Innovations and Emerging Technologies: A Study of the Italian
Intellectual Property Knowledge Database
Annamaria Demarinis Loiotile
1,2 a
, Francesco De Nicolò
1,2 b
, Alfonso Monaco
1c
,
Sabina Tangaro
3d
, Shiva Loccisano
4
, Giuseppe Conti
5
, Adriana Agrimi
6
, Nicola Amoroso
7e
and Roberto Bellotti
1f
1
Dipartimento Interateneo di Fisica, Università degli Studi di Bari Aldo Moro, Bari, Italy
2
Dipartimento di Ingegneria Elettrica e dell'Informazione, Politecnico di Bari, Bari, Italy
3
Dipartimento di Scienze del Suolo, della Pianta e degli Alimenti, Università degli Studi di Bari Aldo Moro, Bari, Italy
4
Behold srl, Bologna, Italy
5
Netval – Network per la Valorizzazione della Ricerca Universitaria, Lecco, Italy
6
Direzione Ricerca, Terza Missione e Internazionalizzazione, Università degli Studi di Bari Aldo Moro, Bari, Italy
7
Dipartimento di Farmacia-Scienze del Farmaco, Università degli Studi di Bari Aldo Moro, Bari, Italy
Keywords: Natural Language Processing, Italian Patents, Intellectual Property Analytics, Clustering, Technology
Transfer, Knowledge Database, Healthcare 4.0.
Abstract: A great mine of innovation is represented by the excellence of the scientific know-how of the Italian
universities and research centers. But very often university patents remain unvalued and unexploited, in the
so-called “Valley of death”. In the framework of Intellectual Property Analytics and Patent Informatics, this
paper analyses the Italian patent database “Knowledge Share” and its proposed classifications (10
technological areas). By means of Natural Language Processing (NLP) techniques, we examined 1694 patents
from 89 Italian Research Institutions and a cluster analysis revealed the existence of 8 homogeneous clusters
instead of the 10 proposed by the platform. Thus, our findings suggest the presence of possible
inhomogeneities within the traditional classifications, probably due to the emergence of novel technologies
or cross-domain areas, e.g., Healthcare 4.0; moreover, these clusters could lead to better performance in terms
of offer/demand matching for the platform users.
1 INTRODUCTION
The so called Fourth industrial revolution is generally
identified with: exponential evolutions, integration of
technologies and holistic system impact across
society, industry, and countries (Schwab, 2017). The
integration of Industry 4.0 technologies within
society is pivotal for resolving many challenges that
the world and its population are currently facing
(Bartoloni et al., 2022). In this multidisciplinary and
a
https://orcid.org/0000-0003-3503-9952
b
https://orcid.org/0000-0003-3036-8108
c
https://orcid.org/0000-0002-5968-8642
d
https://orcid.org/0000-0002-1372-3916
e
https://orcid.org/0000-0003-0211-0783
f
https://orcid.org/0000-0003-3198-2708
complex context, the technology transfer plays a key
role for the adsorption and dissemination of
technologies, resources, and knowledge to transform
each invention into tangible and useful innovation.
Furthermore, the increased data availability
represents an opportunity to better support decision-
making processes and introduce disruptive
technologies (Baglieri & Cesaroni, 2013;
Aristodemou & Tietze, 2018). For this aim,
companies seek effective strategies for developing
Demarinis Loiotile, A., De Nicolò, F., Monaco, A., Tangaro, S., Loccisano, S., Conti, G., Agrimi, A., Amoroso, N. and Bellotti, R.
Innovations and Emerging Technologies: A Study of the Italian Intellectual Property Knowledge Database.
DOI: 10.5220/0011627000003393
In Proceedings of the 15th International Conference on Agents and Artificial Intelligence (ICAART 2023) - Volume 2, pages 75-86
ISBN: 978-989-758-623-1; ISSN: 2184-433X
Copyright
c
 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
75

