Page 54 - AI Ready – Analysis Towards a Standardized Readiness Framework

AI Ready – Analysis Towards a Standardized Readiness Framework

(continued)

Examples Potential AI impacts
Characteristics
16. The Models include Khmer ASR (speech engine), TTS (fast speech), and
chatbot (sentenceBERT for finetuning), for accent handling more data
collection is needed, currently this model is based on central City. [2]
17. Context DB, local information about government services [59]
18. chatbots, LLM models, and AI/ML algorithms for finetuning, recommen-
dation, prediction, and generation [59].
19. Combining AI model results from YOLO models and drone imagery [60]
20. Models optimized for energy consumption, to enable a long battery life-
time (e.g. tinyML models for weather stations [67], optimized for smaller
memory and resource usage [68]).
21. LSTM model, useful for time-series data [71]
22. Explainable AI for interpretation, Ray Tune for fine-tuning [80]
23. Combining images and multi-modal data inputs to predict answers,
Audio-visual speech recognition combines sound and video information
to identify speech content [2]
24. "application and algorithm" dual decoupling service combination of
distributed cluster, based on cloud edge resource collaboration technol-
ogy [1]
1. Alternatives for technology
2. Stepwise scenario. E.g. exchange of information between various actors.
3. Inter-PLMN mobility for connected and automated mobility of enterprise
vehicles.
4. Connected vehicles with competing standards such as DSRC [47] and
IEEE 802.11p [42].
5. MQTT (message queuing telemetry transport) [56]
6. Data formats between different government data portals [49], [50]
7. Coordination of satellite data and orbits [51]
8. Combination of camera and satellite data formats [51]
9. Ad hoc network protocols for coordination between drones from differ-
ent vendors/providers [52]
10. Protocols and interfaces for multiagent systems and transfer learning
approaches for models trained using different ML frameworks [52]
Standards and
Interoperability 11. generalization formats across a diverse variety of disaster scenarios [52].
12. Regulations and guidelines for public services management such as
procurement and tenders [57], [58]
13. Data aggregation, model optimization, and deployment standards [67],
[68], [62].
14. Generalizability and informational guidelines on reusing models with
integrated local expertise
15. National sign languages and regional sign languages [77] [2] and
machine translation for Braille for local languages.
16. Energy usage substitution [80]
17. Standards for AIGC – creation and selection of samples for addressing the
availability of samples and edge/cloud collaboration for federated learn-
ing [1]
18. Annotation formats for interoperable annotations [1].
19. 5G/6G Network Intent input using natural language [93]

47