The field of artificial intelligence (AI) utilizes various techniques such as machine learning, deep learning, and natural language processing (NLP) and neural network (NN) algorithms to make predictions.
Machine learning involves training on structured information to identify patterns, while deep learning has the capability to process unstructured information and learn on its own. NLP/NN algorithms can learn and adapt to new information and situations, allowing them to understand the global structure of problems.
In addition to these techniques, AI systems can use trial-and-error (reinforcement learning) and continuous cause/effect relationships, where memory plays an important role in storing and recalling past experiences and outcomes.
However, I think that there are several gaps that need to be addressed for AI systems to function more effectively.
- convergent thinking, which relies on logical inference and previously learned knowledge
- resolving functions through the mechanical application of precise and deterministic rules
- using information provided by humans and then processed through their manipulation
- semantics, which provides structure for the input of information
- syntax, which is used to derive meaning from that information
- ontology, which concerns the nature of reality or truth
- epistemology, which explores the methods of gaining knowledge about that reality
- commonsense, which refers to our pre-reflective knowledge of the external world based on ordinary experience
- the implicit, which involves reading between the lines to interpret implication and the interpretation of information simply focusing on their individual pieces
To address these gaps and improve AI systems, I think there are several added values that can be incorporated.
- AI systems can be designed to improve divergent thinking, which involves producing unique and effective answers through fluidity, flexibility, originality, and elaboration
- AI systems can also be improved to understand the global structure of problems, shaping information across sectors not necessarily dependent on all available information or logical propositions
- Additionally, AI systems can be designed to understand both syntax and semantics, which can enhance their prediction-making capabilities and increase their potential for discovering non-obvious connections between seemingly unrelated pieces of information
- Incorporating ontological and epistemological principles can help AI systems distinguish between apparent truths and hidden truths from information to make more accurate predictions
- Considering commonsense and implicit knowledge as subsets of an overall cognitive process can equip AI systems to understand and interpret the complexities of the information
Through the following steps, I think:
- relationships with linear cause-effect information
- relationships with apparently unrelated and/or non-evident information
- similarities from the above n 2 (through the cooperation between different layers)
- relationships with apparently distant processes
- similarities from the above n 4 (through the cooperation between different layers)
- relationships with non-adjacent in space and time causal information
- similarities from the above n 6 (through the cooperation between different layers)
- learning the dynamics of the process from the above n 2-4-6
- keeping memory of all the above relationships and similarities
- integrating the Implicit to algorithms
- approaching by trial and error trained on multiple heterogeneous datasets
- learning how to exploit and relate them to lead it to predict what will happen not “only” based on past experience but also by proposing events in a Virtual dimension in which to pro-actively simulate their resolution
Finally, I think that to succeed in the digital-AI world, businesses need a teamwork and agreement approach that includes breaking down complex problems, creating urgency, and generating short-term wins.
It's important to use models to analyze situations, predict change, and inspire teams.
AI involves both technical and soft skills, and those with lifelong learning attitudes and lateral thinking can drive innovation.
People with empathetic curiosity and a constructive critical spirit are necessary for guiding change towards the right vision, purpose, and plan.
The KISS principle and Kotter Change Management Model simplify the process, while a teamwork and agreement approach centers around creating urgency, building coalitions, forming visions and initiatives, empowering, removing barriers, generating short-term wins, tracking lessons, and institutionalizing change.
- Philosophy: from Greek = love of wisdom I Aristotle: Philosophy seems to be of no use because it is released from any bond of servitude, but precisely because of this concept of freedom it is the most ‘noble’ matter.... philosophy comes from the wonder of man for the reality that surrounds him
- Plato: Philosophy is the desire to know ..... those who do not feel they are needy, do not want what they do not feel they need
- Socrates: True wisdom lies in the one who knows he doesn't know
- Descartes: You exist as doubting beings ..... reason is nothing without imagination
- Kant: The madman is a smart dreamer
- Schopenhauer: Life and dreams are sheets of the same book, reading them in order is living, browsing them at random is dreaming
- Sartre: If you're sad when you're alone, you're probably in bad company
- Taleb: Every ‘rule’ of ours that we have in mind is a white swan until it encounters a contingency that disappoints it, the black swan ..... the human being tends to explain the unpredictable events retrospectively from which the inability to read the future
- Inductive reasoning (empiricism) from evidence reaches general conclusions (proceeding from the particular to the general)=cause-effect relationship (probabilistic nature)
- Deductive reasoning (syllogism) from statements about what is known comes to a conclusion (going from general to particular)
- Fluid and crystallized intelligence (Cattell 1966):respectively manipulation of abstract symbols (speed and accuracy for new problems) and stored knowledge
- Staged logical systems of children (Piaget 1972): Cognitive schema “manipulation” or operation of mental objects concerning abstractions and symbols by assimilating new information to pre-existing schemas first and then including new information (accommodation)
- Analogical reasoning (Sternberg 1977) moving from the coding of the terms of the problem and their inference to the mapping of the relationships inferred on other terms up to their application to new situations
- Isomorphic parallelism between symbolic structure and reality: what happens in the model (symbolic structure) must represent what happens in reality(Rosen 1986)=the model predicts through a symbolic representation
- Human Information Processing (HIP): mind as information processor (Cognitive Psychology) capable of operating transformation on symbols ("representational states")=creation of representations of reality and their transformation vs "states of fact" of reality (Pessa 1992)
- Intelligence=ability to learn from experience..increasing the ability to adapt to the environment (Sternberg 2000)
- Cattell (1966) proposed two types of intelligence: fluid and crystallized.
- Fluid intelligence involves manipulating abstract symbols to solve new problems, while crystallized intelligence relies on stored knowledge.
- Piaget (1972) developed staged logical systems for children, which involved the operation of mental objects concerning abstractions and symbols. This process includes assimilating new information into pre-existing schemas and accommodating new information.
- Sternberg (1977) introduced analogical reasoning, which involves coding problem terms, inferring relationships, mapping inferred relationships onto other terms, and applying the relationships to new situations.
- Rosen (1986) described isomorphic parallelism between symbolic structure and reality, where the model predicts through a symbolic representation.
- Pessa (1992) discussed the human information processing (HIP) model, which views the mind as an information processor capable of creating representations of reality and transforming them. HIP distinguishes between representational states and states of fact.
- Sternberg (2000) defined intelligence as the ability to learn from experience and adapt to the environment.