and applying technologies while observing the
constraints of time, budget, and without infringing
any third-party Intellectual Property (IP).
The analysis of patents can meet these companies
needs because intellectual property (IP) related
documents represent valuable and recognized sources
of technological and legal knowledge (Aristodemou
et al., 2017). As a matter of fact, different documents,
such as patents, trademarks and other IP registered in
national and international IP systems, contain
important research results which are of great value for
industry, legal researchers, and policy advocates in
science and technology R&D (Trappey, Trappey &
Chung, 2017).
Intangible assets such as R&D, inventions, artistic
and cultural creations, brands, software, know-how,
business processes and data “are the cornerstones of
today's knowledge economy” (Press release EU,
2020).
As reported by the European Commission, the
volume of annual investment in "intellectual property
assets" has increased by 87 percent in the EU over the
past two decades, in contrast to the volume of tangible
(non-residential) investment. Thus, industries that
make intensive use of intellectual property play an
essential role in the EU economy and provide good
and sustainable jobs for society (Press release EU,
2020). Seemingly, according to a study conducted by
the Ponemon Institute LLC looking at the S&P500
(Standard and Poor's 500) companies, the relative
importance of the intangible assets over the total
patrimonial value of those organizations has
increased dramatically in the last 40 years, passing
from representing about 20% (122B$ intangibles vs.
594B$ tangibles) in 1975 to a ratio of more than 5 to
1 in 2018 (21T$ intangibles vs. 4T$ tangibles)
(Ponemon, 2019).
EU Valorization policy strongly encourages the
use of knowledge and technology, the management of
intellectual property, and the involvement of citizens,
academia, and industry (EU valorisation policy,
2020; Demarinis et al., 2022).
Today, intellectual property assets are both
engines of development and drivers of social
transition. Industries that make intensive use of
intellectual property rights (IPRs), such as patents,
trademarks, industrial designs and copyrights,
generate 45 percent of annual GDP (€6.6 trillion) in
the EU and account for 63 million jobs (29 percent of
all jobs) (EU valorisation policy, 2020).
In recent years, Intellectual Property Analytics
(IPA) has emerged as a multidisciplinary approach
used to gain valuable insight about intellectual
property data (Trippe, 2015; Aristodemou & Tietze,
2018).
Of course, patent data can be analyzed in a variety
of ways to fulfill different purposes. Organizations
analyze patents for, but not limited to, determining
novelty in technologies (Erzurumlu & Pachamanova,
2020; Bonino, Ciaramella & Corno, 2010),
forecasting technological developments in a
particular domain (Ernst, 2003) and technological
road mapping (Phaal, Farrukh & Probert, 2004),
analyzing patent trends, strategic technology
planning, extracting the information from patents for
identifying the infringements, determining patents
quality analysis for R&D tasks, identifying the
promising patents, the technological vacuums,
hotspots, and technological competitors (Abbas,
Zhang & Khan, 2014).
Besides, intellectual property knowledge
databases come in heterogeneous forms (text, data,
images, colors, and smells), and it is becoming
increasingly difficult to analyze, synthesize and
classify their contents (Trappey, Lupu & Stjepandic,
2020).
Often the classification of patents and, therefore,
the search and consultation method, are based on
taxonomies self-defined by experts or database
managers and are not very effective.
Automated approaches, such as natural language
processing for data, text and graph mining, clustering
and neural networks, are increasingly used for IP
knowledge processing and various tools have been
developed for supporting patent analysis experts,
business managers, and technology offices (Trappey
et al., 2009; Lei, Qi & Zheng, 2019; Yoon & Park,
2007; Rodriguez et al., 2016; Puccetti, Chiarello &
Fantoni, 2021; Kang et al., 2020; Trappey et al., 2020;
Song, Ran & Yang, 2022).
Analyzing patent data using the automated tools
to discover the patent intelligence through
visualization, citation analysis, and other techniques,
such as text mining is termed as “patent informatics”
(Trippe, 2003). These techniques can be broadly
classified into text mining techniques and
visualization techniques and involve several steps,
including extracting patents from databases,
extracting information from patents, and analyzing
the extracted information to derive logical
conclusions (Abbas, Zhang & Khan, 2014).
Figure 1 shows a generic workflow that can be
used for patent analysis.
In this scenario the aim of this paper is twofold:
- To define a specific patent analysis
workflow aimed at improving patent
classification and fruition;
ICAART 2023 - 15th International Conference on Agents and Artificial Intelligence
76

- To experiment the proposed workflow on an
Italian database named “Knowledge-Share”
(KS).
As side effect the obtained results can also
outlines the presence of different technology domains
(or “technological areas”) with respect to those ones
originally proposed by the KS platform and thus
suggesting the use of a different taxonomy for patents
classification.
The research questions addressed are:
- is the taxonomy based on 10 technological
areas used by KS platform effective in
classifying the whole landscape of academic
patents collected?
- to which extent the new taxonomy obtained
overlap with those one used by the KS
platform?
The paper is organized as follows: Section 2
illustrates the materials and methods
(Knowledgeshare database and the proposed
approach with NLP and clustering), Section 3
describes the analyses and the results, Section 4 draws
some conclusions and suggests future developments.
2 MATERIALS AND METHODS
2.1 The Knowledgeshare Database
Very often university patents remain unvalued and
unexploited. They remain stuck in what is called the
“Valley of death” which represents “the gap between
where publicly available research funding stops and
where private investment or commercial funding
starts” (Hockaday, 2020). Recently, many
universities tried to enhance the value of their
research results to a greater extent, increase the
valorization of research results and the market uptake
of innovative solutions.
Improving easier access to and sharing of
intellectual assets is “key to increase the valorization
of research results and the market uptake of
innovative solutions”: this is one of the objectives
envisaged in the IP Action Plan 2020 of the European
Commission (EU valorisation policy: making
research results work for society, 2020). The ASTP
Survey Report on KT Activities FY2019 shows a
significant lack of valorization of the patented
technologies from the Universities and Public
Research Centers across Europe, since only 18% of
those inventions are licensed or optioned (ASTP
Report, 2019).
Figure 1: Workflow of patent analysis (from patent
database to pre-processing, processing and post processing
in patent analysis).
The Italian landscape, as a part of this picture,
suffers from the same lack of exploitation. In order to
overcome the difficulties that many Italian
universities face in effectively promoting their
research results and their IP assets, the Knowledge-
Share platform was developed as a joint project
involving Politecnico di Torino, the Italian Patent and
Trademark Office at Ministry for Economic
Development and Netval (the Italian Network for the
Valorization of Public Research). KS is a platform
designed for Italian Universities, Research Centers
(PROs), Scientific Institute for Research,
Hospitalization and Healthcare (IRCCS)
to showcase
their patented technologies and spin-off projects
seeking commercialisation opportunities, and for
businesses to find solutions and expertise to
overcome R&D&I challenges (Technology Transfer
System Handbook, 2019). It is specifically aimed to
“translate” the contents of academic patented
inventions into a self-speaking language which
anybody can understand (the so called “patent
marketing briefs”), thus obtaining three important
results: i) to generate a real social and economic
impact at national level, in accordance with the
objectives of the “Third Mission”; ii) to provide a
tangible support to (not only) Italian businesses to
accelerate their innovation processes; iii) to drive
economic return for Universities and PROs to be re-
invested in new technology transfer activities within
the public research system.
Patent database
Pre‐processing
(Transformation into
structured data)
Processing
(Extraction ofstructures)
Post‐processing‐ Patent analysis
(TrendAnalysis,Technologyforecasting,Strategic
TechnologyPlanning,Novelty detection,Technological
Roadmapping,CompetitorAnalysis,etc)
Innovations and Emerging Technologies: A Study of the Italian Intellectual Property Knowledge Database
77

