Department of Computer Science


Seminars

Seminars and talks are frequently hosted by the Department and associated research groups. Staff and students are warmly invited to attend.

2016 | 2015 | 2014

2016

2015

  • Seminar - Approximation Algorithms for Fully Proportional Representation by Clustering Voters (Seminars) Event as iCalendar
    09 December 2015, 12 - 1pm
    Departmental Seminar on: Approximation Algorithms for Fully Proportional Representation by Clustering Voters
  • Seminar - The Survey Octopus: an approach to teaching Total Survey Error (Seminars) Event as iCalendar
    06 November 2015, 12 - 1pm
    Departmental Seminar on: The Survey Octopus: an approach to teaching Total Survey Error. The talk will describe the Survey Octopus, show how it relates to Groves et als' model of the survey lifecycle, and explain the benefits and problems of each depiction of TSE.
  • Seminar: Departmental Seminar on Adventures in computational social science: New Zealand as a "social laboratory" (Seminars) Event as iCalendar
    26 August 2015, 12 - 1pm
    Over the next two years, under a James Cook fellowship, we will be constructing a simulation-based model of some of the key socio-demographic processes in New Zealand society over the last quarter century, drawing in the first instance on the New Zealand Longitudinal Census, 1981-2013, as its empirical foundation. Potentially this is a powerful instrument of cooperative social inquiry that can be used for policy testing, for scholarly purposes, and also for teaching as well.
  • Seminar: Quantum random walks (Seminars) Event as iCalendar
    16 June 2015, 3 - 4pm
    Quantum walk represents a generalised version of the well-known classical random walk. Regardless of their apparent connection, the dynamics of a quantum walk is often non-intuitive and far deviates from its classical counterpart. A multi-particle quantum walk presents an even richer dynamical system due to intrinsic quantum correlation and interaction. Current research is suggesting potential applications across a wide range of different fields. In this talk, I will give a brief introduction to quantum walks, discuss their potential applications, and consider several physical implementation schemes.
  • Seminar: Can network coding help bridge the digital divide in the Pacific? (Seminars) Event as iCalendar
    10 June 2015, 12 - 1pm
    Network-coded TCP (TCP/NC) can increase goodput under high latency and packet loss, but has not been used to tunnel conventional TCP and UDP across satellite links before. We report on a feasibility study aimed at determining expected goodput gain across such TCP/NC tunnels into island targets on geostationary and medium earth orbit satellite links.
  • SERG Seminar: Performance Analysis using Subsuming Methods: An Industrial Case Study (Seminars) Event as iCalendar
    05 May 2015, 12 - 1pm
    Subsuming methods analysis is a new approach that aggregates performance costs across repeated patterns of method calls that occur in the application’s runtime behaviour. This allows automatic identification of patterns that are expensive and represent practical optimisation opportunities.
  • Seminar: Algorithmically random infinite structures (Seminars) Event as iCalendar
    29 April 2015, 2 - 3pm
    In spite of much work, research on algorithmic randomness has excluded the investigation of randomness for infinite structures such as graphs, trees, and algebras. The reason is that it was unclear how one introduces a meaningful measures into these classes of structures. In this talk we provide a solution to this problem, and initiate the study of algorithmic randomness in various classes of infinite structures.
  • Seminar: The $d$-step conjecture for runs verified (Seminars) Event as iCalendar
    15 April 2015, 12 - 1pm
    In 1999 Kolpakov & Kucherov showed that the maximum number of runs in a string is linear in the length of the string. A run is a maximal repetition. They posited a so-called \emph{runs conjecture} that the upper bound should be at most the length of the string, i.e. that $\rho=\max\{r(x):|x|=n\}\leq n$, where $r(x)$ denotes the number of runs in a string $x$. Many researchers, including the speaker, since investigated the problem and several progressively closer asymptotic lower and upper bounds were achieved. In 2012 Deza and Franek proposed a $d$-step approach and conjectured that $\rho_d(n) = \max\{r(x):|x|=n\ \&\ x\ has\ exactly$ $d\ distinct\ symbols\}\leq n-d$ for any $2\leq d\leq n$. In 2015, Bannai et al. proved the runs conjecture giving a universal upper bound of $\rho(n)\leq n-2$. Refining their method, we proved the $d$-step conjecture showing some interesting structural properties of run-maximal strings.
  • Seminar: Practical Record-and-replay Debugging (Seminars) Event as iCalendar
    25 March 2015, 12 - 1pm
    Mozilla's browser developers find debugging expensive and frustrating, especially when bugs are non-deterministic. Researchers have proposed to expedite debugging by recording, replaying and analyzing program executions, and in theory such techniques are well-understood, but they have not yet been widely adopted.
  • SERG Seminar: Improving Course Relevance: Techniques for Incorporating the Social Value of Computing into your Courses (Seminars) Event as iCalendar
    24 March 2015, 12 - 1pm
    This talk introduces "Computing for the Social Good: Educational Practices" (CSG-Ed). CSG-Ed is an umbrella term meant to incorporate any educational activity, from small to large, that endeavours to convey and reinforce computing's social relevance and potential for positive societal impact.
  • Seminar: Demystifying Big Data Adoption: Paradigm shifts and Complexity Tolerance Theory (Seminars) Event as iCalendar
    04 March 2015, 11am - 12pm
    Amidst the big data hype pushed by the technology vendors and the unexpectedly low adoption rate, this seminar aims to clarify the myths about big data adoption and to offer research directions on big data system development. We will first describe big data management challenges, present 10 paradigm shifts in big data management, and then present the results of an empirical study of 20 large international European companies (average number of employees of our case companies is > 150,000 people).
  • Seminar: Cognitive Crash Dummies: Predictive Human Performance Modeling for Interactive System Design (Seminars) Event as iCalendar
    02 March 2015, 3 - 4pm
    This talk reviews the state of the art of predictive modeling and discusses real-world experiences using cognitive crash dummies.
  • Seminar: Reticulate Evolution: Why Trees Become Networks (Seminars) Event as iCalendar
    25 February 2015, 12 - 1pm
    In this talk, we will investigate problems in evolutionary biology from the perspectives of graph theory and theoretical computer science.
  • Seminar: Real-Time Big Data Stream Analytics (Seminars) Event as iCalendar
    18 February 2015, 12 - 2pm
    We will discuss some advanced state-of-the-art methodologies in stream mining based in the use of adaptive size sliding windows. Finally, we will present the MOA software framework with classification, regression, and frequent pattern methods, and the new Apache SAMOA distributed streaming software.