Seminars

The information on this page is not completed and schedule is tentative. Please check back later for more information.

  • Title: Digitized music - Drum Pads using Arduino
    Speaker: Baabu Aravind Vellaian Selvarajan

    Date: Monday, 23 February 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: In early days, mostly hardware musical instruments like Drums, Guitar, Piano were used to create music. But now-a-days mostly software programs are used to create different type of musical pitch. The idea is to create a simplified small musical instrument. By making use of arduino hardware to create a musical instrument, i.e., drum-pads using some sensors and different drum beat pitches are allotted to those drum-pads using arduino. According to the hit made on sensors on drum-pads, it will create the drum beat. Future goal is to connect arduino with synthesizers to make changes in the effect of drum beats and playing the digitized music using those drum-pads.

  • Title: Cost-sensitive Reducts
    Speaker: Kuifei Liu

    Date: Wednesday, 25 February 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Attribute reducts are an important concept in rough set theory. An attribute reduct preserves certain property of the given data. In real life, data acquisition and processing are associated with costs. The cost can be classified into two categories, the process cost and the result cost. Many existing studies focus on one type of cost and ignore combing both of them into a whole. In this talk, a general definition of a cost-sensitive attribute reduct is proposed. The objective is to find a reduct with the minimal cost, which is interpreted as an optimization problem. Consequently, different reducts are associated with different costs. In designing reduct construction algorithms, a heuristic algorithm is proposed to approximate the optimal cost-sensitive reduct.

  • Title: An Introduction to Pi-Calculus
    Speaker: Kevin Yu

    Date: Friday, 27 February 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Pi-calculus is a model for high-level descriptions of interactions and communications for multiple independent processes. Used in concurrent programming, pi-calculus allows us to explain how different process streams can not only be executed simultaneously, but also interfere and act with one another. I will be presenting an introduction to pi-calculus and will talk about its syntax, applications, and implementation, and history. I will give examples of its use in the field and its effectiveness as well as comparisons to other forms of process calculi.

  • Title: Alternative Representations for Sparse Coding Audio Data
    Speaker: Jordan Ubbens

    Date: Monday, 2 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: A trend in machine learning has emphasized the use of features which are learned algorithmically, in contrast to the hand-engineered features traditionally used in classification tasks. Sparse coding seeks to encode exemplars as a sparse vector of coefficients against a dictionary of basis functions. While sparse coding has yielded state of the art results in many application domains, computational challenges often make it impractical to use. This talk examines the application of sparse coding to the challenging problem of audio classification. The convex optimization problems involved in dictionary learning are discussed, and existing methods are reviewed. Finally, experimental results are presented where pre-processing steps effectively reduce the computational overhead involved in sparse coding audio data.

  • Title: Next Day Gas Price Prediction using Competitor Information
    Speaker: Imran Jawaid

    Date: Wednesday, 4 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Our goal is to predict the start of day price of gas at each store. Gas stores compete for customers by offering gas at lower prices. So before predicting the price of gas at a store, we first identify k competitors of the store, based on OPIS and Gas Buddy data for North American cities. The Vector Autoregressive (VAR) model is used to predict, at the start of day d, what the price of gas will be at the start of day d+1 at a particular store. The available inputs for our VAR model are the retail price, the wholesale price, the spot price, and the profit margin at the store of interest and k competitor stores.

  • Title: A survey of platforms for big data analytics
    Speaker: Khantil Patel

    Date: Friday, 6 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Big data analytics involves the analysis of complex, unstructured, large volume, and high velocity data, which poses limitations on the use of traditional data analysis platforms. In the recent years, big data analytics has drawn interest from many researchers, with the focus on developing novel big data analysis techniques, which has led to the development of many different algorithms and platforms. In this presentation, first an in-depth analysis of the challenges in big data processing and different factors that affects the choice of the platform is provided. Second, set of popular big data platforms that are currently being used in practice are introduced, with the benefits and drawbacks of each of them. Lastly, the case of using the Apache Storm platform for real-time sentiment-based anomaly detection analysis in Twitter streams will be discussed, along with justification for the choice of the platform.

  • Title: Sentiment Analysis - Identifying sentiment from non-sentiment bearing hashtags
    Speaker: Credell Simeon

    Date: Monday, 9 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: The growing popularity of Twitter in recent years has resulted in much research interest in the sentiment analysis of its user-generated content called tweets with its many unique features. One such feature is that of hashtags which are user-defined topics denoted by "#". Hashtags can simultaneously connect related tweets and users, and may contain sentiment information which can contribute significantly to the sentiment analysis of tweets. Therefore, there is a need to automatically distinguish sentiment hashtags from those that contain only topic information. In this presentation, we describe and evaluate a lexicon-based approach that has been applied to solve this problem. Experimental results show that this approach achieves 88.1% precision on sentiment hashtags and 95.9% recall on non-sentiment hashtags.