Particularly the platform’s key objectives are to:
• Become the touchpoint between corporations,
SMEs and public research;
• Create a national standard to foster the exploitation
of intellectual property;
• Create an innovation network for technological
excellence at an international level;
• Provide industry scouting teams with an easy and
effective way to tap into the Italian research
landscape;
• Provide a service for technology transfer offices
(market intelligence);
• Promote and foster events and initiatives related to
innovation and exploitation of research;
• Generate spin-offs and innovative technology
projects.
The KS database contains 1694 patents, uploaded
on the platform by 89 Research Institutions
(Universities, Research Centers, Scientific Institute
for Research, Hospitalization and Healthcare, etc).
The application date of the patents ranges from 1999
to 2021. The patents are categorized according to a
taxonomy based on10 technological areas: Aerospace
and aviation, Agrifood, Architecture and design,
Chemistry, Physics, New materials and Workflows,
Energy and Renewables, Environment and
Constructions, Health and Biomedical, Informatics,
Electronics and Communication System,
Manufacturing and Packaging, Transports. The most
populated technological area is Health and
Biomedical that includes more than 1/3 of the total
number of patents in KS database.
The website (https://www.knowledge-share.eu/,
last accessed 16/12/22) allows the search for patents
according to several criteria: the name, the
organization it comes from (i.e. the patent owner), the
area of application, the events at which they were
shown and free full text research. For each patent,
described in an informative and non-technical
language, the technical features, the applications, uses
and characteristics, and the benefits deriving from the
adoption of the technology are illustrated.
Furthermore, information about the inventors, the
priority number, the priority date, the license, the
commercial rights can be found and it is possible to
download a “marketing annex”, i.e., a sheet that
contains the basic information on the patent, conceived
to be a functional and brief communication tool to
share and circulate outside the platform (Fig. 2).
Figure 2: Marketing annex of each patent uploaded on
Knowledge-Share platform.
There are currently more than 1520 registered
users of the KS platform, consisting of companies,
investors, banks, stakeholders and so on. In recent
years, more than 250 contacts have initiated between
universities and companies and some of them have
already led to signed contracts and multiple forms of
collaborations (co-development agreement, new
research agreement leveraging on the same know-
how of a particular invention, license or option
agreements, etc.).
The main challenges now faced by the KS
platform concern:
1. The improvement of the classification
system to enhance the retrieval of relevant
information.
2. The design of pre-matching strategies to
provide an automated or semi-automated
user support.
3. The identification of trends and forecasting
of emerging technologies.
In this paper we propose a focused patent analysis
workflow to face the first challenge.
ICAART 2023 - 15th International Conference on Agents and Artificial Intelligence
78

2.2 Mining the KS Data Base
The Figure 3 shows the proposed approach for patent
analysis and its application to KS. Natural Language
Processing (NLP) techniques have been applied on
the “marketing annex” of the 1694 patents, and in
particular the sections “introduction”, “technical
features” and “application” were processed. These
fields include all the relevant information about the
patent without substantial redundancies.
We performed some of the main steps of NLP in
the following order:
 Tokenization.
 Stop-words removal.
 Stemming.
In the Tokenization phase, texts were subdivided
into single words (also called “tokens”).
In the Stop-words removal phase all the useless
token (such as articles, prepositions conjunctions,
punctuation, numbers etc.) were removed.
In the Stemming phase, the remaining words were
“stemmed” so that only the root-words were kept; for
example, “fished” and “fishing” were transformed
into their common root-word “fish”.
Accordingly, we applied NLP techniques to our
Italian corpus (i.e., the whole set of our collected
texts).
Figure 3: Proposed workflow for KS patent analysis.
Afterwards, we built the TF-IDF matrix (Beel et
al., 2016). The rows of the TF-IDF matrix represent
the texts of the corpus, while the columns stand for
the words. The generic element of this matrix is a
product of two terms:

𝑇𝐹  𝐼𝐷𝐹




𝑇𝐹




𝐼𝐷𝐹


;
where

𝑇𝐹



𝑓

∑
𝑓

;

