old
/
old-website
2
0
Fork 0
Code Issues 2 Pull Requests 1 Projects Releases Wiki Activity

Adding Elyot talk

This commit is contained in:
Brennan Taylor 2010-07-17 12:34:34 -04:00
parent 4068b22e88
commit e861f19887
1 changed files with 38 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
events.xml
View File

@ -4,6 +4,44 @@
<eventdefs>
<!-- Spring 2010 -->
<eventitem date="2010-07-20" time="04:30 PM" room="MC2066" title="The Incompressibility Method">
<short>
In this talk, we shall explore the incompressibility method---an interesting and
extremely powerful framework for determining the average-case runtime of
algorithms. Within the right background knowledge, the heapsort question can be
answered with an elegant 3-line proof.
</short>
<abstract>
<p>Heapsort. It runs in $\Theta(n \log n)$ time in the worst case, and in $O(n)$
time in the best case. Do you think that heapsort runs faster than $O(n
\log n)$ time on average? Could it be possible that on most inputs,
heapsort runs in $O(n)$ time, running more slowly only on a small fraction
of inputs?</p>
<p>Most students would say no. It "feels" intuitively obvious that heapsort
should take the full $n \log n$ steps on most inputs. However, proving this
rigourously with probabilistic arguments turns out to be very difficult.
Average case analysis of algorithms is one of those icky subjects that most
students don't want to touch with a ten foot pole; why should it be so
difficult if it is so intuitively obvious?</p>
<p>In this talk, we shall explore the incompressibility method---an interesting
and extremely powerful framework for determining the average-case runtime of
algorithms. Within the right background knowledge, the heapsort question
can be answered with an elegant 3-line proof.</p>
<p>The crucial fact is that an overwhelmingly large fraction of randomly
generated objects are incompressible. We can show that the inputs to
heapsort that run quickly correspond to inputs that can be compressed,
thereby proving that heapsort can't run quickly on average. Of course,
"compressible" is something that must be rigourously defined, and for this
we turn to the fascinating theory of Kolmogorov complexity.</p>
<p>In this talk, we'll briefly discuss the proof of the incompressibility
theorem and then see a number of applications. We won't dwell too much on
gruesome mathemtical details. No specific background is required, but
knowledge of some of the topics in CS240 will be helpful in understanding
some of the applications.</p>
</abstract>
</eventitem>
<eventitem date="2010-07-13" time="04:30 PM" room="MC2066" title="Halftoning and Digital Art">
<short><p>Edgar Bering will be giving a talk titled: Halftoning and Digital Art</p></short>
<abstract><p>Halftoning is the process of simulating a continuous tone image