169 lines
7.2 KiB
Markdown
169 lines
7.2 KiB
Markdown
file:: [Sundberg e Holmström - 2023 - Democratizing artificial intelligence How no-code.pdf](file://C:/Users/david/Zotero/storage/GCSEEXQP/Sundberg e Holmström - 2023 - Democratizing artificial intelligence How no-code.pdf)
|
|
file-path:: file://C:/Users/david/Zotero/storage/GCSEEXQP/Sundberg e Holmström - 2023 - Democratizing artificial intelligence How no-code.pdf
|
|
|
|
- a broad subse
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: green
|
|
id:: 6475cd05-9520-4d62-881a-049972e50d98
|
|
- a computer program applies algorithms and statistical models to construct complex patterns of inference within data
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475cd10-3b3f-4117-998e-49d58bef2aee
|
|
hl-stamp:: 1685441811334
|
|
- MLOps
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475cd1d-024b-461c-bd77-26aa6ba58110
|
|
- use graphical user interfaces, which help novice users to train ML models
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475cd3f-4d77-460e-97c4-ef912662690e
|
|
- based systems providing on-demand services to organizations and individuals to deploy, develop, train and manage AI models” (Lins et al., 2021)
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475cd61-0461-4852-b977-11f6d140cdc5
|
|
- no-code or low-code tools can support ML pipelines
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475cd6d-c081-46de-b4f6-89d5f623cf6e
|
|
- AI in digital transformation initiatives
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475cd75-7639-47ac-b086-4be0b7b98d68
|
|
- training AI models without any coding and we argue that they have high potential utility in operationalization of ML
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475cd81-03c6-477f-aa24-72cf3ff29ad3
|
|
- No-code AI platforms can make AI accessible and affordable for organizations that lack the size and technological competence needed for solutions that require advanced coding skills, and guide users through the process of developing and deploying AI models, with no need to learn the detailed workings of complex algorithms.
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475ce42-c509-43fd-b59f-188b83e831e5
|
|
- AI capabilities as a service
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
hl-color:: purple
|
|
id:: 6475ce50-e490-41fa-b1a7-fe5d6533c73e
|
|
- How can no-code AI support Machine Learning Operations (MLOps)?
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
hl-color:: purple
|
|
id:: 6475ce7f-fe93-4f72-8fae-bf0818a6d53a
|
|
- ML has been defined as “a broad subset of artificial intelligence, wherein a computer program applies algorithms and statistical models to construct complex patterns of inference within data” (
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475ee78-c1fa-4afa-b20a-1fc39cfbc9d6
|
|
- many organizations are engaging in systematic ways of developing, deploying and monitoring ML operations
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475eeab-777c-44c6-ae09-759c0974e871
|
|
- . Large tech companies such as Amazon, Google, IBM and Microsoft
|
|
ls-type:: annotation
|
|
hl-page:: 4
|
|
hl-color:: purple
|
|
id:: 6475f13d-419d-43dc-952e-4fdcc81cac84
|
|
- By shifting the required capabilities for developing ML from a domain restricted to data scientists to a much broader domain, no-code AI can thus enable new paths for the development of intelligent systems
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
hl-color:: purple
|
|
id:: 6475f41f-1ea5-48f0-a5c3-b89ae918fe18
|
|
- he development of ML significantly differs from traditional software development as ML models are dynamic and constantly changing (Borg, 2022; Moreschini et al., 2022).
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
hl-color:: purple
|
|
id:: 6475f4d4-415f-4062-9b93-d0cb945b9d78
|
|
- preparation of collected data to train an AI model with an ML algorithm
|
|
ls-type:: annotation
|
|
hl-page:: 5
|
|
hl-color:: purple
|
|
id:: 6475f539-7df5-4fd9-8fbe-22ab46d7acad
|
|
- th
|
|
ls-type:: annotation
|
|
hl-page:: 6
|
|
hl-color:: purple
|
|
id:: 6475f966-e14b-4063-b27d-73408290272e
|
|
- end to degrade over time, unless these changes are considered, and accounted for using new datasets(Constantiou & Kallinikos, 2015), and additional cycles of training and evaluation.