𝐼𝐷𝐹


𝑙𝑜𝑔
𝑁
𝑑 ∈ 𝐷𝑤

∈𝑑
;
where:
- f
ij
= frequency of word j in patent i;
- N = total number of patents in the corpus;
- D = set of patents, so that |D| = N;
- w
j
= j – th word;
- d = document in D.
We obtained a matrix of 1.694 rows (patents) and
12.505 columns (words). It should be noted that, as
pointed out in the literature (Jun, Park & Jang, 2014),
this matrix is sparse.
Accordingly, in order to reduce the sparsity and
make the clustering process less prone to the curse of
dimensionality, we applied the Singular Value
Decomposition (SVD) (Abdi, 2007) for dimension’s
reduction. Finally, we used the k-means clustering
algorithm (Trappey, Trappey & Chung, 2017; Bock,
2007; Trappey et al., 2013).
Since both SVD and k-means depend on user-
defined parameters, optimal values for these
parameters must be chosen. In particular, since SVD
shrinks the dimension of matrices by using linear
combination of columns, it must be decided the
optimal number of these linear combinations; for
what concerns k-means, the most important
parameter to choose is the number of clusters to
retrieve. Accordingly, we performed a grid-search
exploration of the parameters’ space and used the
Silhouette (Jun, Park & Jang, 2014) for optimization.
Following its definition, Silhouette is a
combination of two terms: (1) the mean distance of
each point of a cluster to the other points of the same
cluster; (2) the mean distance of each point of a
cluster to all the points in different clusters.
Nonetheless, since there are many types of distance
function that can be used in computing Silhouette, we
considered three different distance measures
(Euclidean, Manhattan and Cosine) and chose the
function maximizing the Silhouette for any given
value of the SVD and k-means parameters. Therefore,
the grid-search explored three quantities: (1) the
number of linear combinations in SVD; (2) the
number of clusters in K-MEANS; (3) the distance
function in Silhouette. The combination having the
maximum value of Silhouette was the following:
 SVD - number of linear combinations 10,
 K-means - number of clusters 8,
1694 patents
•NLP
Patents – words
frequency
Matrix
•Clustering
Categories /
technological
areas
Innovations and Emerging Technologies: A Study of the Italian Intellectual Property Knowledge Database
79

 Silhouette - cosine distance.
Thus, the results obtained, presented in the next
section, are based on this configuration of parameters.
3 RESULTS AND DISCUSSION
The first research question addressed here deal with
the number of technological areas, i.e. the number of
class in the taxonomy used by KS. Are these 10 areas
sufficient to capture the whole landscape of academic
patents? If not, do the data suggest a different
classification?
Fig. 4 shows that according to Silhouette 8
clusters should be considered.
N°clusters
Figure 4: The variation in Silhouette value as a function of
the number of clusters (the second one shows a
magnification).
To fully answer the previous research question,
we examined each cluster trying to evaluate whether
there was a consistent overlap with the KS
classification. Besides, we examined the keywords
(in terms of frequency) for each cluster, and we
assigned to each cluster a coherent although
subjective classification, Table 1 shows the clustering
results and Figure 5 illustrates the wordcloud for the
8 clusters.
Table 1: Clustering results.
CLUSTER N° patents
1 Technologies 4.0 (mechanics and
robotics)
271
2 Material science 341
3 Cancer treatment 108
4 Optics - Image processing 179
5 Sensor technology - ICT 288
6 New molecules - new compounds
- pharmacology
242
7 Energy/green Technologies 114
8 Biomedical 133
Therefore, the available 1694 patents were finally
grouped into 8 clusters.
The first cluster, observing the most frequent
words that are system, robot, device, material,
biomedical, sensor, etc., seemed to be related to
technologies 4.0 including mechanics and robotics
applied to different field: healthcare, transport,
design, manufacturing, etc. In particular, a careful
analysis of the patents contained in this cluster, in
conjunction with the analysis of the most frequent
words, reveals interesting innovation spanning
different fields of research:
Figure 5: Wordcloud for the 8 clusters.
- automotive/transportation with patents
pertaining to rail safety devices, underwater drones,
innovations for cycling and motorcycling,
- aerospace with patents inherent in anti-icing
systems for aircraft, rotating aerodynamic elements,
drones, intelligent structures for integrity testing etc.
ICAART 2023 - 15th International Conference on Agents and Artificial Intelligence
80

