ANNOTATION GUIDELINES

Warning

GitHub markdown does not fully support visual annotation components (e.g. entity boxes) used below. We invite user interested in the annotation guidelines to download the documents and open it in a development environment supporting extended markdown syntax (e.g. MacDown, PyCharm, etc) and/or save it as a pdf.

Preliminary comments

The patent corpus that we consider for GB has two types of formats and spans the period 1893-1980. The formatting of UK patent documents have evolved in time but only modestly. Typically, until patent number GB2000001, the first paragraph of the text contains most of the relevant information, which is completed by the header, whose content changes slightly over time. See Figures 1, 2 and 3 for different examples.

From GB2000001 onward, the information is located in the front-page of the patent in a structured way (see Figure 4 for an example). This only concerns 23,889 patent documents as we stop the analysis in 1980.

More on GB patents numbering

• Prior to 1916, patent number is given by a number preceded by the year of application
• From 1916, patents are numbered from 100001 to 1605470 and then from 2000001 onward.
• The last application we consider is 2023380

Format 1, from 1894 to 1979

In the first format, from 1894 to 1979, all the information is given in the first paragraph, which starts by "I, " or "We, " and usually ends with "do hereby declare the nature of this invention to be as follows" or "for which we pray that a patent may be granted to us..."

We extract 5 different "entities" from the body of GB patents.

Entity	Content	E.g.
PERS	Person full name	Maxim Hanson Hersey PERS, Lighting Engineer
ORG	Firm full name	We, The Convex Incandescent Mantle Company Limited ORG, Manufacturers
CIT	The origin of the firm or citizenship of the person	a subject of the king of Great Britain and Ireland CIT,
LOC	Location of the person/firm	Maxim Hanson Hersey, Lighting Engineer, of 145, Bethune Road, Amhurst Park, London N. LOC.
OCC	Occupation of the person	Maxim Hanson Hersey, Lighting Engineer OCC.

These entities are tied together with 3 types of relations.

Relation	Content	E.g.
CITIZENSHIP	Links an ORG/PERS to its CIT	Maxim Hanson Hersey PERS-->CITIZENSHIP-->subject of the king of Great Britain and Ireland CIT
LOCATION	Links an ORG/PERS to its LOC	Maxim Hanson Hersey PERS-->LOCATION-->145, Bethune Road, Amhurst Park, London N. LOC
OCCUPATION	Links an PERS to its OCC	Maxim Hanson Hersey PERS-->OCCUPATION-->Lighting Engineer OCC

Specific labelling issues

• In some cases the text of the patent is repeated twice in the same document, once for the provisional specification and once for the complete specification (see e.g. GB132951A). In such case, all relevant entities must be labelled, even if this means labelling the same entities twice. -In some cases, the name of the inventor, the name of the assignee and even its address can appear at the end of the patent. Those entities must not be labelled (e.g. GB509140A).

Other meta-data in GB patents

The header contains some specific information including:

• the publication date
• the acceptance date
• the application number
• the publication number
• the title
• the technological class
• the name of the inventor(s) - in some instance

Format 2

In the second format, restricted to the year 1979, the information is structured in the front page of the patent. For these patents, the identity of the inventor and the assignee are clearly stated, but only the location of the assignee is given.

Entities

Format 1

PERS

General case

The tag PERS refers to the full name of a patentee person which can or cannot be directly presented as the inventor. This name usually follows "I, " or "We, " and is given in capital letters.

Specific cases

• Inventor name in the header: The name of the inventor can also be specified in the header, preceded by the mention "Inventor(s):" and usually in capital letters. In this case, we label the inventor(s) as PERS. See example 2.
• Third party: In the rare case where the inventor uses a third party to file the application (deceased, mandated), we don't tag the third party person. See example 3 where we do not tag "HAROLD WADE" as a PERS because the context tells us that he is not the inventor.

Examples

1. standard case, from patent GB150481

   We, ANTHONY FULFORD READ PERS , 18, Fown Terrace, Brighton, Manufacturers' Agent, and HAROLD NORMAL READ PERS , 18 Down Terrace, Brighton, Manufacturer's Agent.

2. inventor name in the header, from patent GB1222048

   Inventors WALTER BUNGARD PERS and HANS ZEHNPFENNIG PERS
   

   Improvements in or relating to bearings and bearing liners
   

   We, T.H. GOLDSCHMIDT A.G., a body corporate organised under the Laws of Germany,

3. third party, from patent GB191413361

   (A communication from CHARLES LOUIS MICHOD PERS, Manufacturer, of Chicago Heights, Illinois, United States of. America.)
   

   I, HAROLD WADE, Chartered Patent Agent, of 111 and 112, Hatton Garden, London, E.C., do hereby declare...

4. dead, from patent GB1046893

   We, LEVI CLEWS PERS of 140, Finch Road, Birmingham 19, a British Subject, and FRANCES MABEL GROVES, a British subject, of 41 Ettington Road, Aston, Birmingham 6, legal representative of the late Alfred Groves PERS deceased, a British subject of 140 Finch Road, Birmingham 19, do hereby declare...

