Previous Beer and Algorithm Talks



Date Person Topic Links/References
Summer 2009 Talks
Monday 6/08/09 Christine Task Approximation Algorithms Mainly CLRS Chapter 35, results for Vertex Cover and Traveling Salesman
Monday 6/22/09 Lindsey Kuper Relativization and Oracle Turing Machines Sipser (ch 6,9),    Lecture Notes
Monday 6/29/09 Mark Wilson Randomized Algorithms Randomized Algorithms by Motwani and Raghaven (Available free online for IU folks, through the IU library and Books 24x7)
Monday 7/6/09 Josh Bonner Quantum Computing Quantum Computer Science by N. David Mermin
Monday 7/13/09 Nate Skiba Cryptography RSA Link I     RSA Link II     ICSA Guide to Cryptography, by Randall K. Nichols
Monday 7/20/09 Matt Zurschmeide Image Compression Algorithms PNG, JPEG
Monday 7/27/09 Christine Task The Watchman Problem On the Paranoid Watchman Problem (the flawed masters thesis)
Monday 8/3/09 Andy Keep The Matching Problem (Applied) Basic Wikipedia Pages: 1 2 3 4, Journal Articles: 1 2 3 4, and That French Book by Knuth.
Monday 8/10/09 Dave Bender Volumetric Graphics Efficient Synthetic Image Generation of Arbitrary 3-D Objects, Geometric Modeling Using Octree Encoding
Monday 8/17/09 Brenton Bostick The 3n+1 Problem Surveys/Bibilographies: 1, 2, 3, Stats approach, New Entropy Convergence Idea
Monday 8/24/09 Josh Bonner Traffic Analysis (cars, not packets)



6/08/09: Approximation Algorithms
Presenter:   Christine Task
Location:   Runciple Spoon
Description:   Not all NP-Complete problems are created equal. Once we're pretty certain that we can't find a general, correct solution in polynomial time, we start to look at efficient solutions for special cases, or general solutions that deviate from the optimal answer by a predictable (preferably constant) factor. For some NP-complete problems we can can find good approximate solutions, but for others we provably cannot (unless P=NP, in which case it's all sort of a moot point). I talked about vertex cover (which falls into the first category) and the traveling salesman problem (which falls into the second).
References:   Mainly CLRS Chapter 35, results for Vertex Cover and Traveling Salesman


6/22/09: Relativization and Oracle Turing Machines
Presenter:   Lindsey Kuper
Location:   BuffaLouis
Description:   We know the halting problem is undecidable, but what if it weren't? What else would a decider for the halting problem let us do? What problems are decidable 'relative' to the halting problem? Similar questions can be asked about feasibility: if we could solve one problem efficiently (in polynomial time), what other problems could we solve efficiently?
References: Sipser (ch 6,9),    Lecture Notes


6/29/09: Randomized Algorithms
Presenter:   Mark Wilson
Location:   Upland Brewery (notes: too crowded)
Description:   Deterministic algorithms which perform acceptably in the average case are often susceptible to corner-case inputs which trigger unacceptable worst-case performance. (See, e.g., the secretary problem.) This weakness can be especially problematic if exploited by an adversary with control of the input.
Randomized algorithms, such as quicksort with randomly chosen pivots, can alleviate this weakness of deterministic algorithms. But how much can they affect performance? We will look at some models of randomization and the basics of analyzing random algorithms.
References:Randomized Algorithms by Motwani and Raghaven (Available free online for IU folks, through the IU library and Books 24x7)


7/06/09: Quantum Computing

Presenter:   Josh Bonner
Location:   Dat's Cajun Place (notes: no beer :-( )
Description:   [to be filled in]
References: Quantum Computer Science by N. David Mermin


7/13/09: Cryptography
Presenter:   Nate Skiba
Location:   Nick's
Description:   A brief history of cryptography and the issues that lead up to the development of such highly useful public key/private key systems as RSA, along with a clear description of just exactly why RSA can do what it does.
References: RSA Link I     RSA Link II     ICSA Guide to Cryptography, by Randall K. Nichols


7/20/09: Image Compression Algorithms

Presenter:   Matt Zurschmeide
Location:   Runciple Spoon
Description:   Image compression is a specialized form of data compression. This Monday, Matt will be giving a talk on how we can make certain assumptions about image data which allow for specific compression techniques to be used. Among the topics covered will be the TGA,PNG, and JPEG image formats.
References: PNG, JPEG


7/27/09: The Watchman Problem
Presenter:   Christine Task
Location:   Sushi Bar
Description:   Imagine you have a graph, and on that graph are a set of watchmen, and a thief. Time is discrete. In each time increment, the thief and every watchman can move from the vertex they are on to a neighboring vertex. If some watchman 'sees' a thief (lands on a vertex that's neighboring the thief's location), then the thief is caught. For a given graph, how many watchmen does it take to guarantee the thief gets caught? How quickly can they do it? (for instance, every complete graph (edges between every pair of vertices) is watchable by one watchman)
References: On the Paranoid Watchman Problem (the flawed masters thesis)


8/3/09: The Matching Problem (Applied)
Presenter:   Andy Keep
Location:   Trojan Horse
Description:   The stable matching problem, sometimes phrased as the stable marriage problem, is a way to take two, or sometimes more, populations and match individuals from population A and population B into stable matches. A matching between two groups is considered stable if there's no two people who would prefer to be with each other than their assigned partners. At Teach For America, we used this algorithm to match corp members with teaching positions we needed to fill.
References: Basic Wikipedia Pages: 1 2 3 4, Journal Articles: 1 2 3 4, and That French Book by Knuth.


8/10/09: Volumetric Graphics

Presenter:   Dave Bender
Location:   Irish Lion (notes: loud music)
Description:   I'll be talking about representing objects as solids using octrees (comprised of voxels). I'll cover orthographic rendering of these structures and discuss some of their benefits and limitations compared to polygonal methods.
References: Efficient Synthetic Image Generation of Arbitrary 3-D Objects, Geometric Modeling Using Octree Encoding


8/17/09: The 3n+1 Problem
Presenter:   Brenton Bostick
Location:   BuffaLouis
ent)Description:   The 3n+1 problem is a problem that was first stated in the 1930's and is currently open.It asks if for any positive integer n, does this program always terminate:

while (n != 1) {
  if (odd(n)) {
    n = 3n+1;
  } else {
    n = n/2; }
  }

Even though the problem is easily stated and appears simple, it is a source of complex behavior. In my talk, I will discuss the origin, milestones, and state-of-the-art of approaches to solving the 3n+1 problem.
References: Jeff Lagarias: The 3x+1 problem and its generalizations (Nice survey about the problem), 3x+1 Problem Annotated Bibliography (1963-1999), (2000-)
Eric Roosendaal: On the 3x+1 problem (Statistics and data about the problem)
Terence Tao: Moser's entropy compression argument. (Introduction of fundamentally new kind of termination argument)


8/24/09: Traffic Analysis (cars, not packets)
Presenter:   Josh Bonner
Location:   Runciple Spoon (notes: call to reserve room next time)
Description: [to be filled in]
References: [to be filled in]