- architecture and design with innovations in
construction for energy upgrading, seismic isolation
or monitoring complex structures and in the fashion
industry,
- healthcare which is the main area of application
of the innovations in this cluster. The patents present
are truly varied, ranging from innovative wheelchairs
for the disabled, underwater guidance for the blind,
innovative orthotics, endomedicular prostheses,
sensorized heart valve prostheses to devices for
radiotherapy, pneumatic muscles, artificial bladders,
and innovative brassieres and sheaths. But the bulk of
innovations revolve around robotics leading the way:
robotic exoskeletons, robotic platforms for
laparoscopy, wearable robots, biomimetic robots,
robotic surgical simulator, robotic limbs etc.
The second one was dedicated to material
sciences with focus on agrifood, chemistry&physics,
manufacturing and environment sectors. In particular,
a close analysis of the patents contained in this cluster
and the most frequent words, reveals interesting
innovations spanning different fields of research. For
example, we find patents that create new technology
for the production of composite ceramic powders,
biocompatible sandwich panel, invention employing
supercritical carbon dioxide to pasteurize foods or the
realization of a natural product from Rosa canina
seeds obtained by CO
2
extraction in supercritical
phase. Other innovations concern the synthesis of a
pesticide nano-formulation from environmentally
friendly materials or high mechanical performance
materials from stone processing waste, micro-algal
photo-bioreactor, inhibitor preparation of unpleasant
odors from household waste, an economical and
effective method for treating wastewater in liquid
form or a multifunctional hybrid material based on
natural clays for environmental recovery and
bioremediation. Many patents deal with the food
sector: intelligent, active and biocompatible label;
process for manufacturing additives for use in making
antibacterial nanocomposites; innovative coating
composed of a glassy matrix and metal nanoparticles
with antiviral, antibacterial and antifungal properties;
ready-made base for chocolate confectionery
products; device for removing proteins, metals and
other instability agents from wine and vegetable
beverages; production of biopolymers, exploiting
agro-industrial wastes, etc.
The cluster n.4 was focalized on optics and image
processing; in fact we can find here innovations about
packaging of optical signals, automatic machine for
real-time detection of contaminants, optoelectronic
apparatus for measuring position and orientation of
rigid bodies, x-ray concentrator, energy conversion
device, solid-state photodetector, integrated optical
device, automatic immersion ultrasonic system, new
sensor for accurate pH measurements, multi-modal
optical fiber communication system, fully automatic
optical microscope for fast reading of samples
consisting of a transparent dielectric with metal
nanoparticles, use of porous silicon membranes
decorated with silver or gold nanoparticles embedded
in a microfluidic chip. In this cluster are also present
some patents with medical application: textile
electrode device for acquiring electrophysiological
signals from the skin; video-assisted dentistry using
intraoral video cameras; acquisition of surface
electromyographic signals and ultrasound images
from the same portion of muscle; measuring device
for assessing the volume of a breast; treatment of
tumors with ion beams (hadrontherapy).
Analyzing the most frequent words of the cluster
n.5, we could deduce that it is concentrated on sensor
technology and ICT in general. In fact, the main
innovations are: soft-computing techniques for
aerospace, anthropogenic noise control device; low-
cost portable apparatus for characterizing sensor
devices integrated with RFID transponders; RFID
system developed for precise localization and
tracking of objects equipped with low-cost tags;
device prepared for analysis, simulation and
prediction of slope instability phenomena; invention
to protect files from ransomware-type attacks;
ground-based synthetic aperture radar capable of
acquiring both three-dimensional and two-
dimensional images; virtual sensor organized with a
trained neural network. There are also patents with
healthcare application (such as wearable haptic
system to guide the cadence of steps in a person
through vibrotactile stimuli), agricultural application
(technology for disinfection of agricultural soil
through the use of microwaves or radio frequency),
architectural application (innovative IoT system to
manage building cooling through advanced machine
learning techniques), automotive application (logic
control system for automotive; platform, and related
method, for the identification, classification and
subsequent removal of manufacturing defects present
on vehicle components; innovative system that
projects vehicle information and augmented reality
elements to enhance the driving experience) and
economic ones (dynamic and responsive computer
model capable of representing economic interactions
among financial institutions).
Cluster n.7 is the one that best matches
Knowledgeshare's classification; in fact, it is
characterized by patents in energy and green
technologies. Following are some examples: energy
Innovations and Emerging Technologies: A Study of the Italian Intellectual Property Knowledge Database
81

harvester device built into the stem of the paddles;
energy converter, which uses gyroscopic effects to
generate electricity from sea waves; energy
conversion device, which allows electricity to be
generated from wave motion; micro-wind generation
system; enthalpy heat exchanger; solar-derived
thermal energy storage and/or exchange device;
motorized system with parallel kinematics that
enables automatic cleaning of the surface of
photovoltaic panels; spectrally selective solar
absorber coatings with enhanced photo-thermal
performance and stability; device that produces water
from the air; innovative portable device capable of
ionizing water taken from environments outside the
home; new method for the simultaneous treatment of
polluted water and power generation; system to carry
out air humidification and heat recovery in air
conditioning systems; electric generator that
statically converts heat into electricity without the use
of moving parts or matter flows; heat and mass
exchanger made from biocomposite hydrogel. Other
innovations go towards the automotive application:
new cooling solution applicable to electric machines;
integrated system capable of transforming an internal
combustion vehicle into an electric vehicle; hybrid-
electric light aircraft and so on.
Three clusters (n. 3, 6 and 8) were totally devoted
to the health and biomedical sector, broadly
understood.
The cluster n.3 was more related to cancer
treatment with very relevant innovation in this field,
such as: photodynamic therapy as a promising
noninvasive treatment for cancer and nonmalignant
tumors; method for early diagnosis and/or monitoring
of Mucor infection; tumor suppressor of malignant
mesothelioma of the pleura; innovative theranostic
system involving a multifunctional nanoconstruct and
ultrasonic activation set-up capable of treating cancer
cells; use of particular strawberry extracts for the
prevention, treatment, and/or control of the
progression of uterine fibroids; circulating bio-
marker for diagnosis and prognosis of tumor
progression; multi-modular and innovative system
capable of isolating stem cells from small amounts of
adipose tissue; molecular markers predictive of
response to immunotherapy; novel method for
cryopreservation of dental pulp to isolate
mesenchymal stem cells; efficient targeted delivery
system of molecules with therapeutic action (e.g.,
cytotoxic agents) based on adipose stromal stem cells;
method for identifying a biomarker of stemness in
hepatocellular carcinoma cells. Some patents were
focalized on regenerative medicine: non-erodible,
sterilizable, biocompatible hydrogel scaffold for 3D
cell cultures or recombinant protein scaffold for
preparing cell culture plates for use in developing
biomaterials for neuro-regenerative medicine.
The cluster n.6 was more connected with the
formulation of new compounds and the discovery of
new molecules and therapies. Some examples:
medicine designed to counteract the progression of
acute and chronic neurodegenerative diseases; RNA
interference-mediated therapy for neurodegenerative
diseases; synthesis of a leptin antagonist tetrapeptide;
synthetic melanocortins with antimicrobial effects for
the treatment of topical infectious diseases;
pharmaceutical compound for the treatment of
wounds for topical use; pharmaceutical composition,
which contains bactericidal/permeability-increasing
protein and hyaluronic acid with the purpose of
treating different types of arthropathies; new peptide
and its use for the treatment of Alzheimer's disease;
nanostructures capable of delivering oxygen into
hypoxic tissues, which are associated with various
metabolic, ischemic, and infectious diseases;
multifunctional biomaterial consisting of a hydrogel
(hydrogel) that is administered through an injection
directly into the tissue to be treated; use of irisin, a
hormone secreted mainly from skeletal muscle and in
smaller amounts from adipose tissue, as a
drug/strategy for the preservation of the function and
survival of pancreatic islets of Langerhans and, in
particular, the beta-cells that are deputed to insulin
production. Other innovations deal with: antiviral
compounds that find application in the prevention and
treatment of infections caused by coronaviruses; use
of benzofurans as synthetic "natural-like" herbicides,
characterized by high phytotoxic and herbicidal
activity; micro and nanocapsules tannins useful for
the preparation of controlled-release pharmaceutical,
nutraceutical or cosmetic compositions; new yeast
strain that can be used to combat fungal infections in
fungi of agronomic and commercial interest; lentil
extract with cholesterol-lowering and prebiotic
activity, particularly useful in therapeutic
applications and used as a nutraceutical; mixture of
active ingredients from pomegranate seeds useful in
the treatment and/or prevention of obesity and
associated diseases, such as particularly insulin
resistance and type 2 diabetes and hepatic steatosis.
The last cluster (n.8) was named “biomedical” as it
included mainly methods and techniques for disease
diagnosis and monitoring. Some of these are very
interesting: new diagnostic marker for Paget's bone
disease and associated bone tumors; new diagnostic
test to identify the two most common mutations in
chronic myeloid tumors; new test for early detection
ICAART 2023 - 15th International Conference on Agents and Artificial Intelligence
82

