exponentiating operators

So, it is a common thing to apply the same operator over and over again.
say:
op op op op op |x>
So that motivates a short cut:
op^5 |x>

Let's give an example using a simple binary tree:
left |x> => |0>
right |x> => |1>
child |x> => |0> + |1>

left |0> => |00>
right |0> => |10>
child |0> => |00> + |10>

left |1> => |01>
right |1> => |11>
child |1> => |01> + |11>

left |00> => |000>
right |00> => |100>
child |00> => |000> + |100>

left |10> => |010>
right |10> => |110>
child |10> => |010> + |110>

left |01> => |001>
right |01> => |101>
child |01> => |001> + |101>

left |11> => |011>
right |11> => |111>
child |11> => |011> + |111>

Now, lets use it:
sa: child |x>
|0> + |1>

-- without using the notational short-cut:
sa: child child |x>
|00> + |10> + |01> + |11>

-- using it:
sa: child^2 |x>
|00> + |10> + |01> + |11>

sa: child^3 |x>
|000> + |100> + |010> + |110> + |001> + |101> + |011> + |111>

sa: child^4 |x>
|>
-- we are past the bottom of the tree!

Note that child^k only gives you horizontal slices through the tree.
What if you want a vertical component too?
We can do that.

First recall from maths: exp(x)
And that exp(x) has the Taylor Series:
exp(x) = 1 + x + x^2/2 + x^3/3! + x^4/4! + x^5/5! + ...

If we tidy that up (we don't need the n! denominators), we have:
exp[op,n] |x>
is equivalent to:
(1 + op + op^2 + op^3 + ... + op^n) |x>

eg:
-- equivalent to:
-- |x> + child |x> + child^2 |x>
sa: exp[child,2] |x>
|x> + |0> + |1> + |00> + |10> + |01> + |11>

sa: exp[child,3] |x>
|x> + |0> + |1> + |00> + |10> + |01> + |11> + |000> + |100> + |010> + |110> + |001> + |101> + |011> + |111>

sa: exp[child,2] |1>
|1> + |01> + |11> + |001> + |101> + |011> + |111>

sa: exp[left,2] |1>
|1> + |01> + |001>

Finally, what if you want everything, but you don't know how deep to go?
Hence:
exp-max[op] |x>
which is equivalent to:
exp[op,n] |x> such that exp[op,n] |x> == exp[op,n+1] |x>

eg:
sa: exp-max[child] |x>
|x> + |0> + |1> + |00> + |10> + |01> + |11> + |000> + |100> + |010> + |110> + |001> + |101> + |011> + |111>

sa: exp-max[child] |0>
|0> + |00> + |10> + |000> + |100> + |010> + |110>

Finally, the idea of "six degrees of separation"
If we had the data, then eg:
sa: friends^6 |Fred>
sa: exp[friends,6] |Fred>
sa: exp-max[friends] |Fred>

That's more than enough for now. More later!

Update: now with the magic of tables:
sa: load binary-tree-with-child.sw
sa: table[node,child,text] exp-max[child] |x>
+------+----------+-------------------+
| node | child    | text              |
+------+----------+-------------------+
| x    | 0, 1     | start node        |
| 0    | 00, 10   | first child node  |
| 1    | 01, 11   | second child node |
| 00   | 000, 100 | third child node  |
| 10   | 010, 110 | fourth child node |
| 01   | 001, 101 | fifth child node  |
| 11   | 011, 111 | sixth child node  |
| 000  |          |                   |
| 100  |          |                   |
| 010  |          |                   |
| 110  |          |                   |
| 001  |          |                   |
| 101  |          |                   |
| 011  |          |                   |
| 111  |          |                   |
+------+----------+-------------------+



Home
previous: linearity of operators
next: learning indirectly

updated: 19/12/2016
by Garry Morrison
email: garry -at- semantic-db.org