W30 - Early Exploration of Digitalization

Last year I set a goal to study areas related to digitization. Recently I started reading Professor Mai's "Digital Leap," which counts as my first steps into the field of digital transformation.

I'll share some thoughts on what digitization is. What distinguishes those who enable digital capabilities from those who do IT? How does digitization differ from informatization?

The first question is easier to answer; a useful approximation is to ask how those who do IT differ from people who repair computers.

For the second question, based on my current understanding: informatization centers on information-driven dashboards and interfaces, while digitization centers on knowledge-driven solutions and decisions. Informatization typically delivers a human-machine interface; compared with digitization, it does not provide a complete set of solutions and requires users to define strategies and make decisions themselves.

For example, in stock trading, individuals and institutions follow completely different approaches. Individuals generally operate within an informatization framework, making judgments based on market data, moving averages, and other information. Institutions (in the West) generally operate within a digitization framework, doing quantitative trading based on models and data, with most trading decisions executed by machines.

How should we understand information and knowledge? Coincidentally, a few weeks ago I saw an internal blog post about knowledge graphs that explained this well.

To achieve artificial intelligence, one must bridge the gap from information to knowledge. Information is merely a representation of external objective facts; knowledge is the induction and synthesis of those facts from information. Thus knowledge is the network of relationships built on top of information.

A knowledge graph establishes links between pieces of information and helps enable deep natural language understanding. For example, in Meituan search, a user searching for "prayer for blessings" (祈福) is just an activity at the text level; that shallow understanding won't improve search results. If a graph can link "prayer for blessings — place — temple," that relationship can improve search relevance.