CANSSI Newsletter
Advancing research in the statistical sciences in Canada
Canadian Statistical Sciences Institute
Copula Dependence Modeling: Theory and Applications CRT at their December 2014 meeting at CRM.

Congratulations to Nancy Reid, an Officer of the Order of Canada

OC MedalOn December 26, 2014, the Governor General of Canada named Nancy Reid as an Officer of the Order of Canada for her leadership in the field of statistical inference, which has helped to facilitate sound public policy decision making. The Officer level of the Order of Canada recognizes national service or achievement. Every recipient of the Order of Canada receives a fine silver insignia. This insignia is in the shape of a stylized six-point snowflake. At its centre is a red annulus inscribed with the Order’s motto DESIDERANTES MELIOREM PATRIAM (“They desire a better country”). The annulus surrounds a red maple leaf and is topped by the Royal Crown.

The Big Data Program

The Fields Institute thematic program on Statistical Inference, Learning and Models for Big Data, in which CANSSI has collaborated, will soon be coming to a close. We are very grateful to the Fields Institute for inviting us to lead the organization of this program, outside of the Fields CANSSI contribution, so soon after the inception of CANSSI. The six-month program encompassed graduate courses on machine learning and inference for big data, a short course on latent tree graphical models, and a full slate of workshops emphasizing deep learning, statistical learning, optimization, visualization, networks, health and social policy, and physical sciences. Distinguished lectures were given at Fields by Emmanuel Candes, Michael Jordan, Terry Speed and Bin Yu. A Big Data industrial problem-solving workshop was held May 25-29 at Fields. Still to come: the closing conference, June 12-13 in Halifax, co-sponsored by CANSSI and AARMS and presented jointly with the Institute for Big Data Analytics at Dalhousie.
Because of the importance of this initiative to the Canadian statistical sciences community, chief organizer Nancy Reid felt it was important that the thematic program, though primarily held at Fields, should be pan-Canadian. Thanks are due to the organizing committee: Nancy Reid, Yoshua Bengio, Hugh Chipman, Sallie Keller, Lisa Lix, Richard Lockhart, and Ruslan Salakhutdinov - for all their work in bringing about a very successful program coast to coast. We are grateful to PIMS for sponsoring allied workshops in statistical inference for large scale data and big data in environmental science, and to CRM for sponsoring and hosting a workshop on networks, web-mining and cyber-security, and an upcoming summer school on deep learning. There have also been allied activities at Fields, including a CANSSI-sponsored workshop on complex spatio-temporal data structures, and one on big data in commercial and retail banking.
To get an idea of the magnitude of this program and its many offshoots, see more online.

New Composition of CANSSI SAC

CANSSI would like to thank Jim Berger and Anthony Davison who have been members of the Scientific Advisory Committee (SAC) for the last two years, and have stepped down. We welcome Christian Robert, Université Paris-Dauphine and University of Warwick, and Nell Sedransk, Associate Director of NISS, who will be joining the SAC. Steve Fienberg, Peter Guttorp, Susan Holmes, Jack Kalbfleisch, Xihong Lin, Rob Tibshirani and Jeff Wu will continue in 2015.
Spring 2015

Save the Date

June 13, 2015:  The CANSSI 2015 Annual Meeting will take place from 3:30 - 6:00 pm at Dalhousie University the day before the SSC meetings begin.

Check out our calendar of events. If you'd like your event to appear here, let us know!

One of the new CRT projects applies to wildland fire regime modeling.

Scientific Planning Meeting and Annual General Meeting

Member institutions are invited to send a delegate to the Scientific Planning Meeting, to be held Saturday, June 13 from 3:30 to 4:30 pm, and to the Annual General Meeting, to be held on the same day at 4:30 pm. For the delegates, attending in person is preferable, but it will also be possible to attend this meeting and the AGM remotely.

The following topics may be discussed at the Scientific Planning Meeting:
  1. how departments can collaborate in the training of undergraduates and graduate students
  2. MITACS, and how CANSSI can support researchers interested in working with industry
  3. coordinating postdoctoral fellowship opportunities and applicants
The agenda for the Annual General Meeting is be available on the CANSSI website and includes a financial update, discussion of the CANSSI Operating Policies, and the election for the formation of the new CANSSI Board.

How the election will work:

The initial Board of CANSSI was formed in November 2012 by the Executive of the Statistical Society of Canada, and except for one who replaced a resigning member, the current members of the Board have all served for two and a half years. CANSSI is now instituting a process of Board renewal, and is thus proposing a slate of Board members to begin staggered terms of one, two or three years. The Officers of CANSSI (Director, Deputy Director, regional Associate Directors) are being appointed by the Board. The Director and two of the Associate Directors, elected by them, will be members of the Board. As well, the Directors of PIMS, the Fields Institute, and the CRM will be ex-officio members of the Board.

The rest of the Board will be elected by the membership. The Operating Policies specify election of two representatives of the institutional members, to be nominated by them, and this year there are five nominees, from whom two are to be elected. The Operating Policies also specify that there should be six to nine other “at large” members who are representative of the scientific and stakeholder communities, nominated by the Board Nominating Committee. This year, there are fewer than nine “at large” nominees, and thus all will be acclaimed.

Thus the ballot issued at the Annual General Meeting to voting delegates will present the names of the candidates to represent the institutional membership, and ask for each person to vote for up to two. One of those elected, chosen at random, will serve a two-year term and the other will serve a three-year term. The ballot will also present the names and proposed terms for the nominees for the “at large” positions, filled by acclamation. The ballots will be counted at the meeting and the results declared.

