A Fields-CQAM Public Lecture
Affective AI
APPLICATIONS
AND
ITS
had to be secured
to meet capacity.
During the special
lecture on Robustness
to and Approximations
in Stochastic Control
Models, Serdar
Yüksel (Queen's
University) discussed
approximations used
in the formulation
31. 01. 2019
Kang Lee
of stochastic control
problems and devoted
particular attention
to the properties
of robustness. Yüksel considered robustness in
relation to quantized approximations for stochastic
control problems using Borel spaces and presented
the conditions under which finite models can
be used to construct useful approximations. He
showed that weak continuity of the transition
kernel is sufficient to ensure convergence of finite
approximations. Using additional conditions,
he also obtained information about the rate of
convergence to the optimal cost and showed how
the convergence rate changes with quantization.
Artificial intelligence is now everywhere and has the poten-
tial to change all aspects of our lives. However, one major
shortcoming of the AI today is that although it may have
high IQ, it lacks EQ or emotional intelligence. In fact, the
best AI system today does not have the emotional intelli-
gence of a 2-year-old. Kang Lee is Professor and Tier 1 CRC Chair in developmen-
tal neuroscience at the University of Toronto. Drawing on
more than 2 decades of research, his team has developed
a novel imaging system called Transdermal Optical Imag-
ing that uses conventional video cameras to decode human
physiology and emotions.
For AI to transform our lives, AI must reach the next level.
That is, any future AI system must have not only IQ but EQ.
In this talk, Kang Lee will draw insights from a century of
psychological research on human emotion to discuss vari-
ous methods AI developers can use to develop intelligent
systems with affective artificial intelligence, or the ability to
detect, decode, interpret, and simulate human affects. 6.00 PM Thursday 31 January 2019
The Fields Institute, 222 College Street
Interested? Register online:
http://www.fields.utoronto.ca/talks/Affective-artificial-intel-
ligence-and-its-applications
This short course was offered to an interdisciplinary
audience of mathematicians and engineers
and it provided an excellent introduction to
the development of control systems that rely
on a combination of deterministic model
expressions and stochastic terms to account
for uncertainty and random behaviour.
With energy and mining sectors playing a critical
role in the modern economy, the Workshop
on Real Options in Energy and Mining, held
February 10-12, brought together academics
from mathematics, statistics, business, and
engineering. Crucial inputs to the foundation
of Canada's resource-based economy were
viewed through the perspective of real options.
Graduate students and postdoctoral scholars
shared their work with leading international
experts. Academics from institutions in Argentina,
Denmark, France, Germany, Norway, the United
States, and Canada gave 22 talks followed by
questions and discussions to foster further
research collaborations among attendees.
At the Digital Design of Pharmaceutical
Manufacturing Processes Public Lecture,
Salvador Garcia-Munoz (Imperial College London)
spoke about the efforts of pharmaceutical
companies to further new drug discovery and
to bring these drugs to market. In particular, he
discussed the roles of the U.S. Food and Drug
Administration and Health Canada in approving
new drugs and
ensuring baselines
Digital Design OF
of quality. He also
Pharmaceutical
lectured on the
Manufacturing
challenges of scaling
Processes
drug manufacturing
processes from
laboratory-based
experiments to
full commercial
production. On this
point, he noted
12. 02. 2019
that mathematical
Salvador Garcia Munoz
models and complex
multivariate
analysis are frequently employed to render
the process faster and more efficient.
A Fields-CQAM Public Lecture
The Workshop on Quantitative Analysis and the
Digital Turn in Historical Studies sought to address
an urgent need in digitally enabled research:
collaboration across research specializations.
According to the organizers, there are relatively
few opportunities for historical researchers to
meet, share their work, and learn about new
research tools. To this end, they invited 30
historians, demographers, mathematicians, and
statisticians to participate in intensive workshops
on new approaches to working with historical data.
During these workshops, they convened a series
of presentations, discussions, and networking
sessions on the state of the field's use of web
archives and social media streams, geographic
information systems, and network analysis.
During the special lecture on Multiscale Modelling
of Polymer Networks for Conservation of Old
7