Seminars and Colloquia by Series

Friday, November 13, 2015 - 13:05 , Location: Skiles 005 , Marcel Celaya , Georgia Tech , Organizer: Yan Wang
We find a good characterization for the following problem: Given a rational row vector c and a lattice L in R^n which contains the integer lattice Z^n, do all lattice points of L in the half-open unit cube [0,1)^n lie on the hyperplane {x in R^n : cx = 0}? This work generalizes a theorem due to G. K. White, which provides sufficient and necessary conditions for a tetrahedron in R^3 with integral vertices to have no other integral points. Our approach is based on a novel proof of White's result using number-theoretic techniques due to Morrison and Stevens. In this talk, we illustrate some of the ideas and describe some applications of this problem.
Friday, November 6, 2015 - 13:00 , Location: Skiles 005 , Troy Retter , Emory University , Organizer: Yan Wang
For an integer k, the Folkman number f(k) is the least integer n for which there exists a graph G on n vertices that does not contain a clique of size k and has the property that every two coloring of E(G) yields a monochromatic clique of size of size k. That is, it is the least number of vertices in a K_{k+1}-free graph that is Ramsey to K_k. A recent result of Rodl, Rucinski, and Schacht gives an upper bound on the Folkman numbers f(k) which is exponential in k. A fundamental tool in their proof is a theorem of Saxton and Thomason on hypergraph containers. This talk will give a brief history of the Folkman numbers, introduce the hypergraph container theorem, and sketch the proof of the Rodl, Rucinski, and Schacht result. Recent work with Hiep Han, Vojtech Rodl, and Mathias Schacht on two related problems concerning cycles in graphs and arithmetic progressions in subset of the integers will also be presented.
Friday, October 30, 2015 - 13:00 , Location: Skiles 005 , Alejandro Toriello , Georgia Tech , Organizer: Yan Wang
We consider a version of the knapsack problem in which an item size is random and revealed only when the decision maker attempts to insert it. After every successful insertion the decision maker can choose the next item dynamically based on the remaining capacity and available items, while an unsuccessful insertion terminates the process. We propose a new semi-infinite relaxation based on an affine value function approximation, and show that an existing pseudo-polynomial relaxation corresponds to a non-parametric value function approximation. We compare both theoretically to other relaxations from the literature and also perform a computational study. Our new relaxation provides tight bounds over a variety of different instances and surprisingly becomes tighter as the number of items increases. Joint work with Daniel Blado (ACO) and Weihong Hu (ISyE).
Friday, October 23, 2015 - 13:30 , Location: Skiles 005 , Anna Kirkpatrick , Georgia Tech , Organizer: Yan Wang
We introduce a notion of pattern occurrence that generalizes both classical permutation patterns as well as poset containment. Many questions about pattern statistics and avoidance generalize naturally to this setting, and we focus on functional complexity problems – particularly those that arise by constraining the order dimensions of the pattern and text posets. We show that counting the number of induced, injective occurrences among dimension 2 posets is #P-hard; enumerating the linear extensions that occur in realizers of dimension 2 posets can be done in polynomial time, while for unconstrained dimension it is GI-complete; counting not necessarily induced, injective occurrences among dimension 2 posets is #P-hard; counting injective or not necessarily injective occurrences of an arbitrary pattern in a dimension 1 text is #P-hard, although it is in FP if the pattern poset is constrained to have bounded intrinsic width; and counting injective occurrences of a dimension 1 pattern in an arbitrary text is #P-hard, while it is in FP for bounded dimension texts. This framework easily leads to a number of open questions, chief among which are (1) is it #P-hard to count the number of occurrences of a dimension 2 pattern in a dimension 1 text, and (2) is it #P-hard to count the number of texts which avoid a given pattern?
Friday, October 23, 2015 - 13:05 , Location: Skiles 005 , Kevin Lai , Georgia Tech , Organizer: Yan Wang
We resolve an open question from (Christiano, 2014b) posed in COLT'14 regarding the optimal dependency of the regret achievable for online local learning on the size of the label set. In this framework the algorithm is shown a pair of items at each step, chosen from a set of n items. The learner then predicts a label for each item, from a label set of size L and receives a real valued payoff. This is a natural framework which captures many interesting scenarios such as collaborative filtering, online gambling, and online max cut among others. (Christiano, 2014a) designed an efficient online learning algorithm for this problem achieving a regret of O((nL^3T)^(1/2)), where T is the number of rounds. Information theoretically, one can achieve a regret of O((n log LT)^(1/2)). One of the main open questions left in this framework concerns closing the above gap. In this work, we provide a complete answer to the question above via two main results. We show, via a tighter analysis, that the semi-definite programming based algorithm of (Christiano, 2014a), in fact achieves a regret of O((nLT)^(1/2)). Second, we show a matching computational lower bound. Namely, we show that a polynomial time algorithm for online local learning with lower regret would imply a polynomial time algorithm for the planted clique problem which is widely believed to be hard. We prove a similar hardness result under a related conjecture concerning planted dense subgraphs that we put forth. Unlike planted clique, the planted dense subgraph problem does not have any known quasi-polynomial time algorithms. Computational lower bounds for online learning are relatively rare, and we hope that the ideas developed in this work will lead to lower bounds for other online learning scenarios as well. Joint work with Pranjal Awasthi, Moses Charikar, and Andrej Risteski at Princeton.
