From: "Jeffrey E." <jeevacation@gmail.com>
To: Bamaby Marsh "Nowak, Martin" < >,
Joichi Ito
Subject: this seems closer to an AI question .
Date: Tue, 17 May 2016 09:05:03 +0000
IT is part of the AI algorithm used in the austrialian quantum experiemnt. it
The default ramp is a linear sweep from SweepMinBias to SweepMaxBias. If this does not suit your needs you
can actually completely customize the ramps function using:
SweepRampFunction = (lambda function which takes one integer and returns a float between 0.0 and 1.0)
Under the hood lambda functions are actually what are used to implement default ramp. Here we let the user
directly make such a function. This is mostly included just to show off how flexible python is, and how clever
the input method we are using is. If you use this trick make sure the lambda function behaves sanely. It will
check it with a few random input integers, but that's it.
In high dimensional spaces the learners propensity to search the space, and particularly the boundaries might
become highly inefficient. If instead you want the learner to only randomly walk thought the search space
instead of completely freely probing it you can put it on a leash. In practice the leash puts bounds on the
minimizer when searching the predicted landscape based on the initial starting point it is given. This means there
is a maximum distance the learner can travel from any of the points it has previously visited. Using this approach
ensures that the learner can still in principle search the space, however it has to do so in a restricted manner.
The definition you need to add to turn the leash on is
LeashSize = (float between 0.0 and 1.0)
The leash size refers to the fraction of the difference between the minimum and maximum boundaries that the
controller can maximally move in each direction. By default the Leash is set to LeashSize = 1.0, basically
meaning it can go anywhere. But you can set it to be whatever length you want.
Much like the sweep function, if you instead want the leash to grow or shrink as a function of the run number
you can instead specify a LeashFunction:
LeashFunction = (lambda function which takes one integer and returns a float between 0.0 and 1.0
please note
The information contained in this communication is
confidential, may be attorney-client privileged, may
constitute inside information, and is intended only for
the use of the addressee. It is the property of
JEE
Unauthorized use, disclosure or copying of this
communication or any part thereof is strictly prohibited
and may be unlawful. If you have received this
communication in error, please notify us immediately by
return e-mail or by e-mail to jeevacation®gmail.com, and
destroy this communication and all copies thereof,
including all attachments. copyright -all rights reserved
EFTA00827325