Welcome! In this module, you will get acquainted with the symbolic AI techniques characteristic of the 60s and 70s, including the Dendral and Eliza systems.
Those who complete this module will be able to:
- understand the methods used in the second era,
- familiarize the strengths and weaknesses of these techniques,
- get to know the Dendral and Eliza systems.
During the lesson you must read the written explanations and follow the given instructions at interactive elements. You can assess your understanding by taking the quiz. In the second part of the module, you will need to solve a task that requires creativity.
Formal logic and the if-then rule
In this module, we can familiarize with the formal logic related to the second stage of AI development, as well as its connection to AI algorithms.
The second era of AI history, the era of abstract theories and symbolic AI, brought unique advancements in computer science and artificial intelligence theory. During this period, a significant portion of AI research focused on the utilization of logic, symbols and formal rules. Emphasis was placed on algorithm development and knowledge representation, while data and problem analysis were embedded within theoretical frameworks.
Algorithms were manually crafted, and machines generally performed well in narrower domains where rules could be easily applied. One key characteristic of symbolic AI was the use of “if-then” rules: if a certain condition is met, then the machine executes a specific action or inference.
Dendral, Eliza and the ProLog
In this module, the most significant creations of the era are introduced: Dendral, Eliza and the ProLog language.
During this period, several renowned AI programs emerged, employing a symbolic approach. For instance, DENDRAL (video) was capable of modeling chemical structures, whereas ELIZA (video) assumed the role of a psychotherapist, engaging with people in natural language. However, the methods of symbolic AI also encountered limitations: machines struggled to handle uncertainty, subjective interpretation and emotional context.
As researchers developed algorithms, they designed increasingly complex and general models. The PROLOG language, for example, facilitated logic-based programming, while LISP served for efficient handling of lists and recursion. Researchers aimed to create models of human thinking and devised systems capable of logical inference and problem-solving.
The contribution of the era of abstract theories and symbolic AI to the advancement of AI was undeniable, as it laid the groundwork for later domain diversification with fundamental frameworks and algorithms. However, it became apparent that machines faced challenges in dealing with uncertainty and non-linear problems, prompting further research in subsequent eras.
The interactive quiz is accessible by clicking on the link above.
https://www.youtube.com/watch?app=desktop&v=8jGpkdPO-1
Yhttps://www.youtube.com/watch?v=PxavdLkjD1I
Creative task
With the help of artificial intelligence, we can chat online with a psychologist. Let’s see how! You’ll need creativity and English language skills for the following task to converse with Eliza.
https://web.njit.edu/~ronkowit/eliza.html
Open the website above, where you can chat with Eliza, an online psychotherapist chatbot.
Make a conversation with Eliza. Ask for advice! If you don’t have a problem, make up one. For example, inquire about how you could fit more easily into social situations as you experience a lot of anxiety.
https://web.njit.edu/~ronkowit/eliza.html
Learning outcomes:
In this module, we have acquainted with the second chapter of the history of artificial intelligence development, along with its key characteristics. Through the completion of the practical task, the era’s primitive, rule-based and relatively limited knowledge of AI compared to today’s advanced chatbots has become apparent.
Main take-aways:
- The innovations of the second era were crucial.
- Numerous algorithms, that were indispensable in later periods, were developed.
- Handling uncertainty and non-linear problems is challenging for artificial intelligence, leading to further research in subsequent eras.
Brief introduction of the next block:
In the following module, we will learn about the decline of artificial intelligence (1980s – 1990s). This period is also known as the “AI Winter.”
