Human Task Performance

There is not one single and universally accepted definition of intelligence. 

If there is not one single universally accepted definition of intelligence; how can their possibly be one single universally accepted definition of artificial intelligence?

Per the book, Artificial Intelligence, A Modern  Approach (Third Edition) by AI pioneers Stuart Russell and Peter Norvig, on page 2 of 1151 pages, they provide eight definitions of artificial intelligence that fit into a matrix of four quadrants:

What I am trying to achieve, my garden, is to get a machine to perform work that a human accountant, auditor, or analyst would otherwise have to perform.  The objective is not all work of an accountant, auditor, or analyst. Just the repetitive, monotonous, grueling, boring, mindless tasks.  I would call this "human task performance". This is a rather conservative objective or goal. The reason is so that human accountants, auditors, and analysts can focus their efforts or performing higher value add tasks that machines cannot perform.

And so, to  me, artificial intelligence is an attempt to replicate human task performance of an accountant, auditor, or analyst using a machine of some sort to reliably perform work; preferably better, faster, and/or cheaper than a more costly human with less errors than that human and as good as or better than that human.

By "human task performance" I mean work. In my specific case; accounting work, audit work, or analysis work. A task, a process, a project, an entire workflow. The work relates to accounting.  The processes are industrial processes.

An agent is some one or something that perform tasks. Per the AI book, an agent is something that acts. From what I can tell right now, I see four different types of agents:

  • Human agents: Humans create the actual meaning behind words and numbers using context, observations, measurements, and intuition through actual lived experiences.  Humans communicate with other humans or entire "collectives" of humans.  Humans have specific skills and experience; these skills and experiences are considered knowledge. Humans communicate with these other humans using language, emotions, rules they come up with, and a language referred to as "logic". Logic is a communications tool that takes many forms.  Humans tend to innately understand logic and machines can also be made to understand logic. Human agents can exchange meaning with other human agents per a common understanding within a collective of humans; or with machine agents which can interpret meaning represented in machine interpretable form of some sort.
  • Symbolic artificial intelligence agents: Symbolic artificial intelligence agents try and replicate or mimic human intelligence by representing knowledge in the form of symbols that a machine can interpret, creating rules also using symbols. The same formal logic that humans understand can be represented in a form that symbolic artificial intelligence can also understand. Symbolic artificial intelligence is also referred to as rules-based artificial intelligence.
  • Machine learning agents: Machine learning agents try and mimic how humans use intuition and pattern recognition and the rules of probability to interpret, really guess, what things are supposed to mean.  Because the interpretation is based on probability, machine learning agents can never know for sure if they are right.  Machine learning basically clusters things by identifying the category of something.  Machine learning agents can find patterns completely on their own.
  • LLMs and transformer agents: An LLM or large language model and transformer is a special type of machine learning agent which uses a large set of examples of natural language text and then it uses that large set of examples to guess the position of the next word in a sentence based on context. LLM and transformer agents don't really understand the words they are working with, only the position of words based on the position of previous words. A transformer is incredibly good at processing sequences of sentences by looking at all the words at the same time to understand context. An LLM is a transformer model that has been fed a massive portion of the internet's text so it can predict and generate human-like language. It is important to understand that LLMs and transformers use guessing so you can never be sure if they are correct.
Agents don't need to work in isolation.  In fact when agents work together, the strengths of one type of agent can be used by another type of agent to overcome weaknesses of that agent so that the system as a whole works better than the sum of the parts of the system.  This is called a multi-agent system.
Fundamentally, a computer is a machine.  Over estimating or hyping what a computer can do is not useful.  But on the other hand, you don't want to underestimate what these tools are capable of.

Additional Information:

Comments

Post a Comment

Popular posts from this blog

Digital Proficiency

Image
A paradigm shift is occurring.  The territory is changing.  A new map is necessary. Trying to use your old map to understand the new territory will be unsatisfying. This shift is caused by the difference between how "realspace" (the real world, analog) and "cyberspace" (the internet, digital) operate. While many things will stay the same; it is also the case that entirely new business models and products are possible. Probably the biggest mistake one can make is misunderstanding the notion of intelligence and grouping "human intelligence" and "machine intelligence" together. I would contend that it is a fool's errand to try and compare machine-intelligence to human-intelligence.  They are not even in the same ballpark and they won't be for a long time, if ever.  The notion of "superintelligence" in computers is a fantasy. Computers are nowhere close to achieving superintelligence which I would consider as being human-level inte...
Read more

Inhabiting Babel, A Manifesto for Responsible Meaning Engineering

Image
Inhabiting Babel, A Manifesto for Responsible Meaning Engineering is a book written by Nicolas Figay . This book is very practical and easy for a motivated business professional to read. For me, this book has both tempered my expectations as to what is possible and also reassured me of the real possibilities that exist if you do the required work. Accountants and auditors tend to be on the conservative side of the spectrum.  Me included. I am  actually rather conservative in terms of what technology might bring to the table for accounting, reporting, audit, and analysis. I have learned how to "cut the cards" when dealing with sales people and software vendors. What this book does is bring the wisdom from the experiences of an expert in system interoperability, enterprise architectures, and complex systems to helps motivated business professionals and technical professionals that truly want to understand artificial intelligence, cutting through the rhetoric and hype. This boo...
Read more

Example Financial Statement Holon

Image
As I have mentioned in other posts on my blog, a financial statement is a formal semantic structure , in fact a financial statement is a special type of formal semantic structure referred to as a holon . As I mentioned in my blog posts, Kurt Cagle and Chloe Shannon are providing a lot of very good information about holons on their substack. ( Here is Kurt Cagle's and Chloe Shannon's answers to my questions. Here is an example Kurt Cagle created of a financial statement as a DataBook.) I have provided a bunch of good examples of financial reporting frameworks and reports   In this blog post I want to focus on one specific financial statement and see if I can explain that a financial statement is a "container" of  blocks of information and the "information blocks" that the financial statement contains. Cagle and Shannon point out that a holon is made up of four distinct knowledge graphs that maximize coherence and operationalize the holon which are: The Sc...
Read more