Friday, October 9, 2015 - 13:00 , Location: Skiles 005 , Arefin Huq , Georgia Tech , Organizer: Yan Wang
Yannakakis showed that the matching problem does not have a small symmetric linear program. Rothvoß recently proved that any (not necessarily symmetric) linear program also has exponential size. It is natural to ask whether the matching problem can be expressed compactly in a framework such as semidefinite programming (SDP) that is more powerful than linear programming but still allows efficient optimization. We answer this question negatively for symmetric SDPs: any symmetric SDP for the matching problem has exponential size. We also show that an O(k)-round Lasserre SDP relaxation for the metric traveling salesperson problem (TSP) yields at least as good an approximation as any symmetric SDP relaxation of size n^k. The key technical ingredient underlying both these results is an upper bound on the degree needed to derive polynomial identities that hold over the space of matchings or traveling salesperson tours. This is joint work with Jonah Brown-Cohen, Prasad Raghavendra and Benjamin Weitz from Berkeley, and Gabor Braun, Sebastian Pokutta, Aurko Roy and Daniel Zink at Georgia Tech.
Friday, September 25, 2015 - 13:00 , Location: Skiles 005 , Sarah Cannon , Georgia Institute of Technology , Organizer: Yan Wang
Many statistical physics models are defined on an infinite lattice by taking appropriate limits of the model on finite lattice regions. A key consideration is which boundary to use when taking these limits, since the boundary can have significant influence on properties of the limit. Fixed boundary conditions assume that the boundary cells are given a fixed assignment, and free boundary conditions allow these cells to vary, taking the union of all possible fixed boundaries. It is known that these two boundary conditions can cause significant differences in physical properties, such as whether there is a phase transition, as well as computational properties, including whether local Markov chain algorithms used to sample and approximately count are efficient. We consider configurations with free or partially free boundary conditions and show that by randomly extending the boundary by a few layers, choosing among only a constant number of allowable extensions, we can generalize the arguments used in the fixed boundary setting to infer bounds on the mixing time for free boundaries. We demonstrate this principled approach using randomized extensions for 3-colorings of regions of Z2 and lozenge tilings of regions of the triangle lattice, building on arguments for the fixed boundary cases due to Luby Our approach yields an efficient algorithm for sampling free boundary 3-colorings of regions with one reflex corner, the first result to efficiently sample free boundary 3-colorings of any nonconvex region. We also consider self-reducibility of free boundary 3-colorings of rectangles, and show our algorithm can be used to approximately count the number of free-boundary 3-colorings of a rectangle.
Friday, September 11, 2015 - 13:00 , Location: Skiles 005 , Chun-Hung Liu , Princeton University , Organizer: Yan Wang
A graph is a minor of another graph if the former can be obtained from a subgraph of the latter by contracting edges. We prove that for every graph H, if H is not a minor of a graph G, then V(G) can be 3-colored such that the subgraph induced by each color class has no component with size greater than a function of H and the maximum degree of G. This answers a question raised by Esperet and Joret, generalizes their result for 3-coloring V(G) for graphs G embeddable in a fixed surface, and improves a result of Alon, Ding, Oporowski and Vertigan for 4-coloringing V(G) for H-minor free graphs G. As a corollary, we prove that for every positive integer t, if G is a graph with no K_{t+1} minor, then V(G) can be 3t-colored such that the subgraph induced by each color class has no component with size larger than a function of t. This improves a result of Wood for coloring V(G) by 3.5t+2 colors. This work is joint with Sang-il Oum.
Friday, April 17, 2015 - 13:00 , Location: Skiles 005 , Cristóbal Guzmán , Georgia Tech , , Organizer:
How much of space can be filled with pairwise non-overlapping copies of a given solid? This is one of the oldest problems in mathematics, intriguing since the times of Aristotle, and remaining remarkably elusive until present times. For example, the three-dimensional sphere packing problem (posed by Kepler in 1611) was only solved in 1998 by Ferguson and Hales. In this talk, I will provide some historical and modern applications of geometric packing problems, and I will introduce a methodology to derive upper bounds on the maximal density of such packings. These upper bounds are obtained by an infinite dimensional linear program, which is not computationally tractable. However, this problem can be approximated by using tools from sums of squares relaxations and symmetry reduction (harmonic analysis and representation theory), leading to rigorous computational upper bounds on the density. Time permitting, I will present ongoing work with Maria Dostert, Fernando de Oliveira Filho and Frank Vallentin on the density of translative packings of superspheres (i.e., ell_p balls). This is an introductory talk: no previous knowledge of sums of squares relaxations or symmetry reduction is assumed.
Friday, January 16, 2015 - 13:00 , Location: Skiles 005 , Peter Winkler , Dartmouth College , Organizer:

Joint ARC colloquium/ACO student seminar

Pursuit games---motivated historically by military tactics---are a natural for graphical settings, and take many forms. We will present some recent results involving (among other things) drunks, Kakeya sets and a "ketchup graph.'' Lastly, we describe what we think is the most important open problem in the field.