Transcript Slides

PATENT SEARCHES
Istituto Nazionale Fisica Nucleare
17.12.2012
Rossella Osella
Roadmap
INTRODUCTION TO PATENT SEARCHES
ORBIT
Patent searches portal and IP Business Intelligence
SEARCH EXAMPLES
2
The role of patents
Protection and economic
exploitation of the inventions
Sources of information:
technical, scientific, business
publicly available
3
Importance of patent information
Content
Volume
Accessibility
First account
4
Use of patent information
Survey
state of the art
competitors’ activities
Identify
solutions to technical problems
licensing opportunities
duplication of R&D efforts
Avoid
5
infringements
Patent databases
• Worldwide bibliographic
Search fields:
bibliographic data, title and abstract (as published
by the inventor or enriched by the producer)
• Full text
Search fields:
bibliographic data, title, abstract (as published by
the inventor), description, claims
6
Search fields
Bibliographic data
Keywords
Operators
Classification
7
Patent search types
State of the art
Freedom of use
Bibliographic searches
Patentability
Opposition and validity
Landscape and mapping
8
Fampat: patent family database
Full Text Data
Bibliographic
patent data
21 countries
95 countries
Legal Status
information
45 countries
Patent copies
40 countries
One
database
FamPat
Citations
20 countries
US
reassignments
Language technology
Multilingual
wizard
Machine
translations
Key-content
Concepts
Language
technology
IP Business Intelligence
More Charts
Filing trend analysis: multi-dimensional views by publication and priority
date, publication and priority country
Major players identification: Inventors, Assignees, Agents
Technology analysis: IPC, ECLA, USCLASS codes
Concepts analysis
Legal Status analysis
Citation analysis
Full interactivity with search module
Mapping and Clustering based on similarity
•
12
Search examples
Novelty search
Are there relevant patents in the field of
detection of high energy particles by hybrid
diamond-silicon devices?
Landscaping
Overview on radiation dosimeters:
technologies and applications
Search strategy
Search history
NOVELTY SEARCH
Search hit from step 1
LANDSCAPING
Top 30 Assignees
Top 30 Priority countries
Distribution by publication date
Top 75 Concepts
G01T-001/00
Measuring X-radiation, gamma radiation, corpuscular
radiation or cosmic radiation
A61N-005/10
Radiation therapy
X-ray therapy; Gamma-ray therapy; Particle-irradiation
therapy
Mapping
Grazie
www.questel.com
[email protected]
Key content
• Extracted from US, EP, WO, full-text in English
• Identify key sentences describing patent
object, advantages and drawbacks,
independent claims
• Uses morpho-syntactic analysis to spot
important sentences
• Good compromise between conciseness of
bibliographic abstracts and full text
FamPat Concepts
– Field available in FamPat
– Extracted from US, EP, WO, full-text in English
– The Concept Field contains keywords and
multiword phrases that can be used in
searching, browsing and in analysing
Concept extraction
• Main idea: extract concepts rather than words
as a semantic model of patent documents,
weight concepts according to sentence type
• To deal with large volume we need to do this
upfront when loading the data
Semantic concept tagging
• « noun phrases » identification
– Part-of-speech + stemming
– Verb Suppression, some adjectives
– Suppress typical“patent” terminology
• « preferred embodiment », « skilled artisan » ….
– Syntactic normalization:
• « surface of screens »  « screen surface »
• Relevance score computation, based on:
– Field, key sentence morpho-syntactic detection, and
number of occurrences
Concept extraction example
The invention relates generally to molecular level
cleaning of parts by vapor degreasing.
More particularly, the invention relates to a solvent
mixture comprising n-propyl bromide, a [mixture
of low boiling solvents] and, …
The solvent mixture of the invention is nonflammable, non-corrosive, non-hazardous, and has
a low ozone depletion potential.
Normalization examples
the heat conductivity
the conduction of heat
heat conductivity
heat conduction
the user of the cellular telephone
cellular phone users
a user of a mobile telephone
any mobile phone users
their user mobile telephones
HEAT CONDUCTION
MOBILE PHONE USER
Concepts
• Each key concept is “scored”
• Score calculation based on:
– Occurrences of the concept in the doc
– Fields where the concepts appear
– Frequency of the multiword concept in the
database
– Frequency of the concept words in the database
– Non-linear formula (QUESTEL Lingway trade
secret)
– Normalized to 100 (high score)