|
|
ls-type:: annotation
|
|
hl-page:: 6
|
|
hl-color:: purple
|
|
id:: 6475f9fb-1bc1-447d-837b-3b96ed46eb45
|
|
- MLOps is a response to the challenges associated with incorporating ML models into production
|
|
ls-type:: annotation
|
|
hl-page:: 6
|
|
hl-color:: purple
|
|
id:: 64760e8d-3184-45f3-b77c-a67656b1d813
|
|
- Challenge 1: Gaps between domain experts and technical experts
|
|
ls-type:: annotation
|
|
hl-page:: 6
|
|
hl-color:: purple
|
|
id:: 64760f44-5784-4073-89e1-be8d971b052a
|
|
- Challenge 2: Slow problem-solution iterations due to ineffective architectures
|
|
ls-type:: annotation
|
|
hl-page:: 6
|
|
hl-color:: purple
|
|
id:: 64760f8b-3b66-4dc1-8934-941113fda8d7
|
|
- Challenge 3: Infrastructure management
|
|
ls-type:: annotation
|
|
hl-page:: 7
|
|
hl-color:: purple
|
|
id:: 64760f9b-291a-43ef-8a52-314b5fd6763f
|
|
- ML and DL approaches are valuable in the development of drugs for medical uses, but the lack of code-free and user-friendly applications prevent use of such tools by domain expert
|
|
ls-type:: annotation
|
|
hl-page:: 8
|
|
hl-color:: green
|
|
id:: 64770b24-dcd8-4ddd-8f8f-339197c97f39
|
|
- simplifying the data annotation process through intuitive interfaces where domain experts can label text or images, or objects in images, based on their knowledge and experience. This can promote the establishment of cross-
|
|
ls-type:: annotation
|
|
hl-page:: 9
|
|
hl-color:: green
|
|
id:: 64770d07-cde2-467e-bad7-c6e25938af8e
|
|
- generative AI models are increasingly being integrated in different kinds of systems, we expect them to be a part of many AI solutions in the near future.
|
|
ls-type:: annotation
|
|
hl-page:: 9
|
|
hl-color:: green
|
|
id:: 64770dce-da5b-4de0-a69d-cf2c27c75715
|
|
- platform“enables quick prototyping, rapid experimentation and reduces the time to production by
|
|
ls-type:: annotation
|
|
hl-page:: 9
|
|
hl-color:: green
|
|
id:: 64771340-d904-41d6-94f2-17a55682bf15
|
|
- standardizing model building and deployment
|
|
ls-type:: annotation
|
|
hl-page:: 10
|
|
hl-color:: green
|
|
id:: 6477136a-ddbb-4ab4-8402-929bed8b20c8
|
|
- No-code AI can both help AI experts to meet optimization objectives and enable domain experts to test their ideas to generate value for their firm, without dependence on coders
|
|
ls-type:: annotation
|
|
hl-page:: 10
|
|
hl-color:: green
|
|
id:: 647713a6-c740-4f36-b3fd-8e3ed3b84d2b
|
|
- Aiding infrastructure management
|
|
ls-type:: annotation
|
|
hl-page:: 10
|
|
hl-color:: green
|
|
id:: 647714cc-d6cf-4630-8bb4-ef6d9dbc8c2a
|
|
- AI can be a complex process that requires clear understanding of the business goals and objectives, as well as the specific challenges the organization
|
|
ls-type:: annotation
|
|
hl-page:: 12
|
|
hl-color:: green
|
|
id:: 64771629-340b-4593-83c1-688e1f8c0de5
|
|
- code AI platforms allow individuals, regardless of their technical background, to build and deploy
|
|
ls-type:: annotation
|
|
hl-page:: 12
|
|
hl-color:: green
|
|
id:: 64771647-69a0-4492-8ebb-6a957dedded7
|
|
- help organizations to save time and resources, by avoiding the need to hire expensive data scientists or developers to build AI models. Thus, by engaging with no
|
|
ls-type:: annotation
|
|
hl-page:: 12
|
|
hl-color:: green
|
|
id:: 6477166e-015d-44d8-bf90-ae883324541d |