177
Software as a Medical Device: Regulatory and Market Access Implications
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
3. Hidden Markov Model is a statistical Markov model
in which the system being modeled is assumed to be
a Markov process with unobservable (‘hidden’) states.
Wikipedia website. https://en.wikipedia.org/wiki/
Hidden_Markov_model. Accessed 16 February 2021.
4. Symbolic AI: AI system that encodes knowledge
using symbols and structures. CD2 ISO-IEC 22989. 3
September 2020. Symbolic AI performs manipulations
of abstract objects/concepts for logic deduction.
Symbolic AI focuses much less on (mathematical/
numerical) calculation and much more on composition
and manipulation.
5. Logical reasoning is a form of thinking in which
premises and relations between premises are used
in a rigorous manner to infer conclusions that
are entailed (or implied) by the premises and the
relations. Springer website. https://link.springer.com/
referenceworkentry/10.1007%2F978-1-4419-1428-
6_790#:~:text=Logical%20reasoning%20is%20a%20
form,of%20science%20and%20artificial%20intelligence.
Accessed 16 February 2021.
6. Abstract Syntax Tree is a tree representation of the
abstract syntactic structure of source code written in
a programming language. Wikipedia website. https://
en.wikipedia.org/wiki/Abstract_syntax_tree. Accessed
16 February 2021.
7. Probabilistic reasoning combines probability theory
with logic to handle uncertainty of knowledge.
Javatpoint website. https://www.javatpoint.com/
probabilistic-reasoning-in-artifical-intelligence.
Accessed 16 February 2021.
8. Machine learning: process using computational
techniques to enable systems to learn from data or
experience. Source: IEC 23053, 3.16.
9. Knowledge representation is the field of artificial
intelligence dedicated to representing information
about the world in a form that a computer system
can utilize to solve complex tasks such as diagnosing
a medical condition or having a dialog in a natural
language. Wikipedia website. https://en.wikipedia.
org/wiki/Knowledge_representation_and_reasoning.
Accessed 16 February 2021.
10. Natural language processing: system information
processing based upon natural-language understanding
and natural-language generation. CD2 ISO-IEC
22989, 3 September 2020.
11. Artificial Intelligence: A Modern Approach (3rd
Edition), Stuart Russell and Peter Norvig, 2009,
Pearson.
12. Today, in the world of AI there are two schools of
thought: (1) that of Yann LeCun, who thinks we can
reach Artificial General Intelligence via Deep Learning
alone and (2) that of Gary Marcus, who thinks other
forms of AI are needed, notably symbolic AI or hybrid
forms. See Deep learning: A critical appraisal. arXiv
2018 G Marcus. arXiv preprint arXiv:1801.00631, 2019.
13. Deep learning: approach to creating rich hierarchical
representations through the training of neural networks
with many hidden layers. CD2 ISO-IEC 22989, 3
September 2020.
14. COCIR (2020), COCIR Analysis on AI in medical
Device Legislation. September 2020. COCIR website.
https://www.cocir.org/media-centre/publications/
article/cocir-analysis-on-ai-in-medical-device-
legislation-september-2020.html. Accessed 16 February
2021.
15. Samoili S, López Cobo M, Gómez E, De Prato G,
Martínez-Plumed F, and Delipetrev B. AI Watch.
Defining Artificial Intelligence. Towards an operational
definition and taxonomy of artificial intelligence,
EUR 30117 EN, Publications Office of the European
Union, Luxembourg, 2020, ISBN 978-92-76-17045-7,
doi:10.2760/382730, JRC118163. This study assesses
a large set of AI literature from three complementary
perspectives: policy, research, and market. The paper
contains a collection of 46 definitions. Most notable
definitions:
• The High-Level Expert Group (an independent
expert group established by the European
Commission in June 2018) in the context of
the European AI strategy (See AI HLEG
“A definition of AI: Main Capabilities and
Disciplines,” 8 April 2019. Accessed 19 February
2020) provides the following definition: “Artificial
intelligence (AI) systems are software (and
possibly also hardware) systems designed by
humans that, given a complex goal, act in the
physical or digital dimension by perceiving
their environment through data acquisition,
interpreting the collected structured or
unstructured data, reasoning on the knowledge,
or processing the information, derived from this
data and deciding the best action(s) to take to
achieve the given goal. AI systems can either use
symbolic rules or learn a numeric model, and they
can also adapt their behavior by analyzing how the
environment is affected by their previous actions.
[…]”
• Draft IEC 22989 Information Technology:
Artificial Intelligence: Artificial Intelligence
Concepts and Terminology, published 31
October 2019, defines artificial intelligence as
the capability of an engineered system to acquire,
process and apply knowledge and skills. Note:
Knowledge are facts, information, and skills
acquired through experience or education.
16. IMDRF (2020), Artificial Intelligence Medical Devices
(AIMD). IMDRF website. http://www.imdrf.org/
workitems/wi-aimd.asp. Accessed 16 February 2021.
