750 lines
28 KiB
HTML
Executable File
750 lines
28 KiB
HTML
Executable File
<?xml version="1.0" encoding="utf-8"?>
|
||
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
|
||
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
|
||
<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">
|
||
<head>
|
||
<!-- 2025-12-28 Sun 20:49 -->
|
||
<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
|
||
<meta name="viewport" content="width=device-width, initial-scale=1" />
|
||
<title>ise_week_2</title>
|
||
<meta name="generator" content="Org Mode" />
|
||
<style type="text/css">
|
||
#content { max-width: 60em; margin: auto; }
|
||
.title { text-align: center;
|
||
margin-bottom: .2em; }
|
||
.subtitle { text-align: center;
|
||
font-size: medium;
|
||
font-weight: bold;
|
||
margin-top:0; }
|
||
.todo { font-family: monospace; color: red; }
|
||
.done { font-family: monospace; color: green; }
|
||
.priority { font-family: monospace; color: orange; }
|
||
.tag { background-color: #eee; font-family: monospace;
|
||
padding: 2px; font-size: 80%; font-weight: normal; }
|
||
.timestamp { color: #bebebe; }
|
||
.timestamp-kwd { color: #5f9ea0; }
|
||
.org-right { margin-left: auto; margin-right: 0px; text-align: right; }
|
||
.org-left { margin-left: 0px; margin-right: auto; text-align: left; }
|
||
.org-center { margin-left: auto; margin-right: auto; text-align: center; }
|
||
.underline { text-decoration: underline; }
|
||
#postamble p, #preamble p { font-size: 90%; margin: .2em; }
|
||
p.verse { margin-left: 3%; }
|
||
pre {
|
||
border: 1px solid #e6e6e6;
|
||
border-radius: 3px;
|
||
background-color: #f2f2f2;
|
||
padding: 8pt;
|
||
font-family: monospace;
|
||
overflow: auto;
|
||
margin: 1.2em;
|
||
}
|
||
pre.src {
|
||
position: relative;
|
||
overflow: auto;
|
||
}
|
||
pre.src:before {
|
||
display: none;
|
||
position: absolute;
|
||
top: -8px;
|
||
right: 12px;
|
||
padding: 3px;
|
||
color: #555;
|
||
background-color: #f2f2f299;
|
||
}
|
||
pre.src:hover:before { display: inline; margin-top: 14px;}
|
||
/* Languages per Org manual */
|
||
pre.src-asymptote:before { content: 'Asymptote'; }
|
||
pre.src-awk:before { content: 'Awk'; }
|
||
pre.src-authinfo::before { content: 'Authinfo'; }
|
||
pre.src-C:before { content: 'C'; }
|
||
/* pre.src-C++ doesn't work in CSS */
|
||
pre.src-clojure:before { content: 'Clojure'; }
|
||
pre.src-css:before { content: 'CSS'; }
|
||
pre.src-D:before { content: 'D'; }
|
||
pre.src-ditaa:before { content: 'ditaa'; }
|
||
pre.src-dot:before { content: 'Graphviz'; }
|
||
pre.src-calc:before { content: 'Emacs Calc'; }
|
||
pre.src-emacs-lisp:before { content: 'Emacs Lisp'; }
|
||
pre.src-fortran:before { content: 'Fortran'; }
|
||
pre.src-gnuplot:before { content: 'gnuplot'; }
|
||
pre.src-haskell:before { content: 'Haskell'; }
|
||
pre.src-hledger:before { content: 'hledger'; }
|
||
pre.src-java:before { content: 'Java'; }
|
||
pre.src-js:before { content: 'Javascript'; }
|
||
pre.src-latex:before { content: 'LaTeX'; }
|
||
pre.src-ledger:before { content: 'Ledger'; }
|
||
pre.src-lisp:before { content: 'Lisp'; }
|
||
pre.src-lilypond:before { content: 'Lilypond'; }
|
||
pre.src-lua:before { content: 'Lua'; }
|
||
pre.src-matlab:before { content: 'MATLAB'; }
|
||
pre.src-mscgen:before { content: 'Mscgen'; }
|
||
pre.src-ocaml:before { content: 'Objective Caml'; }
|
||
pre.src-octave:before { content: 'Octave'; }
|
||
pre.src-org:before { content: 'Org mode'; }
|
||
pre.src-oz:before { content: 'OZ'; }
|
||
pre.src-plantuml:before { content: 'Plantuml'; }
|
||
pre.src-processing:before { content: 'Processing.js'; }
|
||
pre.src-python:before { content: 'Python'; }
|
||
pre.src-R:before { content: 'R'; }
|
||
pre.src-ruby:before { content: 'Ruby'; }
|
||
pre.src-sass:before { content: 'Sass'; }