Table 2: Comparison between KS technological areas and clusters emerged by the proposed approach.
Knowledgeshare
areas vs Clusters
Technologies
4.0
Material
science
Cancer
treatment
Optics -
Image
processing
Sensor
technology
ICT
New molecules,
compounds,
pharmacology
Energy/green
Technologies
Biomedical TOTAL
Aerospace, aviation et
al
18 9 15 19 2 3 66
Agrifood et al
5 72 10 18 15 5 8 133
Environment and
Constructions et al
54 65 1 17 40 1 26 1 205
Architecture and
design et al
22 4 1 14 6 1 48
Chemistry, Physics,
New materials and
Workflows et al
18 169 6 76 8 34 11 9 331
Energy and
Renewables et al
3 7 5 12 60 87
Informatics,
Electronics and
Communication
System et al
33 1 1 22 140 2 1 5 205
Manifacturing and
Packaging et al
17 6 3 2 1 29
Health and
Biomedical
94 14 100 29 26 188 3 112 566
Transports et al
9 2 12 1 24
TOTAL
273 347 108 180 291 242 117 136 1694
of colorectal cancer which assesses decreased
expression of a protein; noninvasive method suitable
for pancreatic cancer diagnosis at an early stage; fecal
sample testing system able to diagnose major chronic
inflammatory bowel diseases; innovative system for
early diagnosis of acute renal failure; method of in
vitro diagnosis of head and/or neck cancer in tissue
and/or biological fluid samples; new reporter system
that enables early detection of the occurrence of
muscle atrophy; diagnostic for rapid and early
differential diagnosis of ulcerative rectocolitis; next-
generation sequencing techniques to detect specific
molecular signatures of urinary miRNA, which can
be used to distinguish bladder cancer cases from
healthy controls; use of low field nuclear magnetic
resonance for monitoring patients with cystic fibrosis;
innovative method that allows separation of nucleated
fetal cells from maternal peripheral blood at all
gestational ages, etc.
As a second research question, we evaluated to
which extent the new classification overlapped with
the one used by the KS platform. This is strategic if
we want to highlight the presence of emerging
technologies/fields or cross-domain areas. To answer
this question, we built a “contingency table” to
compare KS classifications (rows) with those arising
from our analyses (columns), see Table 2.
Our findings highlight a mismatch between
technological areas defined in KS, as a result of the
self-assignment performed by the inventors, and the
clusters emerged by applying the proposed approach.
This finding suggests that the patent analysis
workflow defined can lead to an alternative and
probably more coherent classification, improving the
offer/demand matching. We are evaluating with
Netval partner the possibility to change the
category/technological areas fixed in KS and/or
create a recommender system able to help the
inventors to correctly assign the patents.
As previously said, it is interesting to note that
patents related to the health sector, originally located
in only one technological area in KS, “Health and
Biomedical” category, that account for a large portion
(about 30 percent of the total), after the processing
proposed in this paper can be rearranged into multiple
and more specialized categories. In fact, in addition
to the three clusters directly related to health, namely
No. 2, No. 6 and No. 8, Cluster No. 1 - Technologies
4.0 - contains several patents connected to new
technologies applied to health. As future work, we
started to extract the most frequent words from the 4
aforementioned clusters dedicated to healthcare and a
complex weighted network was built on them; the
aim is to identify a preliminary “vocabulary”
dedicated to Healthcare 4.0 from the huge mine of
information contained in the Italian patent heritage. In
fact, Healthcare 4.0 represents a great challenge in
Industry 4.0. Integrating innovative technologies
such as the Internet of Health Things (IoHT), cyber-
Innovations and Emerging Technologies: A Study of the Italian Intellectual Property Knowledge Database
83

