Google and Samsung top the list of applicants for AI-related patents at the EPO

More entities are seeking protection for AI-related inventions at the European Patent Office, but while grant rates are also on the rise they continue to lag the overall rate. In this co-published piece, Frances Wilding of Haseltine Lake Kempner offers a guide to key data points

The term “artificial intelligence” is sometimes used in European patent applications and patents without further explanation or elaboration, AI apparently being taken to be no more than a known, off-the-shelf option. It is very unlikely that such applications and patents are concerned with developments in AI.

The International Patent Classification (IPC) helps us here. It has an extensive dictionary in which catchwords are linked to classifications. The catchword “Artificial Intelligence” is linked to only one class: G06N. That at least covers machine learning and neural networks, technologies at the core of developments related to artificial intelligence.

The details of the technologies covered by G06N are given further below. Applications and patents given the classification G06N provide a useful indication of EPO-related trends.

AI applications at the EPO growing

There is huge growth in AI filings at the EPO. The chart below shows the numbers of European patent applications having the classification G06N which were published year by year from 2010 to 2020.

image-20211215145803-1

 

Growth in the numbers of published G06N applications took off from 2014. Taking that year as the base, the number of published G06N applications in 2020 was over 12 times the level of 2014. This means that the percentage growth in published G06N applications over that period has been far greater than the percentage growth in total European patent applications published (shown in a comparison graph below). Also, Covid-19 does not appear to have affected this growth.

image-20211215145850-2

The applicants

Between 2016 and 2020, just over 4,000 AI (G06N) applications were published. Around 1,000 different applicants were responsible for them. The top 25 applicants are listed in the table below. They accounted for just about half of all applications.

Rank

Applicant

No. of AI (G06N) Applications 2016-2020

1

Google

266

2

Samsung Group

187

3

Microsoft Technology Licensing LLC

170

4

Intel Corporation

140

5

Siemens Group

140

6

DeepMind Technologies Limited

91

7

Cambricon Technologies Group

90

8

Fujitsu Limited

81

9

Qualcomm Incorporated

81

10

Huawei Technologies Co. Ltd.

68

11

StradVision, Inc.

67

12

Sony Group

61

13

Robert Bosch GmbH

58

14

Accenture Global Group

46

15

Nokia Group

44

16

Tata Consultancy Services Limited

41

17

Koninklijke Philips N.V.

39

18

Baidu Group

37

19

HRL Laboratories LLC

32

20

Commissariat à l'Energie Atomique et aux Energies Alternatives

31

21

Panasonic Group

31

22

Advanced New Technologies Co., Ltd.

27

23

General Electric Group

25

24

Hitachi Group

24

25

Northrop Grumman Systems Corporation

24

Number of grants increasing but lower than in other fields

Apart from grant as a patent, processing of a European patent application may be concluded by the application being refused, withdrawn or deemed withdrawn. The charts below show year by year numbers of AI (G06N) applications processed to conclusions.

image 184For AI (G06N) applications processed to conclusions, the proportion granted as EP patents has been increasing. For AI (G06N) applications concluded in 2019 and 2020 grant as patents was the most likely outcome but was still the outcome for less than half of the applications. The 40% grant rate for AI (G06N) applications in 2020 well behind the EPO’s overall 69% grant rate in the same year, while the 16% refusal rate in 2020 was far higher than the overall 4% EPO refusal rate.

These statistics are likely to be caused by the classification of AI at the EPO as a mathematical method which is not technical and cannot support an inventive step unless linked to implementation or specific technical application. The percentage outcomes for AI and all European applications are shown clearly below.

image 185Although the statistics indicate that, in general, grant is less likely and refusal more likely for AI applications than for EPO applications overall, outcomes do vary greatly for different AI applicants.

The outcomes of applications processed to conclusions over the five-year period 2016 to 2020 for different AI applicants/patentees are indicated in the chart below. The proportion of patents granted varies from 15% to 74% and the proportion of applications refused ranges from 5% to 35%.