|
||
pre.src-scheme:before { content: 'Scheme'; }
|
||
pre.src-screen:before { content: 'Gnu Screen'; }
|
||
pre.src-sed:before { content: 'Sed'; }
|
||
pre.src-sh:before { content: 'shell'; }
|
||
pre.src-sql:before { content: 'SQL'; }
|
||
pre.src-sqlite:before { content: 'SQLite'; }
|
||
/* additional languages in org.el's org-babel-load-languages alist */
|
||
pre.src-forth:before { content: 'Forth'; }
|
||
pre.src-io:before { content: 'IO'; }
|
||
pre.src-J:before { content: 'J'; }
|
||
pre.src-makefile:before { content: 'Makefile'; }
|
||
pre.src-maxima:before { content: 'Maxima'; }
|
||
pre.src-perl:before { content: 'Perl'; }
|
||
pre.src-picolisp:before { content: 'Pico Lisp'; }
|
||
pre.src-scala:before { content: 'Scala'; }
|
||
pre.src-shell:before { content: 'Shell Script'; }
|
||
pre.src-ebnf2ps:before { content: 'ebfn2ps'; }
|
||
/* additional language identifiers per "defun org-babel-execute"
|
||
in ob-*.el */
|
||
pre.src-cpp:before { content: 'C++'; }
|
||
pre.src-abc:before { content: 'ABC'; }
|
||
pre.src-coq:before { content: 'Coq'; }
|
||
pre.src-groovy:before { content: 'Groovy'; }
|
||
/* additional language identifiers from org-babel-shell-names in
|
||
ob-shell.el: ob-shell is the only babel language using a lambda to put
|
||
the execution function name together. */
|
||
pre.src-bash:before { content: 'bash'; }
|
||
pre.src-csh:before { content: 'csh'; }
|
||
pre.src-ash:before { content: 'ash'; }
|
||
pre.src-dash:before { content: 'dash'; }
|
||
pre.src-ksh:before { content: 'ksh'; }
|
||
pre.src-mksh:before { content: 'mksh'; }
|
||
pre.src-posh:before { content: 'posh'; }
|
||
/* Additional Emacs modes also supported by the LaTeX listings package */
|
||
pre.src-ada:before { content: 'Ada'; }
|
||
pre.src-asm:before { content: 'Assembler'; }
|
||
pre.src-caml:before { content: 'Caml'; }
|
||
pre.src-delphi:before { content: 'Delphi'; }
|
||
pre.src-html:before { content: 'HTML'; }
|
||
pre.src-idl:before { content: 'IDL'; }
|
||
pre.src-mercury:before { content: 'Mercury'; }
|
||
pre.src-metapost:before { content: 'MetaPost'; }
|
||
pre.src-modula-2:before { content: 'Modula-2'; }
|
||
pre.src-pascal:before { content: 'Pascal'; }
|
||
pre.src-ps:before { content: 'PostScript'; }
|
||
pre.src-prolog:before { content: 'Prolog'; }
|
||
pre.src-simula:before { content: 'Simula'; }
|
||
pre.src-tcl:before { content: 'tcl'; }
|
||
pre.src-tex:before { content: 'TeX'; }
|
||
pre.src-plain-tex:before { content: 'Plain TeX'; }
|
||
pre.src-verilog:before { content: 'Verilog'; }
|
||
pre.src-vhdl:before { content: 'VHDL'; }
|
||
pre.src-xml:before { content: 'XML'; }
|
||
pre.src-nxml:before { content: 'XML'; }
|
||
/* add a generic configuration mode; LaTeX export needs an additional
|
||
(add-to-list 'org-latex-listings-langs '(conf " ")) in .emacs */
|
||
pre.src-conf:before { content: 'Configuration File'; }
|
||
|
||
table { border-collapse:collapse; }
|
||
caption.t-above { caption-side: top; }
|
||
caption.t-bottom { caption-side: bottom; }
|
||
td, th { vertical-align:top; }
|
||
th.org-right { text-align: center; }
|
||
th.org-left { text-align: center; }
|
||
th.org-center { text-align: center; }
|
||
td.org-right { text-align: right; }
|
||
td.org-left { text-align: left; }
|
||
td.org-center { text-align: center; }
|
||
dt { font-weight: bold; }
|
||
.footpara { display: inline; }
|
||
.footdef { margin-bottom: 1em; }
|
||
.figure { padding: 1em; }
|
||
.figure p { text-align: center; }
|
||
.equation-container {
|
||
display: table;
|
||
text-align: center;
|
||
width: 100%;
|
||
}
|
||
.equation {
|
||
vertical-align: middle;
|
||
}
|
||
.equation-label {
|
||
display: table-cell;
|
||
text-align: right;
|
||
vertical-align: middle;
|
||
}
|
||
.inlinetask {
|
||
padding: 10px;
