From books to bytes
Metadata, Mining. Challenges and chances with the emerging (Generative) AI revolution.
Information management or information governance [Link: The IGP certification promoted by ARMA]. This field is also know as Management of Information System (MIS).
Data lake-house
(An AIGC from https://deepai.org/machine-learning-model/fantasy-world-generator)
Data lake-house is combination of Data warehouse and Data Lake. I personally like to think it as a lake house, too, which take the surrounding wild nature resources and manufacture them into bakery and goods.
Data lake contains unstructured data (lack of consistent metadata), while data warehouse is the opposite.
The emerging idea is data lake-house that curate raw unorganized lake data into organized warehouse data illustrated by the plot below [Source Link: Microsoft Azure Tech Community].
For instance, Feature learning or Representation learning [Link-wiki] [Paper with code] can potentially contribute tokenize the raw samples and cluster them according to similarities. Similarity quantification are accessible with approaches as simple as cosine similarity or fancier ones such as Principal Component Analysis(PCA) and K-means.
Stop learning AI
This may sounds a bit intimidating as the viral spreading of AI panics.
AI is tapering the learning curve of software (tools for intellectual properties). Copilot that requires license but everyone can use Large (Natural) Language Models (thanks to Rick’s introduction). AI is like cars, smartphones, internet, it is a tool that upgrade the productivity of entire human society. The evolution of AI will only makes it more accessible to general public, more ergonomic.
Stop learning AI, instead, specify the application of AI. Just like no one learns cars, they learn to drive a car, repair a car (Auto mechanic), design a car, and trade a car.