Patent Information Users Group, Inc.

Annual EPO Patent Information Conference would held online November. 3-4, 2020

Plenary presentations:

 Tuesday, 3 November 2020

Pruning to improve quality - an evolution in the EPO's Y-tags for the CPC. Geert Boedt and Victor Veefkind, EPO

The Y-tags for climate change related technologies (e.g. clean energy) have been around for nearly 10 years now. … We will explain the main changes and plans for the future of the Y02/Y04S tags.

Additional information:

The INPADOC classification scheme: lifting the veil over details. Sonia Kaufmann and Christian Soltmann, EPO

…This presentation provides an overview of the detailed structure level of the INPADOC classification scheme and the current status of the classification of INPADOC legal event data on that level.

 Note: Based on results of 2018 users survey [2, 3], EPO has opted to develop “genuine” detailed structure level of INPADOC classification scheme (so named “Option C”), not based on the detailed event list of WIPO ST.27 Standard [5, Annex I] (“Option B”) [2], while of the INDADOC classification scheme categories A-M, P-Y [4, p.7-16] correspond to event categories of the WIPO ST.27 [5, 6]. EPO user’s survey [3]  includes drafts of detailed structure level of INPADOC classification scheme  for 6 of 21 categories (D,G,H, L, R and M, [3, p. 27-32]). For explanation of the detailed scheme and users search strategies in GPI based on it, see [1, p. 22-37] and [3, p. 17-22]
1. Hausser, J. (EPO). 2019. Why is legal event data being categorised? The INPADOC classification scheme and how it helps the patent information user. Presented at Patent Information Fair & Conference, Tokyo, Japan, Nov.8, 2019. 38 p.
2. Soltmann, C. (EPO). The INPADOC classification scheme and WIPO ST.27: An update. Presented at Patent Data Day 2019, 20 March 2019. 16 p.
3. European Patent Office. INPADOC classification scheme: Options for the detailed structure level of the INPADOC classification scheme. EPO. 2018. 32 p. (Authors: Sonia Kaufmann, Christian Soltmann)
4. European Patent Office. INPADOC classification scheme: Description of the 21 categories of the INPADOC classification scheme (v. 1). EPO. 2018, 16 p. (Authors: Sonia Kaufmann, Christian Soltmann)
5.WIPO. Standard ST.27 Recommendation for the Exchange of Patent Legal Status Data.(Rev. July 5, 2019).35 p. (see Key events at p. 3.27.8- 3.27.9 (Nos. 31-33) and Annex I)
6. Seeing the wood despite the trees: WIPO’s new standard on legal event patent data, Patent Information News 2017, No. 4, p.6–7.
See also: An Excel tables of legal event codes categorized on Event Class level, linked to DOCDB/INPADOC weekly coverage update page. For example, W40/2020 table lists 4053 event codes (for 56 patent authorities), 3069 of them have characterized by Event Class [Category] of the Inpadoc Classification Scheme. (Description of legal status codes are available for AU, DE, EP, NZ, UK, USA and FR.)

News from INPADOC. Christian Soltmann, EPO
… recent improvements regarding the INPADOC service includes the availability of more kinds of EP legal event data and JP trial/appeal data.

Wednesday, 4 November 2020 

A forgotten source of prior art? Understanding and searching inventor's certificates Eva Straková and Jürgen Mühl, EPO

“..With several hundred thousands documents, [inventor's certificates from former socialist countries] are still an important source of prior art and might be relevant for novelty and invalidation searches.”  The authors would discusses “challenges in accessing these documents” and “how [to] use Espacenet to access basic information on inventor's certificates from several jurisdictions and some national sources”; “what differences between patents and inventors' rights” and “in which countries do [inventor's certificates] still exist as a major form of IP protection

Additional Information (ab)

DocDB inventor certificate coverage (W 41/2020), cf.  Kind Code Concordance list W 41/2020

Notes: CS Application for Inventor certificates, A1, 2994 docs (1994-1991); CU: A6/B6 (Application for an) Author's certificate for a pharmaceutical product 20& 17 docs. (2009-2012); SU A2 docs starts from 1939, for all years Inventor's certificate of addition; MN  Both A6 and A8 appear to be stands for Inventor's certificate

Using EPO linked open data: opportunities unveiled by the IPLOD project
Dolores Modic, Ljubljana University (Presentation, 16 p.)

