May
31
4:45 pm16:45

the tea time talks: rich sutton

  • University of Alberta

The Tea Time Talks are a series of talks primarily given by the students and faculty studying Artificial Intelligence at the University of Alberta, and provide a comfortable, informal space in which to listen and learn about topics pertaining to machine intelligence and machine learning. 


May
26
12:00 pm12:00

ai seminar - marlos machado

  • University of Alberta

Representation learning and option discovery are two of the biggest challenges in reinforcement learning (RL). Proto-RL is a well known approach for representation learning in MDPs. The representations learned with this framework are called proto-value functions (PVFs). In this paper we address the option discovery problem by showing how PVFs implicitly define options.

    May
    12
    12:00 pm12:00

    ai seminar - russ greiner

    • University of Alberta

    Many tasks require building and using a model --- eg, to relate a patient's disease state with the possible symptoms, underlying causes, and effects of various treatments. These relationships are often probabilistic; eg a disease will often, but not always, manifest certain symptoms. Bayesian Belief Nets (BNs), which provide a succinct way to represent such probabilistic models, are in routine use for a wide range of applications, including medicine, bioinformatics, document classification, image processing and decision support systems.

    May
    5
    12:00 pm12:00

    ai seminar – marcin mizianty

    • University of Alberta

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Apr
    28
    12:00 pm12:00

    ai seminar – noah smith

    • University of Alberta

    Dr. Noah Smith (University of Washington) presents new methods for linguistic structure prediction as continuous generalizations of state machines and probabilistic grammars. He shows how they've led to fast and accurate performance on several syntactic and semantic parsing problems.

    Apr
    7
    12:00 pm12:00

    ai seminar – craig milne

    • University of Alberta

    all you need to know about creating your spin-off or start-up

    Bio: Craig has a PhD in Stem Cell Biology and an MBA in Finance.  Craig is involved in a number of companies in Edmonton, most recently as President of Stream Technologies, a company producing a new class of hyper spectral imaging camera and software.  Craig can assist with the spin-off process, licensing arrangements with TEC Edmonton, business development, financing, grant and operation strategies.  Craig is supporting the Dean's goal, to create more spin-offs with long-term survivability.which will be followed by discussion, and Q&A on this topic.


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations

    ai seminar: robert holte
    Mar
    24
    12:00 pm12:00

    ai seminar: robert holte

    • University of Alberta

    This talk begins with a review of long-established results in heuristic search and the early history of bidirectional heuristic search as well as describing a recent breakthrough in bidirectional heuristic search (the MM algorithm), which challenges long-held assumptions and exposes exciting new research directions.

    ai seminar - michael bowling
    Mar
    17
    12:00 pm12:00

    ai seminar - michael bowling

    • University of Alberta

    DeepStack bridges the gap between AI techniques for games of perfect information—like checkers, chess and Go—with ones for imperfect information games–like poker–to reason while it plays using “intuition” honed through deep learning to reassess its strategy with each decision.

    Mar
    10
    12:00 pm12:00

    ai seminar – kai on wong

    • University of Alberta

    Ethnicity is an important variable in epidemiological research. Canada is an ethnically diverse country, yet its insufficiency of ethnicity data impedes research progress and policy development in public health domains. This data gap negatively affects Aboriginal Canadians in particular as their health inequality may remain hidden. Automated name- and location-based ethnicity classification has shown potential but its applicability within the Canadian context is largely unknown.

    Mar
    3
    12:00 pm12:00

    ai seminar – spencer murray

    • University of Alberta

    In this talk, Amii's Spencer Murray will present an overview of Amii research, showing how we're driving innovation in research and development and making an impact outside of the academic community.

    Dec
    12
    9:30 am09:30

    ai seminar - özlem aslan

    • University of Alberta

    In this talk, Özlem presents a novel reformulation of supervised training of a two-layer architecture by introducing a latent feature kernel, which allows a rich set of latent feature representations to be captured while still allowing useful convex formulations via semidefinite relaxation.

    Dec
    9
    12:00 pm12:00

    ai seminar – filipe mesquita

    how to build (or join) a data science team

    Presenter: Filipe Mesquita, Vice President, Data Science, Mitre Media

    Description: This seminar is for students and professionals interested in building or joining a data science team. I will discuss the qualifications and skills required for different roles within a data science team, relevant interview questions, and how the data science team relates to other teams in an organization.
    If you are interested in joining our data science team, apply at https://www.linkedin.com/jobs/view/232448328

    Bio: Filipe Mesquita is the Vice President of Data Science at Mitre Media, a technology company focused on financial publishing. He has a PhD in Computer Science from the University of Alberta. His interests include data science, information extraction, and data mining.
     