  • Title: Decision-theoretic three-way approximations of fuzzy sets
    Speaker: Shu Wang

    Date: Wednesday, 11 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: A basic idea of three-way decisions is to classify a set of objects into three regions, called the positive, negative and boundary regions, by using an evaluation function and a pair of thresholds. How to determine the thresholds and interpret the meaning of the thresholds are fundamental issues. In this talk, we will focus on a particular application of three-way decisions, that is, three-way approximations of fuzzy sets and shadowed sets. We reviewed the basic concept of fuzzy sets, shadowed sets and three-way decisions. We present a formulation suggested by Deng and Yao (A decision-theoretic formulation of three-way approximations). In this formulation, we determine the thresholds by minimizing the decision cost. We will also discuss future research.

  • Title: Visual Analytics of Social Media
    Speaker: Maha EL Meseery

    Date: Friday, 13 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Social media were originally used to connect friends and family, but recently people are using it to share opinions and comment on recent events. This evolution transformed social media from a networking platform to a valuable source of information for various domains. Analysts are increasingly turning to this source of information to discover events and understand public opinion. However, they require tools to synthesize and analyze the vast amounts of information generated by millions of users every day. Visual analytics is an emerging approach that combines multiple visualizations, intuitive interaction, and automated analysis to support reasoning and understanding massive amounts of data. In the last few years, visual analytics has been increasingly employed to explore and analyze social media.

  • Title: A Method to Reduce Boundary Regions in Three-Way Decision Theory
    Speaker: Cong Gao

    Date: Monday, 16 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: In the three-way decision theory, all the elements are divided into three regions: positive region, negative region and boundary region. In this presentation, a method based on tri-training algorithm to reduce the boundary region will be given. The algorithm builds up three classifiers based on three-way decision. All the data is divided into three parts randomly, aiming to keep the three classifiers different. Then a voting mechanism is adopted to label test samples. Experiments have shown that in most cases, tri-training algorithm is not only benefit for reducing boundary regions but also for improving classification precision. Additionally, some rules are found about the parameters alpha and beta how to affect boundary regions and classification precision.

  • Title: Visual overview of online news streams
    Speaker: Kenneth Emeka Odoh

    Date: Wednesday, 18 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: The problem with list-based representation of existing news interfaces is that recent news will hide the old news in the news stream. Thus the users miss out on important old news as more recent news are available. A visual summary of the news stream allows for pre-attentive processing. Visual analytics is an integrated approach to problem solving that combines the analytical ability of humans with the processing power of computers, thereby allowing the users to incrementally explore the live news stream without losing sight of relevant past news. In this seminar, a survey of existing news visualization methods will inspire the design of a news dashboard that supports real-time visualization of the news stream.

  • Title: Object Oriented Design Pattern: Factory Pattern
    Speaker: Omar Alali

    Date: Friday, 20 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Design patterns represent the best practices used by experienced object-oriented software developers. Design patterns are solutions to general problems that software developers faced during software development. These solutions were obtained by trial and error by numerous software developers over quite a substantial period of time. Factory pattern is one of most used design pattern in object oriented programing languages. This type of design pattern comes under creational pattern as this pattern provides one of the best ways to create an object. In Factory pattern, we create object without exposing the creation logic to the client and refer to newly created object using a common interface.

  • Title: Fullerene structures in musical controllers design
    Speaker: Hanlin Hu

    Date: Monday, 23 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: An isomorphic layout can be used to position notes on a 2D hexagonal grid. This allows constructs (chords, scales, melodies, etc.) take on the same shape regardless of the musical key. According to certain cutting rules of hexagonal grid, this work extends the idea of 2D isomorphic layouts into 3D lattice structures. Based on the full enumerated hexagonal isomorphic layouts and existing musical controller design in 3D(on curved surface or polyhedral planar), this paper explores the possibilities of musical controllers based on general 3D lattices which have been extensively studied in the topology research of fullerenes, polyhedral cages formed by carbon molecules which consist of hexagons and exactly twelve pentagons.

  • Title: A Mobile Approach for Searching and Exploring within Twitter
    Speaker: Radhika Gopi

    Date: Wednesday, 25 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Mobile computing is an evolving domain that focuses on access and use of information in a mobile context. Using mobile applications, people are able to share content whenever and wherever they need. Popular microblogging services like Twitter and Twitter app have made it easy for people to post short messages (tweets) on the go, but limit the ability of users to find information by only providing searching and browsing functionality. In this seminar, methods for extracting topics hidden in the tweets will be proposed. An interactive visual interface will shown to support exploration and inspection of messages, allowing the user to see how the tweets are related by the topics.