… Taking Linked Open EP data from the EPO and SN SciGraph from Springer Nature, we built a new EPO LOD database-centric sub-cloud. Interlinking the data carefully enabled us to put together enriched information on research and innovation activities by individuals and linked with further databases, and to discover even more data via a user-friendly interface.

Additional information:

IP LODB Project: IP Linked Open Data: Building Bridges [between patented information and scientific knowledge] (started in 2018 with support of EPO Academic Research Programme)
Project website: IP LOD Map is a qualitative evaluation of LOD sources that are potentially linkable to EP LOD. (SN SciGraph, KIPO LOD, [USPTO data], DBPedia, GRID (Global Research Identifier), DrugBank, Linked Drugs, Uniport, USPTO patent data, Microsoft Academic Graph (MAG db), etc.) [1, Part 3, slides 6-7]
See an example of IPLodB aligning (and an unique person search interface): Oberländer, Kai Daniel: SciGraph Author & EPO Inventor  country) [1, Part 3, slides 9-10]
IP LodB database, a person-publication dataset.
References:
[1] Modic, D., Johnson, A., Vučkovič, M., Hafner, A., Lužar, B., Rožac, B., Rasmussen, E., 2020. Exploring Opportunities of Linked Open Innovation Data: Part 1, Part 2 and Part 3. Presented Norwegian Intellectual Property Office (NIPO) workshop, February 11, 2020.
[2] Hafner, A., Modic, D., Lužar, B., Johnson, A., Vučkovič, M., Rožac, B., 2020. IPLodB Handbook: A handbook on linked open data for innovation (ver. 1). IP LodB project. July 2020, 21 p.
[3] Modic, D., Johnson, A., Lužar, B., Rasmussen, E., Intellectual Property Linked Open Data Building Bridges (IP LodB) project: Methodology Report. August 2020, 62 p
---------
Linked open EP data [open access bulk download raw data product useful for combining patent / non-patent data, available from April 2018, includes EP / non-EP (simple patent family) applications/publications, patent/NPL citations
Linked  Open Data from EPO (Introduction, 2p.) User guide (v. 1.4, Dec. 2019, 18 p.), Sample SPARQL queries (v. 1.2, June 2020, 10 p.); Documentation: Patent Ontology Overview, Patent Ontology Reference, CPC Ontology Overview, API Overview, API Reference)
Kracker, M., European Patent Information and the CPC Taxonomy as Linked Open Data.in: Ali Khalili, Maria Koutraki (Eds.), Proceedings of the Posters and Demos Track of the 14th International Conference on Semantic Systems (SEMPDS 2018), Vienna, Austria, September 10-13, 2018. Paper No. 96 (4 p.).

Epimorphics, 2016. What is linked data? (White Paper). 2 p. [URI, RDF (Resource Description Framework), RDFS (RDF Vocabulary Language), SPARQL query language to search for information within RDF graphs, SKOS (Simple Knowledge Organization System, controlled vocabularies)]
-------
Springer Nature SciGraph
Kaindl, M. (SpringerNature). 2017. Building a Linked Data Knowledge Graph for the Scholarly Publishing Domain. Presented at the II-SDV 2017, Nice, France, 54 p.
… Springer Nature SciGraph, the new Linked Open Data platform aggregating data sources from Springer Nature and key partners from the scholarly domain. The Linked Open Data platform will initially collate information from across the research landscape, such as funders, research projects, conferences, affiliations and publications. Additional data, such as citations, patents, clinical trials and usage numbers will follow over time. This high quality data from trusted and reliable sources provides a rich semantic description of how information is related, as well as enabling innovative visualizations of the scholarly domain.
SN SciGraph datasets on figshare.com, includes: Dataset: Patents v. 4 19.02.2019  (developed by Dimensions.ai. see below) ( includes information about patents linked to Springer Nature publications. This dataset includes around 330,000 records and is refreshed on a monthly basis.

----See also: Digital Science’s Dimensions: a new scholarly search database… which includes  113 m publications, etc. and  the IFI claims patent database …with 54 million patents and (as of Aug. 2018) 320 million citations within those patents with a further 11 million citations to scholarly articles and 9 million citations from scholarly articles (Hook, D.W., Porter, S.J., Herzog, C., 2018. Dimensions: Building Context for Search and Evaluation. Front. Res. Metr. Anal. 3.)

