparse period, a Baidu-driven acceleration peaking in 2015, and a prolific LandingAI era peaking in 2023 with 10 filings.

Annual Patent Filings

Peak year 2023 with 10 filings; 2019 marked the LandingAI founding surge with 9 filings. 7 filings already recorded in 2025.

Annual Patent Filings by Andrew Y. Ng: 1995=1, 1996=4, 2003=2, 2004=5, 2006=1, 2007=2, 2009=2, 2012=4, 2014=4, 2015=8, 2018=3, 2019=9, 2020=5, 2021=5, 2022=5, 2023=10, 2024=3, 2025=7 Line chart showing Andrew Y. Ng's patent filing activity by year, derived from PatSnap Eureka patent database. Peak year was 2023 with 10 filings. 10 7 5 2 0 1995 2004 2012 2015 2019 2023 2025

Technology Domain Breakdown

Neural networks (G06N3) and speech recognition (G10L15) together represent 48% of the curated portfolio, with computer vision and MLOps each at 18%.

Technology Domain Breakdown for Andrew Y. Ng: G06N3 Neural Networks=24%, G10L15 Speech Recognition=24%, G06V10 Computer Vision=18%, G06N20 ML Systems=18%, G06F11 Distributed Computing=15% Donut chart showing the distribution of Andrew Y. Ng's patents across technology domains based on IPC classification codes from PatSnap Eureka. 33 top-domain patents G06N3 Neural Networks (24%) G10L15 Speech Recognition (24%) G06V10 Computer Vision (18%) G06N20 ML Systems (18%) G06F11 Distributed Comp. (15%)

Search Andrew Y. Ng's full patent landscape in PatSnap Eureka IP

Explore in PatSnap Eureka IP →
Technology Domains

Andrew Y. Ng's Core Areas of Innovation

Andrew Y. Ng's patent portfolio spans five technology domains, tracing a career arc from enterprise distributed systems through speech AI and deep learning infrastructure to industrial computer vision and data-centric MLOps.

Neural Networks & Deep Learning Infrastructure

8 patents

This is the portfolio's core technical domain, covering architecture and training of deep neural networks. Patents address real-time neural text-to-speech, adaptive training dataset refinement through deployment feedback, and guided deep learning error analysis workflows — perennial challenges at the heart of practical AI deployment.

  • Systems and methods for real-time neural text-to-speech (US20180247636A1)
  • Model Management System for Improving Training Data Through Machine Learning Deployment (US20220300855A1)
  • Guided workflow for deep learning error analysis (US12333792B2)
IPC: G06N3

Speech Recognition & Transcription

8 patents

A closely clustered group of patents arising from Andrew Ng's leadership of Baidu's AI research division covers end-to-end deep learning approaches to speech recognition, anchored by the Deep Speech system. These filings replaced traditional pipeline-based ASR architectures with recurrent neural networks trained on large datasets, demonstrating that simpler models trained with sufficient data outperform complex hand-engineered systems.

  • Systems and methods for speech transcription (US20160171974A1)
  • Deep learning models for speech recognition (US11562733B2)
  • Systems and methods for speech transcription — EP grant (EP3180785B1)
IPC: G10L15

Computer Vision & Visual Inspection

6 patents

Patents in this domain reflect LandingAI Inc.'s commercial focus: applying machine learning to manufacturing quality control and agricultural computer vision. Inventions address defect detection, impurity identification, partial labelling for efficient training data creation, and data-centric mislabel detection — extending deep learning theory into industrial hardware systems.

  • AI-optimized harvester configured to maximize yield and minimize impurities (US11412657B2)
  • Partial labeling mechanism for quick and accurate training of machine learning models (US12579793B1)
  • Data centric mislabel detection (US12482242B1)
IPC: G06V10

Machine Learning Systems & MLOps

6 patents

A cluster of patents addresses operational and lifecycle challenges of deploying machine learning models in real environments, covering user-generated visual guides for training data classification, integrated ML and rules platforms for improved accuracy, and radiology image classification from noisy images. These inventions collectively represent the emerging "data-centric AI" philosophy.