Software as a Medical Device: Regulatory and Market Access Implications
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
3. Hidden Markov Model is a statistical Markov model
in which the system being modeled is assumed to be
a Markov process with unobservable (‘hidden’) states.
Wikipedia website. https://en.wikipedia.org/wiki/
Hidden_Markov_model. Accessed 16 February 2021.
4. Symbolic AI: AI system that encodes knowledge
using symbols and structures. CD2 ISO-IEC 22989. 3
September 2020. Symbolic AI performs manipulations
of abstract objects/concepts for logic deduction.
Symbolic AI focuses much less on (mathematical/
numerical) calculation and much more on composition
and manipulation.
5. Logical reasoning is a form of thinking in which
premises and relations between premises are used
in a rigorous manner to infer conclusions that
are entailed (or implied) by the premises and the
relations. Springer website. https://link.springer.com/
referenceworkentry/10.1007%2F978-1-4419-1428-
6_790#:~:text=Logical%20reasoning%20is%20a%20
form,of%20science%20and%20artificial%20intelligence.
Accessed 16 February 2021.
6. Abstract Syntax Tree is a tree representation of the
abstract syntactic structure of source code written in
a programming language. Wikipedia website. https://
en.wikipedia.org/wiki/Abstract_syntax_tree. Accessed
16 February 2021.
7. Probabilistic reasoning combines probability theory
with logic to handle uncertainty of knowledge.
Javatpoint website. https://www.javatpoint.com/
probabilistic-reasoning-in-artifical-intelligence.
Accessed 16 February 2021.
8. Machine learning: process using computational
techniques to enable systems to learn from data or
experience. Source: IEC 23053, 3.16.
9. Knowledge representation is the field of artificial
intelligence dedicated to representing information
about the world in a form that a computer system
can utilize to solve complex tasks such as diagnosing
a medical condition or having a dialog in a natural
language. Wikipedia website. https://en.wikipedia.
org/wiki/Knowledge_representation_and_reasoning.
Accessed 16 February 2021.
10. Natural language processing: system information
processing based upon natural-language understanding
and natural-language generation. CD2 ISO-IEC
22989, 3 September 2020.
11. Artificial Intelligence: A Modern Approach (3rd
Edition), Stuart Russell and Peter Norvig, 2009,
Pearson.
12. Today, in the world of AI there are two schools of
thought: (1) that of Yann LeCun, who thinks we can
reach Artificial General Intelligence via Deep Learning
alone and (2) that of Gary Marcus, who thinks other
forms of AI are needed, notably symbolic AI or hybrid
forms. See Deep learning: A critical appraisal. arXiv
2018 G Marcus. arXiv preprint arXiv:1801.00631, 2019.
13. Deep learning: approach to creating rich hierarchical
representations through the training of neural networks
with many hidden layers. CD2 ISO-IEC 22989, 3
September 2020.
14. COCIR (2020), COCIR Analysis on AI in medical
Device Legislation. September 2020. COCIR website.
https://www.cocir.org/media-centre/publications/
article/cocir-analysis-on-ai-in-medical-device-
legislation-september-2020.html. Accessed 16 February
2021.
15. Samoili S, López Cobo M, Gómez E, De Prato G,
Martínez-Plumed F, and Delipetrev B. AI Watch.
Defining Artificial Intelligence. Towards an operational
definition and taxonomy of artificial intelligence,
EUR 30117 EN, Publications Office of the European
Union, Luxembourg, 2020, ISBN 978-92-76-17045-7,
doi:10.2760/382730, JRC118163. This study assesses
a large set of AI literature from three complementary
perspectives: policy, research, and market. The paper
contains a collection of 46 definitions. Most notable
definitions:
• The High-Level Expert Group (an independent
expert group established by the European
Commission in June 2018) in the context of
the European AI strategy (See AI HLEG
“A definition of AI: Main Capabilities and
Disciplines,” 8 April 2019. Accessed 19 February
2020) provides the following definition: “Artificial
intelligence (AI) systems are software (and
possibly also hardware) systems designed by
humans that, given a complex goal, act in the
physical or digital dimension by perceiving
their environment through data acquisition,
interpreting the collected structured or
unstructured data, reasoning on the knowledge,
or processing the information, derived from this
data and deciding the best action(s) to take to
achieve the given goal. AI systems can either use
symbolic rules or learn a numeric model, and they
can also adapt their behavior by analyzing how the
environment is affected by their previous actions.
[…]”
• Draft IEC 22989 Information Technology:
Artificial Intelligence: Artificial Intelligence
Concepts and Terminology, published 31
October 2019, defines artificial intelligence as
the capability of an engineered system to acquire,
process and apply knowledge and skills. Note:
Knowledge are facts, information, and skills
acquired through experience or education.
16. IMDRF (2020), Artificial Intelligence Medical Devices
(AIMD). IMDRF website. http://www.imdrf.org/
workitems/wi-aimd.asp. Accessed 16 February 2021.