  • Title: Volumetric Remeshing Using Distance Fields
    Speaker: Fatemeh Bayeh

    Date: Friday, 27 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: In computer graphics applications, a mesh is a 3D geometric model that specifies the shape of an object in a simulated scene. Meshes can be created by a person using 3D modelling software, obtained from scanning devices, or generated from implicit isosurfaces in density models. Unfortunately, meshes generated by automated processes may contain many redundant vertices, and thus are not well suited for applications such as animation and simulation, which need real time rendering. Given a mesh, a remeshing technique computes another mesh that approximates the input mesh and satisfies some application-specific requirements. In this presentation, we will discuss our volumetric remeshing technique. The technique is based on distance fields and has the advantage of preserving sharp features.

  • Title: Getting Jiggy with Git
    Speaker: Barret Rennie

    Date: Monday, 30 March 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Version control systems are integral to the work software developers do. They reduce the complexity of developing software in a team environment and provide effective ways for managing the history of a project. The two main types of version control (centralized and distributed) will be discussed, with their respective merits and demerits. Git is a popular distributed version control system that was created by Linus Torvalds, the creator of the Linux kernel. Features of Git and distributed source code management systems, such as lightweight branching and distributed development, will be covered in detail. This talk will serve as an introduction to Git and workflows for using it in a productive manner.

  • Title: The Quantum Age of Computing Approaches
    Speaker: Ian Hauser

    Date: Wednesday, 1 April 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Quantum Computing is a new field where our understanding of the strange world of Quantum Mechanics is used to create a completely new breed of computers. These employ an entirely different method of storing and processing information, and can be used to solve problems that are impossible with conventional computers. This seminar will discuss what exactly a Quantum Computer is, introduce some strange principles behind how and why they work, and explain what they could potentially be used to do in the future. We will also go over the current state of development of these machines, what kinds of problems still need to be overcome, and why you will probably never see a Quantum Computer replace your PC at home.

  • Title: Automatic Music Genre Classification
    Speaker: Brandon Prevost

    Date: Monday, 6 April 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Digital music is one of the most popular data types distributed to the general public. This results in the need for very large music databases. These databases must be organized effectively. Music genre is one of the most popular descriptions of musical content. Manually labeling each piece of music would be an incredibly tedious task. Genre classification can also be subjective for different listeners leading to inconsistencies of genre labels. In this presentation I will explore methods for automatic music genre classification such as techniques used for feature extraction. I will also explore various supervised classification techniques such as Daubechies wavelet coefficient histograms, octave-based modulation spectral contrast, K-nearest neighbor, Gaussian mixture models and support vector machines.

  • Title: Variable and Value Ordering Heuristics for Constraint Satisfaction Problems
    Speaker: Ket Wei Yong

    Date: Wednesday, 8 April 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: Many real world computational problems can be formulated as Constraint Satisfaction Problems (CSPs). Scheduling and timetabling, graph colorability, Boolean satisfiability are some of these examples. Solving a CSP is about searching for a solution in a large search space. Therefore many search algorithms and heuristic techniques have been proposed to solve CSPs effectively. My main objective is to combine some of these heuristics to solve CSPs effectively. More precisely, I will combine backtracking, MAC, dynamic variable ordering (DVO) using ddom/conflicts and look-ahead value ordering (LVO) using min-conflicts (MC) to solve CSPs. A survey of how these heuristics can be combined to solve CSPs effectively will be presented in this seminar.

  • Title: Making Information Consumption more Interactive
    Speaker: Regan Meloche

    Date: Friday, 10 April 2015
    Time: 2:30pm - 3:30pm
    Place: Research and Innovation Centre 208

    Abstract: The huge amount of freely available information on Wikipedia would be much more consumable to the public if it were presented in a more interesting format. Beginning in Fall 2014, I have been working on an iOS app that turns a Wikipedia article into a fill-in-the-blanks quiz. This involves retrieving the article text, parsing out the relevant text, and most important of all, identifying and blanking out words that are suitable for such a quiz. As an app, it must also be straightforward to use and have some entertainment value. Each of these steps presented its own set of programming challenges. I will discuss the various strategies I used to solve these problems.

  • Title: FABRIK: An iterative approach to the Inverse Kinematics problem
    Speaker: Victoria Verlysdonk

    Date: Friday, 10 April 2015
    Time: 3:30pm - 4:30pm
    Place: Research and Innovation Centre 208

    Abstract: The Inverse Kinematics (IK) problem is defined as finding a suitable set of joint configurations that produce the desired positions for the end effectors in a kinematic chain. In the context of computer animation, a solution to the IK problem preferably provides realistic, agile, smooth and accurate motion for the end effectors. In this seminar, we discuss the heuristic method Forward And Backward Reaching Inverse Kinematics (FABRIK) (Aristidou and Lasenby, 2011), which proposes an iterative approach to the IK problem that avoids computation of rotational angles or matrices. We compare this method to other popular methods in terms of reliability and computational cost.



To Top of Page