  • User-generated visual guide for the classification of images (US11182646B2)
  • Systems and methods for radiology image classification from noisy images (US11798159B2)
IPC: G06N20

Distributed Computation & Task Management

5 patents

An earlier cluster, dating from the early 2000s and assigned to SAP Aktiengesellschaft, covers distributed systems for task management and fault recovery in data processing environments. These patents predate Andrew Ng's primary AI research career and reflect earlier work in enterprise systems engineering. The key patent has accumulated 82 citations, indicating sustained relevance to the systems management field.

  • Managing tasks in a data processing environment (US20040226013A1)
IPC: G06F11
Most Cited Patents

Andrew Y. Ng's Highest-Impact IP

The most cited patent — Baidu's speech transcription filing — has accumulated 159 citations, reflecting the Deep Speech programme's documented impact on the ASR field and subsequent voice interface development.

Patent Number Title Year Citations Assignee Status
US20160171974A1 Systems and methods for speech transcription 2015 159 ↑ BAIDU USA LLC active
US7801645B2 Robotic vacuum cleaner with edge and object detection system 2004 153 ↑ SHARPER IMAGE ACQUISITION LLC inactive
US20200211154A1 Method and system for privacy-preserving fall detection 2019 94 ↑ ALTUMVIEW SYSTEMS INC. active
US20040226013A1 Managing tasks in a data processing environment 2003 82 ↑ SAP AKTIENGESELLSCHAFT active
US20080137989A1 Arrangement and method for three-dimensional depth image construction 2007 74 ↑ SAXENA, ASHUTOSH inactive
US20140304335A1 Systems and methods for interactive experiences and controllers therefor 2014 65 ↑ TIMEPLAY ENTERTAINMENT CORPORATION inactive
US20180247636A1 Systems and methods for real-time neural text-to-speech 2018 64 ↑ BAIDU USA, LLC active
US20160055261A1 User-controlled graph analysis system 2014 51 ↑ CRAY INC. active
View All 8 Cited Patents & Full Claim Analysis
Access complete citation analysis, full patent text, legal status history, and assignee chain in PatSnap Eureka IP.
Real-time neural text-to-speech (64 citations) User-controlled graph analysis system (51 citations) + full claim maps
Analyse in PatSnap Eureka IP →
Collaboration Network

Andrew Y. Ng's Research Collaborators

Most Frequent Co-Inventors

Top Co-Inventors of Andrew Y. Ng: Gregory Diamos=15 joint patents, Kai Yang=13, Yu Qing Zhou=12, Sanjeev Satheesh=12 Horizontal bar chart showing the most frequent co-inventors in Andrew Y. Ng's patent portfolio based on PatSnap Eureka data. G. Diamos 15 Kai Yang 13 Yu Qing Zhou 12 S. Satheesh 12

Collaboration Highlights

Andrew Ng's collaboration network reveals two distinct phases: a Baidu-era team anchored by Gregory Diamos and Sanjeev Satheesh working on the Deep Speech ASR programme, and a near-complete personnel transition to a LandingAI team centred on Kai Yang and Yu Qing Zhou for industrial computer vision and data-centric AI tooling. The absence of overlap between these two groups signals how completely Andrew Ng rebuilt his research organisation between his Baidu departure in 2017 and LandingAI's scaling from 2019 onward.

  1. Gregory Diamos 14–15 joint patents
  2. Kai Yang 13 joint patents
  3. Yu Qing Zhou 12 joint patents
  4. Sanjeev Satheesh 12 joint patents
Map the Full Network in PatSnap Eureka IP
Academic Contributions

Research Literature by Andrew Y. Ng

1,841 papers indexed · Research themes span crowdsourcing methodology, medical AI imaging, and computer vision scene understanding — each closely coupled to patent filings by 12–24 months.