Different applicants/patentees also appear to have quite different policies regarding withdrawal of applications – two of the 10 applicants did not positively withdraw any applications as opposed to allowing them to be deemed withdrawn. Other applicants have positively withdrawn good proportions of their applications (which may be as a result of a general policy, or a reaction to poor prospects for success).

image-20211215150744-1

The technologies covered by AI patent applications at the EPO

The IPC can help us again here. The 4,000 European patent applications having G06N IPC classifications published between 2016 and 2020 have lead classifications and, usually, further classifications along with these. The 4,000 lead classifications are spread over more than 700 different individual classifications and the applications have a total of around 15,000 classifications spread over about 2,000 different individual classifications.

The detailed hierarchy of IPC classifications of the technologies covered by G06N is given in the table below, along with the number of occurrences of the classifications across the 4,000 European patent applications. By a clear margin the most frequent classification, G06N 3/04, concerns architecture of neural network models, followed by learning methods for neural network models

International Patent Classification G06N

Computer systems based on specific computational models

(Catchword: “Artificial Intelligence”)

Occurrences of this classification as lead IPC

Total occurrences of this classification

G06N 3/00 Computer systems based on biological models

84

253

• G06N 3/02 using neural network models

77

289

• • G06N 3/04 Architecture, eg interconnection topology

643

1502

• • G06N 3/06 Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons

12

31

• • • G06N 3/063 using electronic means

220

546

• • • G06N 3/067 using optical means

7

8

• • G06N 3/08 Learning methods

247

1258

• • G06N 3/10 Simulation on general purpose computers

15

50

• G06N 3/12 using genetic models

15

83

G06N 5/00 Computer systems using knowledge-based models

47

226

• G06N 5/02 Knowledge representation

85

248

• G06N 5/04 Inference methods or devices

53

255

G06N 7/00 Computer systems based on specific mathematical models

52

314

• G06N 7/02 using fuzzy logic

0

8

• • G06N 7/04 Physical realisation

2

5

• • G06N 7/06 Simulation on general purpose computers

3

5

• G06N 7/08 using chaos models or non-linear system models

3

5

G06N 10/00 Quantum computers, i.e. computer systems based on quantum-mechanical phenomena

97

149

G06N 20/00 Machine learning

82

522

• G06N 20/10 using kernel methods, eg support vector machines [SVM]

6

48

• G06N 20/20 Ensemble learning

26

110

G06N 99/00 Subject matter not provided for in other groups of this subclass

214

555

G06F 15/18 Digital computers/data processing equipment in which a programme is changed according to experience gained by the computer itself during a complete run; Learning machines

 

Most of subclass G06N was added to the IPC in the year 2000 but groups and subgroups G06N 10/** and G06N 20/** were not added until 2019. Parts of the subject-matters of what are now G06N 10/** and G06N 20/** were formerly included in G06N 99/00 but machine learning was formerly included in G06F 15/18 (now subsumed into G06N 20/**).

17

40

Where AI is being used, according to EPO applications 

AI has found use across many fields, from internet search engines through self-driving cars, to medical diagnostics, finance and even agriculture. If potential fields of use of a development in AI are only mentioned in the description in a European patent application, they may not be reflected in an IPC classification applied to the application. Nonetheless, and particularly if a potential field of use appears in a claim of the application, this may be reflected in an IPC classification applied to the application, either as lead classification or as a further classification, which is not a G06N classification.

The most common non-G06N classifications applied to the 4,000 European patent applications published between 2016 and 2020 are indicated in the table below. The most frequent classifications relate to pattern recognition and image analysis. Of course, these techniques can, in turn, be used in many different contexts.

