There are 2 more that I would add, both books I read as an undergraduate nearly 60 years ago. First, Marvin Minsky's Computation: Finite And Infinite Machines (published 60 years ago). Not the latest word in automata theory, but as a first year student, I could read it and understand the issues he was addressing. Second, reading the First Edition of v.1 of The Art of Computer Programming (which of course Knuth has revised several times) was what convinced me that there was a difference between computer programming and computer science. Needless to say, there was much in TAOCP that I didn't understand, but showing me the relationship between computer programming, computer science, and mathematics really did change my life.