physical CPS (medical CPS) systems, health cloud,
health fog, big data analytics, machine learning,
blockchain, and intelligent algorithms, it has the goal
of providing better, value-added, and cost-effective
health care services to patients and improving
interactions between patients, stakeholders,
infrastructure, and the value chain (Al-Jaroodi,
Mohamed & Abukhousa, 2020). Furthermore
Healthcare 4.0 focuses on the implementation of
integrated healthcare platforms with progressively
virtualized, distributed and real-time healthcare
services for patients, professionals and formal and
informal care givers. It proposes a shift from a
practitioner-oriented hospital model to a distributed,
patient-centered model (Thuemmler & Bai, 2017), as
well as the integration, sharing and optimization of
the use of health service resources, practitioners and
systems management to improve operations and
reduce costs (Al-Jaroodi, Mohamed & Abukhousa,
2020).Thanks to this research stream, it will be
possible to identify promising and emerging
technologies to become useful applications and
profitable products in healthcare 4.0 (Wang & Lin,
2022).
4 CONCLUSION
In this work we propose a specific workflow for
patent analysis together with its experimentation on
the “Knowledge Share” platform, an Italian patent
database, that includes 1694 patents from 89
Research Institutions (Universities, Research
Centers, Scientific Institute for Research,
Hospitalization and Healthcare, etc.), classified in 10
technological areas. Using NLP techniques and
clustering analyses, the 1694 patents were re-
arranged into 8 clusters, namely: Technologies 4.0,
Material science, Cancer treatment, Optics - Image
processing, Sensor technology – ICT, New molecules
- new compounds – pharmacology, Energy/green
Technologies, Biomedical. Our findings highlight a
mismatch between the technological areas defined in
KS and the clusters obtained by using proposed
workflow. This mismatch is probably due to the self-
assignment performed by the inventors when the
patents were uploaded on the platform. An automated
classification could be more coherent and could
therefore lead to better performance in terms of
offer/demand matching.
The potential benefits are countless, ranging from
the possibility for companies and investors to take
advantage of innovations produced by research
institutions through an improved matchmaking
system to the possibility of introducing novel
technological areas focused on the actual innovation
landscape, thus easing the development of cross-
domain technologies and/or emerging technologies.
This study shows that Italian patents represent an
extraordinary source of innovation that unfortunately
is not yet fully "exploited" as all these inventions do
not reach the market and the population. A great
effort is still needed for a more intelligent
management of intellectual assets that are the only
factor that can foster innovation, creativity,
knowledge sharing and improve the chances that
knowledge reaches the market and brings faster
benefits to society. This is especially true and
strategic for the Healthcare sector that represents a
main challenge in the Industry 4.0 for the final
improvement of quality life and wellbeing of
community and territory.
In conclusion:
 this research aims to emphasize the
importance of technology transfer in the fourth
industrial revolution, from the perspective of
the quadruple helix model which describes
university-industry-government-public
interactions within a knowledge economy;
 patent analysis represents a key tool for
organizations able to analyze, synthesize data
and documents for insights, forecasts,
technology trends, technology infringement,
IP related management etc;
 companies and investors could take advantage
of innovations produced by research
institutions through a matchmaking system
provided by Knowledge share platform;
 “Knowledge share” platform aims at helping
innovators and researchers to make the most
of their results and inventions thereby
generating societal impact, as recommended
by the EC Intellectual Property Action Plan;
 the potential application of a quantitative
framework for innovation are countless,
ranging from scientific policy to R&D
strategies for firms, regions and even nations;
it can be connected to socioeconomic data, to
products and can be embedded in frameworks
for industrial development.
REFERENCES
Abbas, A., Zhang, L., & Khan, S. U. (2014). A literature
review on the state-of-the-art in patent analysis. World
Patent Information, 37, 3-13.
ICAART 2023 - 15th International Conference on Agents and Artificial Intelligence
84