5. assignees of, from patent GB664753**

   We, EASTMAN KODAK COMPANY, a Corporation organised under the laws of the State of New Jersey, United States of Aiuevica, of 343, State Street, Rochester, New York, United States of America (Assignees of Fred Waller PERS, a citizen of the United States of America, of 1925, New York Avenue, Huntington Station, New York, United States of America)

ORG

General case

The tag ORG refers to the full name of the organisation which owns the patent. This name usually follows "We, " and is given in capital letters.

Specific cases

• Third party: Similarly to the tag PERS, we do not tag a third party as an ORG if the context tells us that this is not a patentee.
• Former name: Do not label the former name of the company when it is given. See example 3.

Examples

1. standard case, from patent GB848511

   We, LONZA ELECTRIC AND CHEMICAL WORKS LIMITED ORG, a Swiss Body Corporate of Aeschenvorstadt 72, Basel, do hereby declare the invention for which we pray...

2. standard case, from patent GB757350

   We, W.S. BARRETT & SON LIMITED ORG a British Company of 106-108, West Street, Boston, Lincolnshire, do hereby declare...

3. former name, from patent GB786015

   We, THE SCHOLL MFG Co LIMITED ORG, formerly The Scholl Manufacturing Company Limited, a British Company, of 190 St John Street, London, E.C l, England, do hereby declare...

CIT

General case

The tag CIT refers to the citizenship of a PERS or by the origin of a ORG. In the first case, it is usually given in the form "A British citizen" or "A subject of the King of Britain". In the second case, it is usually given in the form "A company of Sweden". The full sequence must be tagged, that is, including "a citizen", "a subject" or "a company".

Specific cases

• ORG from US: When a company is registered in the US, the sequence can be long and include the state of origin. See example 3.

Examples

1. origin of ORG, from patent GB784551

   We, PROGRESS MERCANTILE COMPANY LIMITED, a British Company CIT, formerly of 19 Malden Crescent London, N.W.1 ...

2. origin of PERS, from patent GB500752

   I, HAROLD FREDERICK MAGNUS, of 79 to 82, Fore Street, London E.C.2, British Subject CIT, do hereby declare...

3. ORG from the US, from patent GB388752

   We, ASSOCIATED TELEPHONE & TELEGRAPH COMPANY, of 1033, West Van Buren Street, Chicago, Illinois, United States of America, a corporation organised under the laws of the State of Delaware, United States of America CIT, do hereby declare...

OCC

General case

The tag OCCrefers to the occupation of a PERS or in some rare case of the type of a firm.

Examples

1. OCC of PERS from patent GB163765

   I, HENRY ART KING, Mechanical Draftsman OCC, residing at No. 2012, Linden Avenue et the City of Baltimore, and State of Maryland...

2. OCC of PERS and ORG from patent GB145878

   We, M. HOWLETT AND COMPANY LIMITED, of 140 Hockley Hill, Birmingham, Manufacturers OCC, and JAMES DOLPHIN of 23, Carless Avenue, Harborne, Birmingham, Works Manager OCC, do hereby declare...

LOC

General case

The tag LOC refers to the full location sequence either of a tag PERS or a tag ORG. The address can be given as a full sequence with street number, street name, city, county and country. It can also be simply given by the name of the city/town/village and county (see example 2), or by a postcode (see example 3). In some cases, the location refers to a specific building (see example 4) or university (example 5) and in some other cases, the name of a nearby city is specified (see example 6).

Specific cases

• Non-patentee location: the tag LOC should only be used to label the address of the inventor or the assignee based on the context (i.e. an entity PERS or ORG).

Examples

1. full address, from patent GB1910000882

   Improvements in or relating to Tobacco Pipes, Cigar and Cigarette Holders.

   I, FRANK WOOD, of 4, Rawes Street, Burnley, in the County of lancaster LOC, Commission Agent, do hereby ...

2. city+, from patent GB850480

   We, DEPARTMENT of MINES, a Department of the Provincial Government of Quebec, Quebec City, Province of Quebec, CanadaLOC,, do hereby...

3. post-code from patent GB1254482

   Improved Cylinder Lock Mechanism.

   We OY WARTSILA AB, a Finnish Company of Box 10230, Helsinki 10, Finland LOC, do hereby...

4. building+, from patent GB937358

   We, MARCONI'S WIRELESS TELEGRAPH COMPANY LIIMITED of English Electric House, Strand, London, W.C.2 LOC , a British Company, do hereby declare...

5. university, from patent GB332692

   I, ARTHUR SIMEON WATT, a citizen of the United States of America, of Ohio University, in the City of columbus, State of Ohio, United States of AmericaLOC , do hereby declare...

6. former address, from patent GB1018822

   I, HUSSAIN ALI MOONTASIR, a citizen of the British Commonwealth, of 29, Beechwood Avenue, Kew, Surrey LOC, (formerly of 409, Mistery Chambers opposite Strand Cinema, Colaba, Bombay 5, India), do hereby declare...

7. nearby city, from patent GB1114180**

   I, DENNIS ROBERT CHASE, a British Subject, of 29 St John's Road, Locksheath, Near Southampton, Hampshire LOC.

Relationships

See the common annotation guidelines.

Examples

Figure 1: GB309428A

Figure 2: GB979428A

Figure 3: GB1309428A

Figure 3: GB2016002A