next up previous
Next: Experimental Results for Both Up: Experiments: 60 Successive Tasks Previous: Incremental Learning: First Solve


All of the above was implemented by a dozen pages of code written in C, mostly comments and documentation: Multitasking and storage management through an iterative variant of round robin Try (Section 4.1); interpreter and 62 basic instructions (Section A); simple user interface for complex declarations (Section A.3); applications to $1^n2^n$-problems (Section 6.3) and Hanoi problems (Section 6.2). The current nonoptimized implementation considers between one and two million discrete unit time steps per second on an off-the-shelf PC (1.5 GHz).

Juergen Schmidhuber 2004-04-15

Back to OOPS main page