|
||
border: 2px solid gray;
|
||
margin: 10px;
|
||
background: #ffffcc;
|
||
}
|
||
#org-div-home-and-up
|
||
{ text-align: right; font-size: 70%; white-space: nowrap; }
|
||
textarea { overflow-x: auto; }
|
||
.linenr { font-size: smaller }
|
||
.code-highlighted { background-color: #ffff00; }
|
||
.org-info-js_info-navigation { border-style: none; }
|
||
#org-info-js_console-label
|
||
{ font-size: 10px; font-weight: bold; white-space: nowrap; }
|
||
.org-info-js_search-highlight
|
||
{ background-color: #ffff00; color: #000000; font-weight: bold; }
|
||
.org-svg { }
|
||
</style>
|
||
|
||
<link rel="stylesheet" href="/assets/styles/style.css" />
|
||
<link rel="stylesheet" href="/assets/styles/bigger-picture.min.css" />
|
||
<link rel="stylesheet" href="/assets/styles/media.css" />
|
||
|
||
<script src="/assets/scripts/script.js" defer></script>
|
||
<script src="/assets/scripts/search.js" defer></script>
|
||
<script src="/assets/scripts/bigger-picture.min.js" defer></script>
|
||
<script src="/assets/scripts/svg-pan-zoom.min.js" defer></script>
|
||
<script src="/assets/scripts/gallery-init.js" defer></script>
|
||
</head>
|
||
<body>
|
||
<div id="preamble" class="status">
|
||
|
||
<div class="banner-header">
|
||
<div class="banner-left">
|
||
<a href="/20241210233721-brain_moc.html"> <img src="/assets/gr.png" alt="Site Logo" class="banner-logo" /> </a>
|
||
<div id="updated">Updated: 2025-12-28 Sun 19:48</div>
|
||
</div>
|
||
|
||
<div class="banner-search">
|
||
<input id="search-box"
|
||
type="search"
|
||
placeholder="Search notes…"
|
||
aria-label="Search notes"
|
||
autocomplete="off" />
|
||
<div id="search-results"></div>
|
||
</div>
|
||
|
||
<button id="close-all">Close All</button>
|
||
</div>
|
||
</div>
|
||
<div id="stack-root">
|
||
<div class="stack-track">
|
||
<article class="stack-pane pane-root" data-url="/20250329121843-ise_week_2.html">
|
||
<div id="content" class="content">
|
||
<div class="title-section">
|
||
<div class="title-controls">
|
||
<button class="pane-fullscreen" aria-label="Fullscreen">F</button>
|
||
<button class="pane-edit" aria-label="Edit pane">E</button>
|
||
<button class="pane-close" aria-label="Close pane">×</button>
|
||
</div>
|
||
<h1 class="title">ise<sub>week</sub><sub>2</sub></h1>
|
||
<div class="title-metadata">
|
||
<span class="metadata-item">
|
||
<span class="metadata-label">planted:</span>
|
||
<span class="metadata-value">2025-03-29</span>
|
||
</span>
|
||
<span class="metadata-item">
|
||
<span class="metadata-label">last tended to:</span>
|
||
<span class="metadata-value">2025-12-28</span>
|
||
</span>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-orgac78bd8" class="outline-2">
|
||
<h2 id="orgac78bd8"><a href="#orgac78bd8"><span class="done DONE">DONE</span> Different Configuration Sampling Methods (pages 8-19)</a></h2>
|
||
</div>
|
||
<div id="outline-container-orgfc73a5c" class="outline-2">
|
||
<h2 id="orgfc73a5c"><a href="#orgfc73a5c"><span class="done DONE">DONE</span> Configuration Encodings (pages 21-28)</a></h2>
|
||
</div>
|
||
<div id="outline-container-org09131b6" class="outline-2">
|
||
<h2 id="org09131b6"><a href="#org09131b6"><span class="done DONE">DONE</span> Single Environment Learning (DaL) (pages 39-46)</a></h2>
|
||
</div>
|
||
|
||
<div id="outline-container-orgff77265" class="outline-2">
|
||
<h2 id="orgff77265"><a href="#orgff77265">2.1</a></h2>
|
||
<div class="outline-text-2" id="text-orgff77265">
|
||
</div>
|
||
<div id="outline-container-org210abb7" class="outline-3">
|
||
<h3 id="org210abb7"><a href="#org210abb7">Configuration Sampling</a></h3>
|
||
<div class="outline-text-3" id="text-org210abb7">
|
||
<p>
|
||
In general machine learning problem, we dont care where the data comes from , but here we do.