IPC Classification

Occurrences of this classification as Lead IPC

Total occurrences of this Classification

 

G06K 9/00

99

320

Methods or arrangements for reading or recognising printed or written characters or for recognising patterns; eg, fingerprints

G06K 9/46

28

147

Extraction of features or characteristics of the image

G06K 9/62

56

337

Methods or arrangements for recognition using electronic means

G06F 17/30

49

113

Database structures for information retrieval for digital computing or data processing equipment or methods

H04L 29/06

49

164

Communication control/Communication processing characterised by a protocol

G06F 9/50

46

91

Allocation of resources; eg, of the central processing unit [CPU] in arrangements for program control; eg, control units

G06F 3/01

37

84

Input arrangements or combined input and output arrangements for interaction between user and computer

G06Q 10/06

29

70

Resources, workflows, human or project management; eg, organising, planning, scheduling or allocating time, human or machine resources; Enterprise planning; Organisational models

B25J 9/16

27

46

Program controls for program-controlled manipulators

G06F 17/27

26

60

Automatic analysis; eg, parsing, orthograph correction, when handling natural language data

H04L 29/08

17

91

Transmission control procedure; eg, data link level control procedure, for communication control/communication processing

G06Q 30/02

25

86

Marketing; eg, market research and analysis, surveying, promotions, advertising, buyer profiling, customer management or rewards; Price estimation or determination

G06T 7/00

19

77

Image analysis

A61B 5/00

15

64

Measuring for diagnostic purposes in medical or veterinary science

Some of the recent AI applications granted at the EPO

Below are the 32 AI (G06N) patents granted in November 2021, with links to the European Patents Register.

That use for AI has been found across many fields could be illustrated by the patents granted in November, which have concerns from cell (biological) analysis and motor vehicle loss assessment (for insurance purposes), through optimisation of mobile phone networks, to detecting whether a self-driving vehicle is travelling in a one-way street.

AI (G06N) Patents Granted in November 2021

Patent No.

Patentee

Title

3262417

Cellanyx Diagnostics, LLC

Cell Imaging And Analysis To Differentiate Clinically Relevant Sub-Populations Of Cells

3444758

Cambricon Technologies Corporation Limited

Discrete Data Representation-Supporting Apparatus And Method For Back-Training Of Artificial Neural Network

3471005

Nokia Technologies Oy

Artificial Neural Network

3520045

Advanced New Technologies Co., Ltd.

Image-Based Vehicle Loss Assessment Method, Apparatus, And System, And Electronic Device

3292471

Hasan, Syed Kamran

Method And Device For Managing Security In A Computer Network

3407265

Cambricon Technologies Corporation Limited

Device And Method For Executing Forward Calculation Of Artificial Neural Network

3619652

Midea Group Co., Ltd.

Adaptive Bit-Width Reduction For Neural Networks

3392809

Accenture Global Solutions Limited

Quantum Computing Machine Learning Module

3469496

Neura, Inc.

Situation Forecast Mechanisms For Internet Of Things Integration Platform

3557484

Shanghai Cambricon Information Technology Co., Ltd

Neural Network Convolution Operation Device And Method

3557487

ZF Friedrichshafen AG

Generation Of Validation Data With Generative Contradictory Networks

3636001

Huawei Technologies Co., Ltd.

Optimizing Cellular Networks Using Deep Learning

3686837

StradVision, Inc.

Learning Method And Learning Device For Reducing Distortion Occurred In Warped Image Generated In Process Of Stabilizing Jittered Image By Using GAN To Enhance Fault Tolerance And Fluctuation Robustness In Extreme Situations

3703332

Advanced New Technologies Co., Ltd.

Graph Structure Model Training And Junk Account Identification

3706267

ABB Schweiz AG

Artificial Intelligence Monitoring System Using Infrared Images To Identify Hotspots In A Switchgear

3719447

Honeywell International Inc.

Deep Neural Network-Based Inertial Measurement Unit (IMU) Sensor Compensation Method

2973248

PPG Industries Ohio, Inc.

Systems And Methods For Determining A Coating Formulation

3192017

Northrop Grumman Systems Corporation

Tunable Transmon Circuit Assembly

3291090

Deutsche Telekom AG

Method And System For Forming A Digital Interface Between Terminal And Application Logic Via Deep Learning And Cloud

3301611

STMicroelectronics S.r.l.

Artificial Neural Networks For Human Activity Recognition

3343392

INTEL Corporation

Hardware Accelerator Architecture And Template For Web-Scale K-Means Clustering

3398295

Dish Technologies L.L.C.