 
Fighting Coronavirus - background to the EPO's response. Nigel Clarke, EPO (Presentation, 33 p.)

…The EPO's response [to the COVID crisis] included the creation of an online patent resource for scientific and clinical researchers, containing free-to-use targeted searches, patent statistics and downloadable datasheets. The presentation will explore some of the Fighting Coronavirus content.

Resource details:

EPO Fighting Coronavirus: Includes more 180 search strategies (combination of CPC codes and keywords, listed in hyperlinked Excel files below), created, as reported, by “EPO patent examiners who are experts in the technologies directed to the fight against coronavirus infections, together with [EPO] expert patent analysts” on the following areas of research:
[1]Vaccines & Therapeutics: vaccines (10 [search strategies]), candidate therapies for coronavirus/COVID-19 (2), candidate antiviral  and  symptomatic therapeutics (61+18), nucleic acids and antibodies therapeutics (6+1);
[2] Diagnostics: Nucleic acids and Immunological analyses (5+5), Analytical protocols (virus in samples [sampling/devices/microscopy], in-vivo [technique/devices], image processing (20+35+3);
[3] Informatics: Bioinformatics (4), Health informatics (14).

What happens if something is happening? Force majeure in Asian patent legislation and its impact for the patent system. Jutta Haußer, EPO

(examples of interpretation legal events for extensions of deadlines, restorations of rights related to Fukushima earthquake and Covid 19 pandemic)      

Irene Kitsara, WIPO

In 2018, the World Intellectual Property Organization embarked on a journey into the world of in 2019. WIPO developed an AI taxonomy, built an AI search strategy, and analysed nearly 340 000 patents, over 1.5 million scientific publications related to AI, and other non-patent data. In this session, some interesting findings, AI's relationship to intellectual property, and what all these can mean for the future of people, tools and administrations.

Additional Information
WIPO, 2019. WIPO Technology Trends 2019 – Artificial Intelligence. World Intellectual Property Organization, 159 p. https://www.wipo.int/edocs/pubdocs/en/wipo_pub_1055.pdf
The report draws on commissioned background research, based on search strategy and methodology developed by the core analytics team led by Irene Kitsara (WIPO) ((p.12; see below)
AI Clustering Scheme [based on the 2012 ACM Computing Classification System]: Three Clusters (Techniques, Functional applications, Application fields, further divided on 34 categories and 71 subcategories (see p.24-27 Fig. 1.1-1.3,, and ‘Chapter 1 figures.xlsx’ from Raw data). For each (sub) categories search queries has been developed, as demonstrated in PATENTSCOPE Artificial Intelligence Index implementation.

Glossary: Selected AI categories and terms (p.146-149)

See also [PATENTSCOPE Artificial Intelligence Index, which implements for Patentscope the search strategy developed  in the Bakground paper for FAMPAT database, and could be used in other databases.  The index also provides key phrases, CPC (Cooperative Patent Classification) and IPC (International Patent Classification) symbols that can be used to carry out searches in PATENTSCOPE and other patent databases.
Note, that complex Patentscope AI queries is precoded  [for, example query=AItechniqueFuzzyLogic] and appear to be run without displaying Patentscope Query Tree, details what exactly has been searched.

Discussion Rounds

Discussion round 4  To what extent is AI learning to "understand" patents?

The session will address three areas of patent searching and AI: unstructured text, patent figures and the problem of multiple languages

Discussion round 5       AI implemented inventions as prior art

... How does one … search for technical details [of AI implemented inventions] that are difficult to explain or that are simply unavailable? Add to this the ever-expanding public domain repositories of computer generated data...

Discussion round 6        CPC website: what do you need to avoid getting lost in the CPC information labyrinth? [to be revamp in 2021]

Discussion round 7       Non-patent literature: what are the searcher's needs and are they being met?

...how searching non-patent literature is different to searching patent literature, what data is available and where, and what today's tools can do to retrieve it. ...[W]here the gaps are, how they could be met and by whom.

https://www.epo.org/news-events/events/conferences/2020/pi-conference/programme.html