|
||
</p>
|
||
|
||
<p>
|
||
Configuration sampling is used to select representative samples for learning performance models.
|
||
</p>
|
||
|
||
<ul class="org-ul">
|
||
<li>Types of options:
|
||
<ul class="org-ul">
|
||
<li>Binary (e.g., on/off)</li>
|
||
<li>Numeric (e.g., value ranges)</li>
|
||
</ul></li>
|
||
<li>Goal: Balance model accuracy with sampling effort.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org34412d1" class="outline-3">
|
||
<h3 id="org34412d1"><a href="#org34412d1">Binary Sampling Strategies</a></h3>
|
||
<div class="outline-text-3" id="text-org34412d1">
|
||
</div>
|
||
<div id="outline-container-org0bf1908" class="outline-4">
|
||
<h4 id="org0bf1908"><a href="#org0bf1908">Option-wise Strategy</a></h4>
|
||
<div class="outline-text-4" id="text-org0bf1908">
|
||
<ul class="org-ul">
|
||
<li>Each binary option is selected at least once in some configuration.</li>
|
||
<li>Minimize other options to reduce unknown interaction effects.</li>
|
||
<li>Size: Linear in the number of binary options.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org60c6569" class="outline-4">
|
||
<h4 id="org60c6569"><a href="#org60c6569">T-wise Strategy</a></h4>
|
||
<div class="outline-text-4" id="text-org60c6569">
|
||
<ul class="org-ul">
|
||
<li>Covers all T-wise combinations of options (T ≥ 2).</li>
|
||
<li>Example (2-wise): {001}, {010}, {100}, {111}</li>
|
||
<li>Size: Exponential in T.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org80959ab" class="outline-4">
|
||
<h4 id="org80959ab"><a href="#org80959ab">Negative Option-wise Strategy</a></h4>
|
||
<div class="outline-text-4" id="text-org80959ab">
|
||
<ul class="org-ul">
|
||
<li>For each option: one configuration where it is disabled, all others enabled.</li>
|
||
<li>Adds one all-yes configuration.</li>
|
||
<li>Size: Linear.</li>
|
||
<li>Example (3 options): {110}, {101}, {011}, {111}</li>
|
||
<li>you can see the 4th one is an all-yes configuration</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-orgf8338c9" class="outline-4">
|
||
<h4 id="orgf8338c9"><a href="#orgf8338c9">Random (Binary)</a></h4>
|
||
<div class="outline-text-4" id="text-orgf8338c9">
|
||
<ul class="org-ul">
|
||
<li>Select n configurations randomly.</li>
|
||
<li>Simple but may be less representative.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org132588a" class="outline-4">
|
||
<h4 id="org132588a"><a href="#org132588a">Difference Between Option-wise and Negative Option-wise Strategies</a></h4>
|
||
<div class="outline-text-4" id="text-org132588a">
|
||
<p>
|
||
Both strategies are used for sampling configurations in systems with binary options, but they focus on different aspects of option selection.
