Chapter 2: Connectionist vs. Symbolic AI, Darthmouth Workshop Flashcards
When was the The McCulloch & Pitts artificial neuron introduced and what is it exactly?
The McCulloch & Pitts artificial neuron was introduced in 1943. It is a mathematical model of a biological neuron that can be used to create artificial neural networks. The model consists of a set of inputs and a binary output, and it uses a threshold function to determine the output based on the input values and the weights assigned to each input.
Explain the components of a BTAN.
The BTAN consists of a set of binary inputs (i.e., either 0 or 1), a binary output (i.e., either 0 or 1), and a set of weights (i.e., either -1, or 1). The threshold Activation a is weighted sum of inputs, a = SUM(weight(i) * input(i))
Draw a BTAN that can represent a XOR
2 inputs, 2 neuron as intermediate layer, with each neuron connecting to both inputs with weight 1. One of the neurons has t = 1, and the other t = 2. The neuron with t = 1 is connected to the output neuron with weight 1, the neuron with t = 2 is connected to an intermediate neuron with weight -1. This intermediate neuron has t = 0. The intermediate neuron connects to the output neuron with weight 1. The threshold of the output neuron is t = 2.
The neuron with t = 1 is technically a logical OR gate. The neurons with t = 2 are technically logical AND gates. The neuron with t = 0 is a logical NOT gate.
Explain symbolic AI
Symbolic AI performs explicit knowledge representation by the use of symbols. Relations can be encoded using symbol structures/expressions. Reasoning is resolving expressions into conclusions using rules. First-order predicate logic is an example, where the different expressions and symbols are called knowledge. Reduces reasoning to a search problem.
When was Whitehead & Russell’s Principia Mathematica introduced, and what does it entail?
Tried to ground foundation of mathematics using formal logic. They proved 1 + 1 = 2 in there :O
What is the logic theorist, by who and when was it created?
Newell, Simon, Shaw: Logic Theorist (1956), was considered the first AI program and encoded axioms and rules from Principia Mathematica. Managed to prove 38 of the 52 theorems of Chap. 2 of PM. This was a clear example of symbolic AI.
What is arguably the first programmable computer and when was it created? Who was the first to create programs for it?
Analytical engine, created in 1837. Lady Ada Lovelace (1815-1852) developed programs for it and is often called history’s
first computer programmer.
When was the Turing machine invented?
1936
What were the first uses for computers?
First uses were often numerical and for military purposes, like code-breaking, ballistic trajectories, computations for hydrogen bombs. Later on non-numerical problems were explored like chess (Shannon, 1950) and logic (Newel & Simon, 1955)
Who were present at the darthmouth workshop, give their names and what they are known for.
Shannon: Computer chess
Marvin Minsky: Artificial neural network
Nathaniel Rochester: chief architect of the IBM 701, the first mass produced scientific computer
John McCarthy: Created the Lisp programming language.
Who were present at the Darthmouth workshop, give their names and what they are known for.
Shannon: Computer chess
Marvin Minsky: Artificial neural network
Nathaniel Rochester: chief architect of the IBM 701, the first mass produced scientific computer
John McCarthy: Lisp
What were the main goals of the Darthmouth workshops, when was it held and for how long?
An attempt was made to find how to make machines use natural language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. The workshops were organized in 1956, for a summer long.
