Page 37 - AI Ready – Analysis Towards a Standardized Readiness Framework

AI Ready – Analysis Towards a Standardized Readiness Framework

Table 1: Characteristics of the AI Readiness factors (continued)

AI Readiness Characteristics Notes/Description
factor
APIs descriptions with rest- Availability of API descriptions with standard,
ful description languages well-accepted restful description languages.

Structured or unstructured Availability of structured or unstructured data.
Distance between the serv- Number of hops including wireless hops,
ing system and SINK weightage according to latencies, and other
logical costs incurred.
Data robustness Effectively eliminating noise from valuable
data.
Number of lines of Code Number of lines of Code in Reference imple-
mentations of algorithms related to the use
cases under study.
Number of code reposito- Number of code repositories with reference
ries implementations of algorithms related to the
use cases under study.
Number of Opensource Number of open-source projects e.g. Linux
projects Foundation, Eclipse Foundation, Apache soft-
ware foundation, Python software foundation,
Developer Open-source initiative, Mozilla Foundation,
Ecosystem data, toolsets, and hardware boards to host
created via the edge data processing.
Opensource
Number of marketplaces, Number of marketplaces, play stores, app
play stores, app stores, IoT stores, IoT gateways (hosting 3 party appli-
rd
gateways cations, APIs, and SDKs; LoRa gateway and
applications, SDKs, and APIs).
Usage statistics of open Usage statistics for the Cloud APIs used for
source repositories and subscribing/publishing data from portals [49].
hosted APIs

Hosted applications inte- Developer ecosystem to the serving models in
grating the models the cloud [68].
Number of papers Domain-specific research (Collision avoidance,
published and cited Driver attention, Human detection, Local
Innovation, (e.g. Patents, publications, local
research); local laws and regulations.
Maturity levels (validation, Technology Readiness Level (TRL), Number of
Access to
Research standards compliance, collaborative industry engagements.
certifications, labs)
Number of foundational Foundational models created by research
models teams working in the domains related to the
use cases, (which leads to domain-specific fine-
tuned models).

30