|
||
</p>
|
||
</div>
|
||
<ul class="org-ul">
|
||
<li><a id="org78494e1"></a><a href="#org78494e1">Option-wise Strategy</a><br />
|
||
<div class="outline-text-5" id="text-org78494e1">
|
||
<ul class="org-ul">
|
||
<li><b><b>Goal:</b></b> Ensure each option is enabled (selected) at least once across configurations.</li>
|
||
<li>For every binary option, create a configuration where it is <b><b>on</b></b>.</li>
|
||
<li>Other options are minimized to avoid unknown interactions.</li>
|
||
<li><b><b>Focus:</b></b> Testing the <b><b>presence</b></b> of each option.</li>
|
||
<li><b><b>Example (3 options):</b></b>
|
||
<ul class="org-ul">
|
||
<li>{100} → Option 1 enabled, others off</li>
|
||
<li>{010} → Option 2 enabled, others off</li>
|
||
<li>{001} → Option 3 enabled, others off</li>
|
||
</ul></li>
|
||
</ul>
|
||
</div>
|
||
</li>
|
||
<li><a id="orga8add10"></a><a href="#orga8add10">Negative Option-wise Strategy</a><br />
|
||
<div class="outline-text-5" id="text-orga8add10">
|
||
<ul class="org-ul">
|
||
<li><b><b>Goal:</b></b> Ensure each option is disabled (deselected) at least once.</li>
|
||
<li>For each option, create a configuration where it is <b><b>off</b></b>, and <b><b>all others are on</b></b>.</li>
|
||
<li>Also includes a configuration where all options are <b><b>on</b></b>.</li>
|
||
<li><b><b>Focus:</b></b> Testing the <b><b>absence</b></b> of each option.</li>
|
||
<li><b><b>Example (3 options):</b></b>
|
||
<ul class="org-ul">
|
||
<li>{110} → Option 3 disabled</li>
|
||
<li>{101} → Option 2 disabled</li>
|
||
<li>{011} → Option 1 disabled</li>
|
||
<li>{111} → All options enabled</li>
|
||
</ul></li>
|
||
</ul>
|
||
</div>
|
||
</li>
|
||
<li><a id="org4e8e372"></a><a href="#org4e8e372">Comparison Summary</a><br />
|
||
<div class="outline-text-5" id="text-org4e8e372">
|
||
<table border="2" cellspacing="0" cellpadding="6" rules="groups" frame="hsides">
|
||
|
||
|
||
<colgroup>
|
||
<col class="org-left" />
|
||
|
||
<col class="org-left" />
|
||
|
||
<col class="org-left" />
|
||
</colgroup>
|
||
<thead>
|
||
<tr>
|
||
<th scope="col" class="org-left">Feature</th>
|
||
<th scope="col" class="org-left">Option-wise</th>
|
||
<th scope="col" class="org-left">Negative Option-wise</th>
|
||
</tr>
|
||
</thead>
|
||
<tbody>
|
||
<tr>
|
||
<td class="org-left">Focus</td>
|
||
<td class="org-left">Presence of each option</td>
|
||
<td class="org-left">Absence of each option</td>
|
||
</tr>
|
||
|
||
<tr>
|
||
<td class="org-left">What is varied</td>
|
||
<td class="org-left">Each option enabled once</td>
|
||
<td class="org-left">Each option disabled once</td>
|
||
</tr>
|
||
|
||
<tr>
|
||
<td class="org-left">Other options in config</td>
|
||
<td class="org-left">Typically disabled</td>
|
||
<td class="org-left">Typically enabled</td>
|
||
</tr>
|
||
|
||
<tr>
|
||
<td class="org-left">Additional config?</td>
|
||
<td class="org-left">Not required</td>
|
||
<td class="org-left">Yes, includes all-on config</td>
|
||
</tr>
|
||
|
||
<tr>
|
||
<td class="org-left">Use case</td>
|
||
<td class="org-left">Minimal presence testing</td>
|
||
<td class="org-left">Influence of removing options</td>
|
||
</tr>
|
||
</tbody>
|
||
</table>
|
||
</div>
|
||
</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org8304e90" class="outline-3">
|
||
<h3 id="org8304e90"><a href="#org8304e90">Non-Binary (Numeric) Sampling Strategies</a></h3>
|
||
<div class="outline-text-3" id="text-org8304e90">
|
||
</div>
|
||
<div id="outline-container-org13877de" class="outline-4">
|
||
<h4 id="org13877de"><a href="#org13877de">One-Factor-At-A-Time (OFAT)</a></h4>
|
||
<div class="outline-text-4" id="text-org13877de">
|
||
<ul class="org-ul">
|
||
<li>Assumes no interactions among options.</li>
|
||
<li>Varies one option at a time, others fixed at center values.</li>
|
||
<li>Size: Linear in number of options.</li>
|
||
<li>Example (values = 1,3,5): {333}, {533}, {133}, {353}, {313}, {331}, {335}</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org3df11fe" class="outline-4">