The full and final slate of candidates will be published June 6. Proxy votes will be accepted at the meeting, and a proxy form is available online.

Announcement of Second Round CRTs

Evolving Marked Point Processes with Application to Wildland Fire Regime Modeling

W. John Braun, University of British Columbia-Okanagan, and Douglas Woolford, Wilfrid Laurier University, are the team leaders. This project will develop statistical methods for quantifying and mapping fire risk, in particular methodology to model marked point process data aggregated over moderate to large temporal and spatial scales.

Specific aims:
  • using marked point process methodology to model ignition points in historical data from Alberta and other provinces, taking into account weather conditions and features of the natural and built environments
  • modeling smouldering and fire spread using underlying interacting particle systems
  • collaboration with provincial ministries and the Institute for Catastrophic Loss Reduction to develop visualization tools for fire managers and property insurers
Read more about this project on our website.

Statistical Inference for Complex Surveys with Missing Observations

David Haziza, Université de Montréal, is the team leader. This project will carry out research and training in survey data analysis necessitated by the advent of increasingly extensive data sets and challenges in data collection resulting from non-response and missing values.

Specific aims:
  • collaboration with researchers at Statistics Canada and Westat on problems in large scale complex and high-dimensional survey data with missing values
  • extending the techniques of fractional imputation and doubly robust methods to dealing with missing values in high dimensional data
  • developing inference from incomplete functional survey data, with application to large data sets on electricity consumption
Read more about this project on our website.

Modern Spectrum Methods in Time Series Analysis: Physical Science, Environmental Science and Computer Modeling

David J. Thomson, Queen’s University, is the team leader. Natural time series often have complex stochastic structures and spectra with “many lines” that are not generally captured by low-order parametric models. This project will pursue methods to model processes with such features as nearly periodic components, nonlinear coupling, non-stationarity and non-Gaussian distributions to devise appropriate statistical tests for their frequency domain parameters.

Specific aims:
  • international collaboration with physicists, engineers, and modelers, including researchers in Natural Resources Canada and Health Canada, to develop methods for exploratory data analysis of time series using multi-taper and related methods
  • application to models of seismic “noise” background, solar gravity modes, environmental solar effects, pollutants and meteorological phenomena
  • developing spectral estimation methods for multi-fidelity computer models
Read more about this project on our website.

Update on First Round CRTs

Copula Dependence Modeling: Theory and Applications

A CRM-CANSSI workshop entitled “New Horizons in Copula Modeling” was held in Montréal, December 15-18, 2014. This 3-½-day workshop was a highly successful event. There were 65 participants, 18 talks, and 15 posters that were divided into two sessions. Four of the poster presentations were made by students working on the project under the supervision of CRT members. The workshop also included an organizational meeting for the team. See the picture at the top of this newsletter.

Statistical Modeling of the World: Computer and Physical Models in Earth, Ocean and Atmospheric Sciences

The group at UBC are investigating the assessment of the Community Multiscale Air Quality computer (numerical) model against a physical network of ozone monitoring stations in Greater Vancouver and the Lower Fraser Valley. At SFU, development of rigorous design and analysis procedures for the calibration of glacier mass-balance models has begun. Thus far, a new experimental design and analysis approach was proposed for studies on non-convex regions (e.g., glacier surfaces). At Acadia, team members have begun work on combining deterministic computational models that simulate tidal flow in Minas Passage with tidal current observations made using the acoustic Doppler current profiler.

Advancements to State-Space Models (SSMs) for Fisheries Science

In 2014, team members presented 5 talks related to the project. Two papers related to the project have been submitted. Plans for the coming year include (at Memorial) evaluating the sensitivity of state-space model approaches with an emphasis on implementation, and (at Dalhousie) investigating the possible robustification of a model used by DTU Aqua to assess fish stocks.

This team is hosting their second workshop on Advancements in State-Space Models for Fisheries Science from June 17-19, 2015 at Dalhousie University. See the poster at the bottom of this newsletter or click on this link to learn more.

CHSG Conference Update

The 4th Annual Canadian Human and Statistical Genetics Meeting (CHSGM) took place in Vancouver, BC from the 18th to the 21st of April, 2015. CANSSI sponsored two sessions. The Prediction and Data Integration session was chaired by Shelley Bull (Lunenfeld-Tanenbaum Research Institute, University of Toronto) and Jinko Graham (Simon Fraser University). The next day, Brad McNeney and Jinko Graham (both of SFU) chaired a session called Interpreting Genomic Variation through Modeling Advances in Population and Statistical Genetics. 350 participants attended the conference at the Pinnacle Hotel Vancouver Harbourfront. The mostly Canadian participants were joined by researchers from as far away as Taiwan and Saudi Arabia. Next year, the CHSGM will take place from April 16-19 in Halifax, NS.

Employment Ads

Our website now includes a section for employment ads. Our member institutes are welcome to send us their ads.

There are two employment opportunities at HEC Montréal:
Statistics Lecturer - applications due as soon as possible
Statistics Professor - deadline: Oct. 15, 2015
Like us on Facebook
Keep Up With Us
Like us on Facebook to get updates on important events and interesting topics right in your news feed!
Copyright © 2015 CANSSI / INCASS, All rights reserved.