DFA and NFA are types of finite automata used to recognize patterns in different applications like text editors and programming languages. Here are the key differences between DFA and NFA:

DFA (Deterministic Finite Automata):

• For each symbolic representation of the alphabet, there is only one state transition in DFA.
• DFA cannot use Empty String transition.
• DFA can be understood as one machine.
• In DFA, the next possible state is distinctly set.
• DFA is more difficult to construct.
• DFA rejects the string in case it terminates in a state that is different from the accepting state.
• Time needed for executing an input string is less.
• All DFA are NFA.

NFA (Nondeterministic Finite Automata):

• For a given state, on a given input, we can reach more than one state.
• NFA can use Empty String transition.
• NFA can be understood as multiple little machines computing at the same time.
• Each pair of state and input symbol can have many possible next states.
• NFA is easier to construct.
• NFA rejects the string in the event of all branches dying or refusing the string.
• Time needed for executing an input string is more.
• Need to convert NFA to DFA in the design of a compiler.

In terms of computational power, both DFA and NFA are equivalent as they both can recognize the exact same set of languages which are the regular languages. However, NFA can sometimes be more expressive and efficient than DFA.