|
||
<h4 id="org3df11fe"><a href="#org3df11fe">Box-Behnken Design (BBD)</a></h4>
|
||
<div class="outline-text-4" id="text-org3df11fe">
|
||
<ul class="org-ul">
|
||
<li>Captures quadratic effects and 2-wise interactions.</li>
|
||
<li>Uses subset of 3<sup>k</sup> full factorial (min, center, max).</li>
|
||
<li>Size: Exponential in number of options.</li>
|
||
<li>Example: {111}, {113}, {115}, {131}, {151}, etc.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-orgf54868e" class="outline-4">
|
||
<h4 id="orgf54868e"><a href="#orgf54868e">Central Composite Design (CCD)</a></h4>
|
||
<div class="outline-text-4" id="text-orgf54868e">
|
||
<ul class="org-ul">
|
||
<li>Combines:
|
||
<ul class="org-ul">
|
||
<li>2<sup>k</sup> factorial points</li>
|
||
<li>2k axial points at α-distance</li>
|
||
<li>1 center point</li>
|
||
</ul></li>
|
||
<li>Captures curvature and interactions.</li>
|
||
<li>Example: 8 full factorial + 6 axial + {333}</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org955eacd" class="outline-4">
|
||
<h4 id="org955eacd"><a href="#org955eacd">Plackett-Burman Design (PBD)</a></h4>
|
||
<div class="outline-text-4" id="text-org955eacd">
|
||
<ul class="org-ul">
|
||
<li>Focus on main effects, assumes negligible interactions.</li>
|
||
<li>Uses predefined seeds, e.g., PBD(9,3)</li>
|
||
<li>First config from seed, rest by right-shifting seed.</li>
|
||
<li>Uses indices only for values.</li>
|
||
<li>Example: If O = {1,100,1000,10000,100000}, index 3 could mean 1, 1000, 100000</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-orge585105" class="outline-4">
|
||
<h4 id="orge585105"><a href="#orge585105">Random (Non-Binary)</a></h4>
|
||
<div class="outline-text-4" id="text-orge585105">
|
||
<ul class="org-ul">
|
||
<li>Random selection of numeric configurations.</li>
|
||
<li>Risk of non-uniformity and clustering.</li>
|
||
<li>Can negatively impact learning performance.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org524185d" class="outline-3">
|
||
<h3 id="org524185d"><a href="#org524185d">Mixed Variable Sampling</a></h3>
|
||
<div class="outline-text-3" id="text-org524185d">
|
||
<p>
|
||
Some systems include both binary and non-binary (numeric) configuration options.
|
||
These are referred to as <b><b>mixed systems</b></b>.
|
||
</p>
|
||
|
||
<ul class="org-ul">
|
||
<li>Requires hybrid or combined strategies to ensure representative coverage.</li>
|
||
<li>One approach: <b><b>Permute over the mixed space</b></b> by combining possible binary and numeric value combinations.</li>
|
||
<li>This can grow combinatorially, so sampling techniques may be needed to reduce the total number of permutations.</li>
|
||
</ul>
|
||
</div>
|
||
<div id="outline-container-org66a3a49" class="outline-4">
|
||
<h4 id="org66a3a49"><a href="#org66a3a49">Example</a></h4>
|
||
<div class="outline-text-4" id="text-org66a3a49">
|
||
<ul class="org-ul">
|
||
<li>Non-binary configs: {0.1, 0.4, 5}, {0.2, 0.4, 7}, {0.2, 0.7, 5}</li>
|
||
<li>Binary configs: {1,0}, {1,1}</li>
|
||
<li>Full mixed permutations:
|
||
<ul class="org-ul">
|
||
<li>{0.1, 0.4, 5, 1, 0}</li>
|
||
<li>{0.1, 0.4, 5, 1, 1}</li>
|
||
<li>{0.2, 0.4, 7, 1, 0}</li>
|
||
<li>{0.2, 0.4, 7, 1, 1}</li>
|
||
<li>{0.2, 0.7, 5, 1, 0}</li>
|
||
<li>{0.2, 0.7, 5, 1, 1}</li>
|
||
</ul></li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org371117a" class="outline-2">
|
||
<h2 id="org371117a"><a href="#org371117a">2.2</a></h2>
|
||
<div class="outline-text-2" id="text-org371117a">
|
||
</div>
|
||
<div id="outline-container-orgbdd8e35" class="outline-3">
|
||
<h3 id="orgbdd8e35"><a href="#orgbdd8e35">Single Environment Learning: DeepPerf</a></h3>
|
||
<div class="outline-text-3" id="text-orgbdd8e35">
|
||
<p>
|
||
Source: Ha & Zhang, ICSE 2019
|
||
</p>
|
||
|
||
<p>
|
||
DeepPerf is an early approach using deep neural networks (>3 layers) to predict software performance in configurable systems.