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

    Dec
    2
    12:00 pm12:00

    ai seminar – garrett nicolai

    "lemmatization and reinflection: two sides of the same coin"

    Presenter: Garrett Nicolai, PhD student, Department of Computing Science, University of Alberta

    Abstract: In English, a single word generally only appears in a handful of forms: see, sees, seeing, seen, saw. In other languages, however, words need to be marked for a large number of features such as number, mood, tense, case, etc. My talk will consist of two parts. First, I will present our ACL 2016 paper which describes a machine-learning method for simplifying inflections by leveraging publicly-available inflection tables, such as those found on Wiktionary. In the second part, I will discuss the surprising results of the First Shared Task on Morphological Reinflection in Berlin, which demonstrate the growing importance of deep learning and neural networks in language processing.

    Bio: Garrett Nicolai is a senior PhD student working on problems in Natural Language Processing with Greg Kondrak. He has published 10 papers since starting his PhD program.


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

    Nov
    25
    12:00 pm12:00

    ai seminar – carlo montemagno

    "getting to mind: the foundation development of a biological neuronal ionic computer"

    Presenter: Carlo Montemagno, Director, Ingenuity Lab; Professor, Department of Chemical and Materials Engineering, University of Alberta; Strategic Chair of Bionanotechnology, Alberta Innovates – Technology Futures; Program Lead, Biomaterials Program, National Institute for Nanotechnology; Canada Research Chair in Intelligent Nanosystems

    Abstract: "Living systems are more than a product of matter manipulation at the molecular scale; the richness of functionality associated with living systems is a direct product of the information generated from both the interactions between molecules and the overall supra-molecular structure of the system. In essence living systems are “living” because of the fusion of nanotechnology and informatics. Living systems result from the precision assembly of matter with prescribed modalities for the transport and transduction of information among supra-molecular clusters.

    The ability to manipulate matter a single molecule at a time renders many things possible that were impossible before. Living systems do this on a regular basis. The core challenge is how to transform a labile molecule that exists in a fragile living organism and to transfer that functionality into a stable system that is economically scalable. The most significant difficulties revolve around environmental stability and the inherent structural limitations of the molecule.

    Presented is a new technology that stabilizes biological molecules maintaining their function for months at application relevant environmental conditions transitioning additive manufacturing from 3D space to a four-dimensional, functional space. This has enabling the synthesis of a new class of printable “inks”. These “ink” have stabilized and active biological molecules as integrated elements of synthesized polymer constructs to create a new class of materials and devices that now includes biologic function, both action and reporting as intrinsic properties.

    We will present efforts to build Excitable vesicles, nanometer scale structures that produce and transmit ionic currents. Their design, modeling and experimental results associated with the fabrication and engineering of Excitable Vesicles, a nano-sized building block with the ability to intrinsically process information will be discussed. Excitable vesicles are designed and engineered to create a complex system that self-organizes on multiple length scales to manifest emergent behaviors. Hopefully providing an experimental tool for studying emergent function."

    Bio: Driven by the principles of excellence, honor and responsibility and an unwavering commitment to education as an engine of economic prosperity, Carlo Montemagno, PhD has become a world-renowned expert in nanotechnology and is responsible for creating groundbreaking innovations which solve complex challenges in the areas of informatics, agriculture, chemical refining, transportation, energy, and healthcare.

    He was Founding Dean of the College of Engineering and Applied Sciences at University of Cincinnati; received a Bachelor of Science degree in Agriculture and Bio Engineering from Cornell University; a Master’s Degree in Petroleum and Natural Gas Engineering from Penn State and a PhD in Civil Engineering and Geological Sciences from Notre Dame. He is now in Alberta as the Director of Ingenuity Lab, professor in the Department of Chemical and Materials Engineering at the University of Alberta, AITF Strategic Chair of Bionanotechnology, Program Lead of the Biomaterials Program at the National Institute for Nanotechnology and Canada Research Chair in Intelligent Nanosystems.

    “Research and education are critical to success because the transfer of knowledge creates economic prosperity.” — C. Montemagno

    Carlo Montemagno has been recognized with prestigious awards including the Feynman Prize (for creating single molecule biological motors with nano-scale silicon devices); the Earth Award Grand Prize (for cell-free artificial photosynthesis with over 95% efficiency); the CNBC Business Top 10 Green Innovator award (for Aquaporin Membrane water purification and desalination technology); and named a Bill & Melinda Gates Grand Challenge Winner (for a pH sensing active microcapsule oral vaccine delivery system which increased vaccine stability and demonstrated rapid uptake in the lower GI tract.)