Title Year Citations Affiliation / Source
Evaluating non-expert annotations for natural language tasks 2008 1,510 ↑ Stanford University / Dolores Labs
Deep learning for chest radiograph diagnosis: CheXNeXt 2018 1,019 ↑ Stanford University / Duke University
Grounded Compositional Semantics for Finding and Describing Images with Sentences 2014 725 ↑ Stanford University / Google Inc.
Deep-learning-assisted diagnosis for knee MRI: MRNet 2018 540 ↑ Stanford University
3-D Depth Reconstruction from a Single Still Image 2007 529 ↑ Stanford University

Crowdsourcing & Data Annotation

The landmark 2008 study on non-expert annotations via Amazon Mechanical Turk (1,510 citations) is among the most cited in NLP methodology and directly prefigures the data-centric AI philosophy embedded in Andrew Ng's LandingAI patents on labelling, error analysis, and mislabel detection.

Medical AI & Clinical Imaging

The CheXNeXt chest radiograph algorithm (1,019 citations) and MRNet for knee MRI interpretation (540 citations), produced through Stanford's Center for AI in Medicine and Imaging, have directly influenced how AI is evaluated for clinical deployment and generated patent-proximate inventions including US11798159B2.

Computer Vision & Scene Understanding

The 3D depth reconstruction from single images work (529 citations) maps directly to Stanford-assigned depth image construction patents in the portfolio. The grounded compositional semantics paper (725 citations) bridges vision and language understanding — a precursor to modern multimodal AI systems.

Global Footprint

Patent Jurisdictions

Andrew Y. Ng's portfolio spans 8 jurisdictions, with the US-WO-EP-KR-CN axis for the Baidu speech AI family reflecting the global commercial importance of voice technology in those markets.

Patent Jurisdictions for Andrew Y. Ng: United States=46, PCT/WIPO=9, European Patent Office=9, South Korea=4, Australia=4, China=4, India=3, Canada=1 Horizontal bar chart showing the distribution of Andrew Y. Ng's patents by country/jurisdiction based on PatSnap Eureka data. United States 46 PCT / WIPO 9 European Patent Office 9 South Korea 4 Australia 4 China 4 India 3 Canada 1

Filing Markets

The United States accounts for 46 of the 77 curated patents, reflecting the US-centric operations of Stanford, Baidu USA LLC, and LandingAI Inc. The PCT and EP filings (9 each) are concentrated in the Baidu speech AI family, consistent with the global voice assistant and telecommunications market. More recent LandingAI filings concentrate heavily in the US, consistent with a startup focused on domestic industrial clients maintaining a leaner international prosecution budget.

🇺🇸 United States · 46 🌐 PCT / WIPO · 9 🇪🇺 European Patent Office · 9 🇰🇷 South Korea · 4 🇦🇺 Australia · 4 🇨🇳 China · 4 🇮🇳 India · 3 🇨🇦 Canada · 1
For IP Professionals

Why Andrew Y. Ng's Portfolio Matters

Strategic implications for patent attorneys, in-house IP teams, and R&D strategists working in AI-adjacent fields.

⚠️

FTO Considerations

The speech recognition and neural text-to-speech clusters held by Baidu USA LLC remain active in the US and EP. Developers building voice interfaces, ASR pipelines, or neural audio synthesis products should assess these patent families carefully. The Deep Speech approach — end-to-end RNN-based ASR without phoneme dictionaries — is broad enough to warrant thorough claim mapping before commercialisation. LandingAI Inc.'s growing cluster covering training data management, visual guide-based labelling, error analysis workflows, partial labelling mechanisms, and mislabel detection constitutes a coherent and defensible IP position in the data-centric AI space.

🔍

Prior Art Relevance

The most cited patent — US20160171974A1 (159 citations) — is substantive prior art for any ASR, voice interface, or neural audio system. The 3D depth image construction patent (74 citations) from Stanford has been cited by robotics, autonomous vehicle, and augmented reality developers. The distributed task management patent from 2003 (82 citations) retains relevance as prior art decades after filing. Inventor name disambiguation is critical: standard searches for "Andrew Ng" will over-include — jurisdiction filtering and assignee cross-referencing against Baidu USA LLC, LandingAI Inc., and Stanford is recommended.