Systems And Methods For Bandwidth Estimation In Oscillating Networks

3399431

ServiceNow, Inc.

Shared Machine Learning

3399716

ServiceNow, Inc.

Network Security Threat Intelligence Sharing

3490449

Tata Consultancy Services Limited

System And Method For Aiding Communication

3542322

Google LLC

Management And Evaluation Of Machine-Learned Models Based On Locally Logged Data

3611472

Mobileye Vision Technologies Ltd.

Controlling Host Vehicle Based On Detected Parked Vehicle Characteristics

3627213

Eagle Technology, LLC

Multi-Channel Laser System Including An Acousto-Optic Modulator (AOM) With Beam Polarization Switching And Related Methods

3631623

Microsoft Technology Licensing, LLC

Tensor Processor Instruction Set Architecture

3662515

International Business Machines Corporation

Josephson Junctions For Improved Qubits

3671526

Accenture Global Solutions Limited

Dependency Graph Based Natural Language Processing

3783477

Cambricon Technologies Corporation Limited

Integrated Circuit Chip Device

AI patents being opposed at the EPO

Since 2010, only nine AI (G06N) European patents have been opposed, as listed in the table below.

Two oppositions have been finally decided by first instance decisions (1825424 – Opposition rejected, no appeal; 2748686 – Opposition rejected, appeal withdrawn). No oppositions have yet been finally decided after appeal.

The last four oppositions listed in the table appear to be “straw man” oppositions.

Patent No.

Proprietor

Title (en)

IPC

Opponent

1825424

Becton, Dickinson and Company

Graphical User Interface For Use With Open Expert System

G06N 5/02

Beckman Coulter, Inc.

2449510

Dow AgroSciences LLC

Application Of Machine Learning Methods For Mining Association Rules In Plant And Animal Data Sets Containing Molecular Genetic Markers, Followed By Classification Or Prediction Utilizing Features Created From These Association Rules

G06N 5/02

KWS SAAT SE & Co. KGaA

2458178

General Electric Company

Turbine Performance Diagnostic System And Methods

F02C 9/00
G06N 7/00

Siemens Aktiengesellschaft

2582341

Fred Bergman Healthcare Pty Ltd

Method For Analysing Events From Sensor Data By Optimization

A61F 13/42
G08B 19/00
G08B 23/00
G06N 5/04
G06N 99/00

Ontex BVBA

2748686

Robert Bosch GmbH

Method For The Creation Of A Function For A Control Device

G05B 17/00
G06N 7/00
G06F 17/50
G05B 13/04

FEV GmbH

2807526

Omron Corporation

Autonomous Mobile Robot For Handling Job Assignments In A Physical Environment Inhabited By Stationary And Non-Stationary Obstacles

G05B 19/18
G05D 1/02
G06N 3/00

Verschelde, Claire

3111380

Rigetti & Co., Inc.

Processing Signals In A Quantum Computing System

G06N 10/00
G06F 13/36
G06F 13/40
G06F 15/80

Ueberfluss, Eva U.

3111381

Rigetti & Co., Inc.

Operating A Multi-Dimensional Array Of Qubit Devices

G06N 10/00
G06F 13/36
G06F 13/40

Schorr, Frank

 

3217336

Rigetti & Co., Inc.

Impedance-Matched Microwave Quantum Circuit Systems

G06N 99/00
H01L 39/02
H01L 39/22
H03H 7/38
G06F 15/80
G06N 10/00
H01L 27/18
H01P 1/201
H01P 5/02

Ueberfluss, Eva U.

Frances Wilding is a partner of Haseltine Lake Kempner, based in the firm’s London office.

Previous articles by Haseltine Lake Kempner authors in this series can be accessed here:

How to secure AI patents in Europe

Drafting AI patent applications for success at the EPO – eligibility and claim formulation

Drafting AI patent applications for success at the EPO – drafting the full specification

Technology trends – why patent your hidden AI?

Get unlimited access to all IAM content