Abdi, H. (2007). Singular value decomposition (SVD) and
generalized singular value decomposition.
Encyclopedia of measurement and statistics, 907-912.
Al-Jaroodi, J., Mohamed, N., & Abukhousa, E. (2020).
Health 4.0: on the way to realizing the healthcare of the
future. Ieee Access, 8, 211189-211210.
Aristodemou, L., & Tietze, F. (2018). The state-of-the-art
on Intellectual Property Analytics (IPA): A literature
review on artificial intelligence, machine learning and
deep learning methods for analysing intellectual
property (IP) data. World Patent Information,
55, 37-51.
Aristodemou, L., Tietze, F., Athanassopoulou, N., &
Minshall, T. (2017). Exploring the future of patent
analytics: a technology roadmapping approach.
ASTP Report (2019) https://www.astp4kt.eu//
assets/documents/Report%20-
%20ASTP%20Survey%20on%20KT%20Activities%2
0FY2019.pdf (last accessed 16/12/2022)
Baglieri, D., & Cesaroni, F. (2013). Capturing the real value
of patent analysis for R&D strategies. Technology
Analysis & Strategic Management, 25(8), 971-986.
Bartoloni, S., Calò, E., Marinelli, L., Pascucci, F., Dezi, L.,
Carayannis, E., ... & Gregori, G. L. (2022). Towards
designing society 5.0 solutions: The new Quintuple
Helix-Design Thinking approach to technology.
Technovation, 113, 102413.
Beel, J., Gipp, B., Langer, S., & Breitinger, C. (2016). Paper
recommender systems: a literature survey.
International Journal on Digital Libraries, 17(4), 305-
338.
Bock, H. H. (2007). Clustering methods: a history of k-
means algorithms. Selected contributions in data
analysis and classification, 161-172.
Bonino, D., Ciaramella, A., & Corno, F. (2010). Review of
the state-of-the-art in patent information and
forthcoming evolutions in intelligent patent
informatics. World Patent Information, 32(1), 30-38.
Demarinis Loiotile, A., De Nicolò, F., Agrimi, A.,
Bellantuono, L., La Rocca, M., Monaco, A., ... &
Bellotti, R. (2022). Best Practices in Knowledge
Transfer: Insights from Top Universities.
Sustainability, 14(22), 15427.
Ernst, H. (2003). Patent information for strategic
technology management. World patent information,
25(3), 233-242.
Erzurumlu, S. S., & Pachamanova, D. (2020). Topic
modeling and technology forecasting for assessing the
commercial viability of healthcare innovations.
Technological Forecasting and Social Change, 156,
120041.
EU valorisation policy (2020) https://research-and-
innovation.ec.europa.eu/system/files/2020-
03/ec_rtd_valorisation_factsheet.pdf (last accessed
16/12/2022)
EU valorisation policy: making research results work for
society (2020). https://ec.europa.eu/info/research-and-
innovation/research-area/industrial-research-and-
innovation/eu-valorisation-policy_en (last accessed
16/12/2022)
Hockaday, T. (2020). University Technology Transfer:
What it is and how to Do it. JHU Press.
Jun, S., Park, S. S., & Jang, D. S. (2014). Document
clustering method using dimension reduction and
support vector clustering to overcome sparseness.
Expert Systems with Applications, 41(7), 3204-3212.
Kang, Y., Cai, Z., Tan, C. W., Huang, Q., & Liu, H. (2020).
Natural language processing (NLP) in management
research: A literature review. Journal of Management
Analytics, 7(2), 139-172.
Lei, L., Qi, J., & Zheng, K. (2019). Patent analytics based
on feature vector space model: A case of IoT. Ieee
Access, 7, 45705-45715.
Phaal, R., Farrukh, C. J., & Probert, D. R. (2004).
Technology roadmapping—A planning framework for
evolution and revolution. Technological forecasting
and social change, 71(1-2), 5-26.
Ponemon Report (2019) https://www.aon.com/
getmedia/60fbb49a-c7a5-4027-ba98-0553b29dc89f/
Ponemon-Report-V24.aspx (last accessed 16/12/2022)
Press release European Commission (2020). “Commission
adopts Action Plan on Intellectual Property to
strengthen EU's economic resilience and recovery”,
https://ec.europa.eu/commission/presscorner/detail/en/
ip_20_2187 (last accessed 16/12/2022)
Puccetti, G., Chiarello, F., & Fantoni, G. (2021). A simple
and fast method for Named Entity context extraction
from patents. Expert Systems with Applications, 184,
115570.
Rodriguez, A., Tosyali, A., Kim, B., Choi, J., Lee, J. M.,
Coh, B. Y., & Jeong, M. K. (2016). Patent clustering
and outlier ranking methodologies for attributed patent
citation networks for technology opportunity discovery.
IEEE Transactions on Engineering Management,
63(4), 426-437.
Schwab, K. (2017). The fourth industrial revolution.
Currency.
Song, K., Ran, C., & Yang, L. (2022). A digital analysis
system of patents integrating natural language
processing and machine learning. Technology Analysis
& Strategic Management, 1-17.
Technology Transfer System Handbook https://www.
polito.it/imprese/trasferimento/TTS_handbook.pdf
(last accessed 15/12/2021)
Thuemmler, C., & Bai, C. (2017). Health 4.0: application
of industry 4.0 design principles in future asthma
management. In Health 4.0: How virtualization and big
data are revolutionizing healthcare (pp. 23-37).
Springer, Cham.
Trappey, A. J., Lupu, M., & Stjepandic, J. (2020). Embrace
artificial intelligence technologies for advanced
analytics and management of intellectual properties.
World Patent Information, 61, 101970.
Trappey, A. J., Trappey, C. V., & Chung, C. L. (2017). IP
portfolios and evolution of biomedical additive
manufacturing applications. Scientometrics, 111(1),
139-157.
Trappey, A. J., Trappey, C. V., Chiang, T. A., & Huang, Y.
H. (2013). Ontology-based neural network for patent
knowledge management in design collaboration.
Innovations and Emerging Technologies: A Study of the Italian Intellectual Property Knowledge Database
85

International Journal of Production Research, 51(7),
1992-2005.
Trappey, A. J., Trappey, C. V., Hsu, F. C., & Hsiao, D. W.
(2009). A fuzzy ontological knowledge document
clustering methodology. IEEE Transactions on
Systems, Man, and Cybernetics, Part B (Cybernetics),
39(3), 806-814.
Trappey, A. J., Trappey, C. V., Wu, J. L., & Wang, J. W.
(2020). Intelligent compilation of patent summaries
using machine learning and natural language
processing techniques. Advanced Engineering
Informatics, 43, 101027.
Trippe, A. (2015). Guidelines for preparing patent
landscape reports. Patent landscape reports. Geneva:
WIPO, 2015.
Trippe, A. J. (2003). Patinformatics: Tasks to tools. World
Patent Information, 25(3), 211-221.
Wang, Y. H., & Lin, G. Y. (2022). Exploring AI-healthcare
innovation: natural language processing-based patents
analysis for technology-driven roadmapping.
Kybernetes.
Yoon, B., & Park, Y. (2007). Development of new
technology forecasting algorithm: Hybrid approach for
morphology analysis and conjoint analysis of patent
information. IEEE transactions on engineering
management, 54(3), 588-599.
ICAART 2023 - 15th International Conference on Agents and Artificial Intelligence
86