Questions
Why do Simon and Newell claim that computer science is an empirical science? Why would it be important that computer science is an empirical science? What would it mean for the field of artificial intelligence research?
- computer science is an empirical science because it involves building machines and programs to observe their behavior, which provides insights into their functioning.
- For the AI research, it means we improve the machine, the model based on observation of the pratical experimentation.
- we adjust the parameters, test it, and evaluate the result, and adjust the parameter again, until we satisfiy with the result.
- it tells us that we can’t only rely on the abstract principles, the theories, but we also need to take the real world application into consideration.
- We won’t know the results if we never try it.
How do Simon and Newell characterize artificial intelligence research in computer science as empirical? In what sense is it an inquiry into nature as they suggest? Is their analysis convincing? Why or why not?
- They characterize that it involves building systems (like machines and programs) to observe and analyze their behavior which leads to new knowledge about intelligence and cognition
- They suggest that this inquiry into nature occurs through experiments that allow researchers to gather evidence and insights about how intelligent action can emerge from symbol systems.
- I think it is quite convincing, because their argument effectively illustrates the dynamic relationship between theory and practice in understanding artificial intelligence.
- This approach contrasts with purely theoretical studies, making their perspective more grounded in practical experience.
Why do you think they compare their discoveries to those of the cell doctrine in biology, the germ theory of disease, plate tectonics and the doctrine of atomism? Are they good comparisons? Why or why not?
- Newell and Simon compare their discoveries to the cell doctrine, germ theory, plate tectonics, and the doctrine of atomism to highlight how foundational laws in various sciences establish a framework for understanding complex systems. These comparisons are effective because they illustrate how their Physical Symbol System Hypothesis and Heuristic Search Hypothesis serve as essential, guiding principles in artificial intelligence, similar to how those doctrines shaped biology and geology. However, the effectiveness of the comparisons may vary, as the complexities of human cognition and machine intelligence could differ significantly from the biological and physical principles in the other fields. Overall, while the comparisons are apt in emphasizing foundational theories, the specificities of each field’s phenomena may limit direct parallels.
- Conclusion: Laws of qualitative structure are seen everywhere in science. Some of our greatest scientific discoveries are to be found among them. As the examples illustrate, they often set the terms on which a whole science operates.
What is thought and intelligence for Simon and Newell? Is their account of thought and intelligence convincing? Why or why not?
According to Simon and Newell, thought and intelligence are rooted in the ability of physical symbol systems to process symbols, which represent objects, processes, or other symbols. They argue that intelligence involves extracting information from a problem domain and using this information to guide problem-solving through heuristic search methods.
Their account is convincing as it outlines how intelligence can be modeled and understood through these systems, backed by empirical evidence from artificial intelligence research. They provide a structured framework for understanding how symbol systems function in problem-solving scenarios . However, some may argue it lacks deep theoretical explanations for certain complexities, relying heavily on empirical findings.
How do they view psychology and philosophy in so far as they are now offering alternative accounts of trying to understand mind, thought and intelligence? How might they see the future of philosophy of mind and psychological research?
Newell and Simon suggest that psychology should be understood through the lens of empirical inquiry, particularly by using symbol systems to explain human cognition and intelligence. They view psychology as increasingly intertwined with artificial intelligence, emphasizing that understanding human behavior requires modeling it with physical symbol systems. They imply that the future of philosophy of mind and psychological research will involve more empirical methods and a deeper understanding of symbolic processing rather than abstract theorizing alone
Symbols and Physical Symbol Systems
- symbol lies at the root of intelligent action
- symbols are the core of intellignece
- We measure the ability of an intelligent system based on how well it can adapt and overcome the challenges in the enviroments.
- When the task is simple, limited with predictable conditions, the intelligent of computer is less aparent.
- It becomes more obvious as we extend computers to more global, complex and knowledge-intensive tasks.
- The study of intelligent systems progresses slowly, as intelligence does not come from a single principle but from multiple structural requirements.
- One crucial requirement is the ability to store and manipulate symbols.
The Physical Symbol System Hypothesis
- A physical symbol system has the necessary and sufficient means for general intelligent action.
- The hypothesis suggests that any intelligent system will be a physical symbol system upon analysis, and any physical symbol system of sufficient size can be organized to exhibit general intelligence.
Heuristic Search Hypothesis
- symbol systems solve problems by generating potential solutions and testing them, that is, by searching.
- The system usually find the solution by creating symbolic expressions and modifying them repeatedly until they satisfy the conditions for a solution.
- Symbol systems cannot appear intelligent when they are surrounded by pure chaos.
- They exercise intelligence by extracting information from a problem domain and using that information to guide their search, avoiding wrong turns and circuitous bypaths. The problem domain must contain information, that is, some degree of order and structure, for the method to work.