Page 56 - AI Ready – Analysis Towards a Standardized Readiness Framework

AI Ready – Analysis Towards a Standardized Readiness Framework

(continued)

Examples Potential AI impacts
Characteristics
1. Reference implementations of algorithms
2. Opensource toolsets
3. 3 party applications, APIs, and SDKs
rd
4. Open data, collected from various crowd-sourced mechanisms
5. Transfer learning mechanisms for wider applications
6. Opensource boards to host the edge data processing
Developers
and 7. MQTT [56] and OM2M [22] applications
Opensource 8. IoT gateways such as LoRa gateway and applications, SDKs, and APIs
ecosystems 9. Cloud APIs for subscribing/publishing data from portals [49]
10. Satellite data usage in fire propagation model [51]
11. Keras libraries for data augmentation [60]
12. Opensource tools such as AutoML for automating algorithm selection,
feature generation, hyperparameter tuning, iterative modeling, and
model assessment [77].
13. Open multi-modal data for accessibility.

1. LoRa-based IoT system for peatland management and detection was
deployed in Raja Musa Forest Reserve (RMFR) in Kuala Selangor, Malaysia
[48]
2. According to the World Economic Forum, the pilot study of agricul-
ture-related AI technology on 7000 farmers in the Khammam district of
Telangana (India) showed promising results, where the net income of the
farmers using the AI technology had doubled ($800 per acre) from the
average income in 6 months [33]
3. Possible PoC in IMEC (India-Middle East-Europe Economic Corridor) [50]
is considered in the future.
4. [51] has pilots in India, Portugal, and Brazil, currently monitoring more
than 117 million hectares of Brazilian wetlands, currently studying the
involvement of communities from Amazon (tribes).
5. Simulation approaches such as (sim2real [53]) – including sim2real transfer
(for leveraging simulated data) and curriculum learning (for achieving a
Experi- smoother learning curve from simple to complex scenarios) are used in
mentation, combination with drone-based disaster management [52].
Deployment 6. Pilot deployments include deployment (https:// asr .idri .edu .kh/ ) https://
capability hal .science/ hal -03865538/ in Partnership Ministry of Post and telecom-
munication and publications for Khmer ASR are available. [2]
7. Greenhouse deployments [60] for drone swarms.
8. AI-enabled Soil Analysis and Weather Station for Local farmers using
tinyML models [67].
9. Pilot deployments in hospitals [68] with background noise and low-quality
microphone conditions.
10. Pilot deployment of water conservation use case in the Dodoma region,
selected for its representative soil types, crop varieties, and climate
conditions prevalent in Tanzanian agriculture [71]
11. Application and deployment of AI-based accessibility solutions as
described in [77] for specific domains such as affordable healthcare.
12. Deployment of energy consumption and prediction models [80] in Zheji-
ang China.
13. The pilot models deployed on China Mobile's internal network for feder-
ated learning [1].

49