Unlock Licensing & Monitoring Insights
Access full FTO analysis, licensing opportunity mapping, and portfolio monitoring for Andrew Y. Ng's patent families in PatSnap Eureka IP.
Licensing opportunities Portfolio monitoring alerts Full FTO report
Access in PatSnap Eureka IP →
FAQ

Frequently Asked Questions About Andrew Y. Ng's Patents

The total patent dataset returns 1,714 records attributed to "Andrew Ng" or variants of this name. The curated sample of 77 unique base patents most closely associated with Andrew Y. Ng — based on verified assignees including Baidu USA LLC, LandingAI Inc., and Stanford University — forms the basis of this analysis. Filings span 1995 to 2025 across 8 jurisdictions. The broader count includes filings from other inventors sharing the same common name across different jurisdictions and assignees.
Andrew Y. Ng's verified patent portfolio concentrates in five areas: (1) deep neural network infrastructure and training (G06N3, 8 patents), (2) end-to-end speech recognition and neural text-to-speech synthesis (G10L15, 8 patents), (3) computer vision and industrial visual inspection (G06V10, 6 patents), (4) data-centric machine learning tooling including labelling, error analysis, and mislabel detection (G06N20, 6 patents), and (5) earlier work in enterprise distributed systems and 3D scene understanding (G06F11, 5 patents).
Within the verified patent sample, the most frequent collaborators are Gregory Diamos (14–15 joint patents, spanning Baidu USA LLC and LandingAI Inc. periods), Kai Yang (13 joint patents, LandingAI Inc.), Yu Qing Zhou (12 joint patents, LandingAI Inc.), and Sanjeev Satheesh (12 joint patents, LandingAI Inc. and Stanford). The Baidu-period team also included Bryan Catanzaro, Adam Coates, and Awni Hannun on the Deep Speech research programme.
The primary assignees in the verified sample are LandingAI Inc. (17 patents), Baidu USA LLC (13 patents), SAP Aktiengesellschaft/SAP SE (covering early enterprise computing work), and The Board of Trustees of the Leland Stanford Junior University (covering Stanford-era computer vision research including the 3D depth image construction patents). The concentration at LandingAI Inc. — his most recent personally led venture — signals that Andrew Ng's most focused and controlled IP strategy has emerged during the 2019–2025 period.
The most cited patent in the sample is US20160171974A1 — "Systems and methods for speech transcription" (Baidu USA LLC, 2015) — with 159 citations, reflecting the Deep Speech programme's impact on the ASR field. The second most cited is US7801645B2, a robotic vacuum cleaner detection system with 153 citations assigned to Sharper Image Acquisition LLC. The managing tasks in a data processing environment patent (US20040226013A1, SAP, 2003) has accumulated 82 citations from enterprise computing filers.
Verified filings span 8 jurisdictions: the United States (46 patents), PCT/WIPO (9), the European Patent Office (9), South Korea (4), Australia (4), China (4), India (3), and Canada (1). The US-EP-KR-CN axis for the Baidu speech AI family reflects the global commercial importance of voice technology in those markets. More recent LandingAI filings concentrate heavily in the US, consistent with a startup focused on domestic industrial clients maintaining a leaner international prosecution budget.

Analyse Andrew Y. Ng's Full Patent Landscape in PatSnap Eureka IP

Access inventor disambiguation tools, claim-level analysis, citation network mapping, and competitive landscape assessments across deep learning, speech AI, and industrial computer vision.

🤖
PatSnap Eureka IP AI Patent Intelligence · 208M+ patents
Ask anything about Andrew Y. Ng's patents.
PatSnap Eureka searches 208M+ patents and papers to answer instantly.
Try asking