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

    Oct
    28
    12:00 pm12:00

    ai seminar – ryan cottrell

    *details coming soon*

    Presenter: Ryan Cottrell, Ph.D Student, Computer Science Department, Johns Hopkins University


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

    Oct
    21
    12:00 pm12:00

    ai seminar – geoff sutcliffe

    "the tptp world – infrastructure for automated reasoning"

    Presenter: Geoff Sutcliffe, Professor and Chair, Department of Computer Science, University of Miami

    Abstract: "The TPTP World is a well known and established infrastructure that supports research, development, and deployment of Automated Theorem Proving (ATP) systems for classical logics. The data, standards, and services provided by the TPTP World have made it increasingly easy to build, test, and apply ATP technology.

    This talk reviews the core features of the TPTP World, describes key service components of the TPTP World, presents some successful applications, and gives an overview of the most recent developments."

    Bio: Geoff Sutcliffe is a Professor and the Chair of the Department of Computer Science at the University of Miami. He received a BSc(Hons) and MSc from the University of Natal in South Africa, and a PhD in Computer Science from the University of Western Australia. His research is in the area of Automated Reasoning, particularly the development and ongoing maintenance of the TPTP problem library, and organizing the CADE ATP System Competition - the world championship for classical logic automated reasoning systems.


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

    Oct
    14
    12:00 pm12:00

    ai seminar – bernardo ávila pires

    "probability maximization"

    Presenter: Bernardo Ávila Pires, Ph.D Candidate, Department of Computing Science, University of Alberta


    Abstract: The theory developed by Vapnik and Chervonenkis provides a solid foundation to study the learnability of binary classification. For example, it is known that the so-called VC dimension of a hypothesis class precisely characterizes the amount of training data needed by optimal learning algorithms to compete against the best hypothesis in that class. It would be interesting to have a concept similar to the VC dimension for other discrete prediction problems. In fact, researchers have been after such notions of dimension for other discrete prediction problems, including multiclass classification, for many years by now, but so far their efforts have failed.

    In order to study learnability in discrete prediction problems, we will look into the problem of "Probability Maximization". This is a discrete prediction problem that we define and which generalizes well-known problems, including binary and multiclass classification, multilabel prediction, relation learning and some flavors of clustering. I will show that in some instances of probability maximization an appropriate notion of dimension does not exist. This previously-unknown phenomenon raises the question of whether the search for appropriate notions of dimension in other discrete prediction problems could also be futile.
     


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

    Oct
    7
    12:00 pm12:00

    ai seminar - perry kinkaide

    "the beginning of the end of the knowledge economy"

    Presenter: Perry Kinkaide, President, Kinkaide Enterprises Inc. (KEI); CEO, Alberta Council of Technologies (ABCtech)

    Abstract: "Analytics, big data, and the cloud were the for-runners to the emergence of machine learning, sensing and deciding apps, and pending artificial intelligence. This is but one of a host of emerging technologies that will significantly effect learning and working, shopping and play, plus deciding. This presentation will investigate the forces and sources of change and their current and anticipated socio-economic implications."

    Bio: Perry Kinkaide, MSc, PhD, CMC is an accomplished manager and since retiring from KPMG Consulting in 2001, he has remained active as an advisor and director for several diverse public and private organizations. His most senior career positions include: former Managing Director of KPMG Consulting in Edmonton and Assistant Deputy Minister with the Alberta Government. He received his PhD in Brain Research from the University of Alberta.

    Perry is the President of Kinkaide Enterprises Inc. (KEI), allocating venture capital and assigning operating officers to emerging technology-based, enterprises in Western Canada.  He is also currently the CEO of the Alberta Council of Technologies (ABCtech) that he founded in 2005. ABCtech is a non-profit Society of over 12,500 followers and embodies Perry’s continuing interest in the socio-economics of emerging technologies.

    Perry is called on throughout Canada and the United States for his insight and foresight – often humorous and always provocative, on the socio-economic implications of emerging technologies. He was recognized in 2007 by Alberta Venture magazine as one of the 50 "most influential" Albertans.


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

    Sep
    30
    12:00 pm12:00

    ai seminar - dale schuurmans

    "exploiting the relationship between forward and inverse reinforcement learning"

    Presenter: Dale Schuurmans, Professor, Department of Computing Science, University of Alberta; Principal Investigator, Amii; Principal Investigator, RLAI

    Abstract: "I will discuss a new approach to structured output learning that exploits a simple connection between forward and inverse reinforcement learning. 

    The key observation is that these two problems can be expressed as minimizing an identical Bregman divergence but in opposite directions.  The connection is not merely theoretical: it allows one to draw a precise relation between tempered log-likelihood and regularized expected reward, revealing that their difference is a simple variance term. 

