With IPscreener anyone is able to explore and understand the tech knowledge hidden in patents. By only using a plain text input the semantic AI presents a dashboard of the innovation landscape, identifying similar documents and pointing out relevant paragraphs. Use IPscreener for a smarter way to validating your ideas.
2. MANUAL
The
searcher
reads
the
text
to
extract
keywords
&
classes,
then
creates
a
boolean
strategy
using
combinaBons
of
different
operators
&
fields
AUTOMATIC
The
engine
creates
a
fingerprint
reflecBng
the
content
of
a
text,
then
matched
using
trained
data
for
finding
similar
documents
AI
vs
manual,
what
is
the
difference?
3. The
elements
of
prior
art
searching
Right
Document(s)
Right
Passage(s)
Right
InterpretaBon(s)
Current
focus
of
search
tools
8. Client files
idea
Build
decision
support
processes
IP-
analysis
Finds match
- the idea/concept
lacks novelty
Finds similar
- the idea/concept
is partially known
Manual
search
Finds nothing
– the idea/concept
seems new
Large search
suggested
Minor search
suggested
Auto- classification of
application
No further search
suggested
9. Search
Analysis
Time
Search
Analysis
The
future
role
of
the
Searcher?
Today
Tomorrow
Strategy
10. Customer
metrics;
recall
on
porVolio
0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
60.00%
A61K
A61P
C07D
C07C
C07K
Recall
100
long
Recall
100
short
Recall
ra(o
of
number
X/Y
cita(ons
from
patent
office
report
found
by
IPscreener
within
top
100
hits.
The
scores
are
based
on
3907
patent
applica(ons
from
latest
10
years,
(tle/abstract
vs
full
text
input.
Input
format
Recall
10
Recall
25
Recall
50
Recall
100
Avg
length
Long
31,29%
39,39%
44,05%
48,88%
120293
Short
23,99%
30,59%
34,86%
39,72%
2007
11. Thanks
For
Listening!
QuesBons?
Thanks
for
listening!
QuesBons?
Torsten
Lindholm
Head
of
Customer
Success
Torsten@Ipscreener.com
Book
a
demo
@:
hhps://meeBngs.hubspot.com/torsten