|
||
</p>
|
||
|
||
<ul class="org-ul">
|
||
<li>Designed to address:
|
||
<ul class="org-ul">
|
||
<li>Small data size: Limited measurements available.</li>
|
||
<li>Feature sparsity: Only a few configuration options significantly impact performance.</li>
|
||
<li>Network instability: Tackled with tailored hyperparameter tuning.</li>
|
||
</ul></li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-orgace2d30" class="outline-3">
|
||
<h3 id="orgace2d30"><a href="#orgace2d30">Limitation of DeepPerf</a></h3>
|
||
<div class="outline-text-3" id="text-orgace2d30">
|
||
<ul class="org-ul">
|
||
<li>Does not handle sample sparsity, a major issue in configuration performance prediction.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org0a0ec40" class="outline-3">
|
||
<h3 id="org0a0ec40"><a href="#org0a0ec40">Improved Approach: Divide-and-Learn (DaL)</a></h3>
|
||
<div class="outline-text-3" id="text-org0a0ec40">
|
||
<p>
|
||
Source: Gong & Chen, ESEC/FSE 2023
|
||
</p>
|
||
</div>
|
||
<div id="outline-container-org212aebe" class="outline-4">
|
||
<h4 id="org212aebe"><a href="#org212aebe">Key Problem: Sample Sparsity</a></h4>
|
||
<div class="outline-text-4" id="text-org212aebe">
|
||
<ul class="org-ul">
|
||
<li>Caused by:
|
||
<ul class="org-ul">
|
||
<li>Inherited feature sparsity.</li>
|
||
<li>Small configuration changes leading to drastic performance shifts.</li>
|
||
<li>Not all configurations being valid.</li>
|
||
</ul></li>
|
||
<li>Training data is sparse due to expensive measurements.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org6be0e50" class="outline-4">
|
||
<h4 id="org6be0e50"><a href="#org6be0e50">Key Properties of Configuration Landscape</a></h4>
|
||
<div class="outline-text-4" id="text-org6be0e50">
|
||
<ol class="org-ol">
|
||
<li>Intra-division smoothness: Configurations in the same division show smooth performance variations.</li>
|
||
<li>Inter-division sharpness: Cross-division configurations differ significantly, possibly on key options.</li>
|
||
</ol>
|
||
|
||
<p>
|
||
Risk: Limited data might lead to overfitting within divisions.
|
||
</p>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org6872259" class="outline-4">
|
||
<h4 id="org6872259"><a href="#org6872259">Architecture of DaL</a></h4>
|
||
<div class="outline-text-4" id="text-org6872259">
|
||
<p>
|
||
Three Goals:
|
||
</p>
|
||
<ol class="org-ol">
|
||
<li>Divide the configuration data into meaningful divisions → function ϕ</li>
|
||
<li>Learn a local model for each division → function μ</li>
|
||
<li>Assign new configurations to the correct local model → using ϕ and μ</li>
|
||
|
||
<li>Implementation:
|
||
<ul class="org-ul">
|
||
<li>CART (Decision Tree) is used for dividing.</li>
|
||
<li>DeepPerf models are trained within each division.</li>
|
||
<li>Random Forest is used for classifying unseen configurations into divisions.</li>
|
||
</ul></li>
|
||
</ol>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org50a058d" class="outline-4">
|
||
<h4 id="org50a058d"><a href="#org50a058d">Trade-off: Number of Divisions</a></h4>
|
||
<div class="outline-text-4" id="text-org50a058d">
|
||
<ul class="org-ul">
|
||
<li>More divisions → better at tackling sparsity, but less data per model → risks underfitting.</li>
|
||
<li>Need to balance:
|
||
<ul class="org-ul">
|
||
<li>Generalizability vs.</li>
|
||
<li>Specialization</li>
|
||
</ul></li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org5ca0e35" class="outline-4">
|
||
<h4 id="org5ca0e35"><a href="#org5ca0e35">Results</a></h4>
|
||
<div class="outline-text-4" id="text-org5ca0e35">
|
||
<ul class="org-ul">
|
||
<li>DaL outperforms or matches state-of-the-art in 33 out of 40 cases.</li>
|
||
<li>Achieves up to 1.94× improvement.</li>
|
||
<li>Needs fewer training samples for same accuracy.</li>
|
||
<li>Especially beneficial in complex systems or with more training data.</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org92505b7" class="outline-2">
|
||
<h2 id="org92505b7"><a href="#org92505b7">2.3</a></h2>
|
||
<div class="outline-text-2" id="text-org92505b7">
|
||
</div>
|
||
<div id="outline-container-org7ae9d98" class="outline-3">
|
||
<h3 id="org7ae9d98"><a href="#org7ae9d98">Single Environment Learning: Encoding</a></h3>
|
||
<div class="outline-text-3" id="text-org7ae9d98">
|
||
<p>
|
||
Source: Gong & Chen, MSR 2022
|
||
</p>
|
||
|
||
<p>
|
||
A study conducted by the lab investigates how different encoding schemes impact the software performance learning pipeline.