    This observation suggests new approaches for exploiting supervised data to efficiently estimate the expected (regularized) reward of a given policy.  We have been exploiting this connection to develop a principled yet practical structured output training method where standard supervised learning (maximum likelihood) is applied to edited target labellings that have been sampled proportionally to their exponentiated scaled reward.  Recently, we have also been using this perspective to develop a principled unification of actor-critic methods."

    Joint work with Mohammad Norouzi, with the assistance of Samy Bengio, Zhifeng Chen, Navdeep Jaitly, Mike Schuster, and Yonghui Wu.
     


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.

     

    Sep
    23
    12:00 pm12:00

    ai seminar - dale schuurmans

    "deep learning games"

    Presenter: Dale Schuurmans, Professor, Department of Computing Science, University of Alberta; Principal Investigator, Amii

    Abstract: "I will discuss a reduction of supervised learning to game playing that reveals new connections and provides some alternative training methods. 

    For convex one-layer problems, one can establish an equivalence between global minimizers of the training problem and Nash equilibria in a simple game. We have shown how the game can be extended to general acyclic neural networks with convex gates, and demonstrate a bijection between the Nash equilibria and critical (or KKT) points of the deep learning problem, assuming differentiability. Given this relationship we have been investigating alternative learning methods based on the no-regret algorithms used in computer game playing. 

    An interesting finding is that "regret matching'', a classical method from the game theory and economics literature, can achieve competitive training performance while producing sparser models than current deep learning algorithms."

    Joint work with Martin Zinkevich


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations

    Sep
    22
    5:00 pm17:00

    exploring entrepreneurship

    "venture capital and you"

    Presenter: Shivon Zilis, Partner and Founding Member, Bloomberg Beta

    Shivon invests in new companies that have machine intelligence at the heart of their business. Shivon will give insight into the things investors look for in a new venture, what it means to be an investable company and some common mistakes made when taking an idea from academia to industry.

    "from academics to entrepreneurship"

    Presenter: Foster Provost, Professor, Data Science & Information Systems, New York University

    Foster is currently working on his fourth startup in fields relating to computing science. He knows what it’s like to be an entrepreneurial academic, and he will share some of his experiences taking an idea from academia to commercial market.

    "starting up in edmonton"

    Presenter: Tiffany Linke-Boyko, CEO, Startup Edmonton

    Tiffany connects entrepreneurs and product builders with skills, community and space to support them as they take ideas to reality. Tiffany will talk about the diversity of programs, services and resources available to local entrepreneurs and startups plus let you know how to take those first steps into launching your idea.


    exploring entrepreneurship series

    Exploring Entrepreneurship is an ongoing series of informal talks giving computing science students a chance to explore topics in business planning and creation in a friendly and supportive environment.

    Talks feature speakers with a range of entrepreneurial experience and cover such topics as venture capital, business scalability and the availability of resources for local startups, among others.

    Exploring Entrepreneurship is hosted by Amii and the Department of Computing Sciences.

    Sep
    16
    12:00 pm12:00

    ai seminar - colin bellinger

    "beyond the boundaries of smote: a framework for manifold-based synthetic oversampling"

    Presenter: Colin Bellinger, Postdoctoral Fellow, Department of Computing Science, University of Alberta

    Abstract: "Problems of class imbalance appear in diverse domains, ranging from gene function annotation to spectra and medical classification. On such problems, the classifier becomes biased in favour of the majority class. This leads to inaccuracy on the important minority class, such as the disease or gene function of interest.

    Synthetic oversampling mitigates this by balancing the training set, whilst avoiding the pitfalls of random under and oversampling. The existing methods are primarily based on the SMOTE algorithm, which employs a bias of randomly generating points between nearest neighbours. The relationship between the generative bias and the latent distribution has a significant impact on the performance of the induced classifier.

    Our research into gamma-ray spectra classification has shown that the generative bias applied by SMOTE is inappropriate for domains that conform to the manifold property, such as spectra, text, image and climate change classification. To this end, we propose a framework for manifold-based synthetic oversampling, and demonstrate its superiority in terms of robustness to the manifold with respect to the AUC on three spectra classification tasks and 16 UCI datasets."


    ai seminar series

    Fridays at noon, Amii and the Department of Computing Science host AI Seminars, engaging presentations on topics in the broad field of artificial intelligence. With speakers from the University of Alberta and other world-leading groups, the talks give AI enthusiasts a friendly way of engaging with the latest trends and topics in research and development.

    Seminars are open to the public, and no registration is required, though seating is limited and on a first-come-first-served basis. Topics range from foundational theoretical work to innovative applications of artificial intelligence technologies.

    If you would like to present at an upcoming AI Seminar, please contact Colin Bellinger.

    Join the AI Seminar mailing list to stay up-to-date on all the latest presentations.