|
||
</p>
|
||
</div>
|
||
<div id="outline-container-org4923f2e" class="outline-4">
|
||
<h4 id="org4923f2e"><a href="#org4923f2e">Three Common Encoding Schemes</a></h4>
|
||
<div class="outline-text-4" id="text-org4923f2e">
|
||
<ul class="org-ul">
|
||
<li>Label encoding</li>
|
||
<li>Scaled label encoding (e.g., max-min normalization)</li>
|
||
<li>One-hot encoding</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org5143990" class="outline-3">
|
||
<h3 id="org5143990"><a href="#org5143990">Encoding Schemes Explained</a></h3>
|
||
<div class="outline-text-3" id="text-org5143990">
|
||
</div>
|
||
<div id="outline-container-org54b9c21" class="outline-4">
|
||
<h4 id="org54b9c21"><a href="#org54b9c21">Label Encoding</a></h4>
|
||
<div class="outline-text-4" id="text-org54b9c21">
|
||
<ul class="org-ul">
|
||
<li>Converts configuration options into numeric values.</li>
|
||
<li>Example:
|
||
<ul class="org-ul">
|
||
<li>Configuration: (cache<sub>size</sub>, interval, ssl, data<sub>strategy</sub>)</li>
|
||
<li>Values: cache<sub>size</sub> = (1, 10, 10000), interval = (1–4), ssl = (0, 1), data<sub>strategy</sub> = (strategy<sub>1</sub>, strategy<sub>2</sub>, strategy<sub>3</sub>)</li>
|
||
<li>Encoded: (10000, 2, 1, 1) → (2, 1, 1, 1) → data<sub>strategy</sub>: (0, 1, 2)</li>
|
||
</ul></li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org11a624a" class="outline-4">
|
||
<h4 id="org11a624a"><a href="#org11a624a">Scaled Label Encoding</a></h4>
|
||
<div class="outline-text-4" id="text-org11a624a">
|
||
<ul class="org-ul">
|
||
<li>Similar to label encoding but normalizes all values to the range [0, 1].</li>
|
||
<li>Example (10000, 2, 1, 1) becomes (1, 1/3, 1, 0.5)</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org785447a" class="outline-4">
|
||
<h4 id="org785447a"><a href="#org785447a">One-Hot Encoding</a></h4>
|
||
<div class="outline-text-4" id="text-org785447a">
|
||
<ul class="org-ul">
|
||
<li>Transforms each categorical value into a binary vector.</li>
|
||
<li>Example: (10000, 2, 1, 1) becomes (0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0)</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-orge613ed8" class="outline-3">
|
||
<h3 id="orge613ed8"><a href="#orge613ed8">Community Debate and Justifications</a></h3>
|
||
<div class="outline-text-3" id="text-orge613ed8">
|
||
<ul class="org-ul">
|
||
<li>Categorical features (e.g., cache<sub>mode</sub> = memory, disk, mixed):
|
||
<ul class="org-ul">
|
||
<li>Label encoding implies false ordering (1, 2, 3)</li>
|
||
<li>One-hot encoding avoids this but may introduce multicollinearity.</li>
|
||
</ul></li>
|
||
|
||
<li>Numeric options (e.g., cache<sub>size</sub> = 1, 10, 10000):
|
||
<ul class="org-ul">
|
||
<li>Label encoding maintains order but struggles with large scale differences.</li>
|
||
<li>Scaled label encoding improves numeric stability but weakens interaction with binary features.</li>
|
||
</ul></li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div id="outline-container-org735fb45" class="outline-3">
|
||
<h3 id="org735fb45"><a href="#org735fb45">Study Protocol</a></h3>
|
||
<div class="outline-text-3" id="text-org735fb45">
|
||
<ul class="org-ul">
|
||
<li>Evaluated using 7 learning algorithms across 5 software systems.</li>
|
||
</ul>
|
||
|
||
|
||
|
||
<p class="backlinks-section" id="backlinks">
|
||
Backlinks
|
||
</p>
|
||
<ul class="org-ul backlinks-list">
|
||
<li><a href="20250329114733-ise.html#ID-c69e4c4d-2fb4-4cf1-a835-a235cf6db8e9">ise</a></li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
</div></article></div></div>
|
||
<div id="postamble" class="status">
|
||
<footer>
|
||
<div class="copyright-container">
|
||
<div class="copyright">
|
||
Copyright © 2022-2025 Zaine Qayyum. All rights reserved unless otherwise noted.</div></div>
|
||
<div class="generated">
|
||
Created with <a href="https://www.gnu.org/software/emacs/">Emacs</a> 30.2 (<a href="https://orgmode.org">Org</a> mode 9.7.11) on <a href="https://www.archlinux.org/">Arch</a> <a href="https://www.gnu.org">GNU</a>/<a href="https://www.kernel.org/">Linux</a>
|
||
</div>
|
||
</footer>
|
||
</div>
|
||
</body>
|
||
</html>
|