Last update: 06/09/2017

Speakers

Confirmed Speakers

 

 

Usama Fayyad, Ph.D

CEO Open Insights / www.open-insights.com

He reactivated Open Insights, after leaving Barclays in London. He is also Interim CTO for Stella.AI a Mountain View, CA VC-Funded startup in AI for recruitment. He is acting as Chief Operations & Technology Officer for MTN’s new division: MTN2.0 aiming to extend Africa’s largest telco into new revenue streams beyond Voice & Data.

See more about Mr. Fayyad here.

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

Peter Flach

Professor of Artificial Intelligence
Intelligent Systems Laboratory, Department of Computer Science 

University of Bristol, UK

"The value of evaluation: towards trustworthy machine learning"

Machine learning, broadly defined as data-driven technology to enhance human decision making, is already in widespread use and will soon be ubiquitous and indispensable in all areas of human endeavour. Data is collected routinely in all areas of significant societal relevance including law, policy, national security, education and healthcare, and machine learning informs decision making by detecting patterns in the data. Achieving transparency, robustness and trustworthiness of these machine learning applications is hence of paramount importance, and evaluation procedures and metrics play a key role in this.

In this talk I will review current issues in theory and practice of evaluating predictive machine learning models. Many issues arise from a limited appreciation of the importance of the scale on which metrics are expressed. I will discuss why it is OK to use the arithmetic average for aggregating accuracies achieved over different test sets but not for aggregating F-scores. I will also discuss why it is OK to use logistic scaling to calibrate the scores of a support vector machine but not to calibrate naive Bayes. More generally, I will discuss the need for a dedicated measurement theory for machine learning that would use latent-variable models such as item-response theory from psychometrics in order to estimate latent skills and capabilities from observable traits.

See more about Prof. Flach here.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Francisco Herrera

Head of Research Group SCI2S
Soft Computing and Intelligent Information Systems

Universidad de Granada, Spain

"A tour on Imbalanced big data classification and applications"

Big Data applications are emerging during the last years, and researchers from many disciplines are aware of the high advantages related to the knowledge extraction from this type of problem.

The topic of imbalanced classification has gathered a wide attention of researchers during the last several years. It occurs when the classes represented in a problem show a skewed distribution, i.e., there is a minority (or positive) class, and a majority (or negative) one. This case study may be due to rarity of occurrence of a given concept, or even because of some restrictions during the gathering of data for a particular class. In this sense, class imbalance is ubiquitous and prevalent in several applications. The emergence of Big Data brings new problems and challenges for the class imbalance problem.

In this lecture we focus on learning from imbalanced data problems in the context of Big Data, especially when faced with the challenge of Volume. We will analyze the strengths and weaknesses of various MapReduce-based algorithms that address imbalanced data. We will present the current approaches presenting real cases of study and applications, and some research challenges.

See more about Prof. Herrera here.

 

 

 

 

 

 

 

 

 

 

 

 

 

Witold Pedrycz

Professor and Canada Research Chair IEEE Fellow
Professional Engineer Department of Electrical and Computer, University of Alberta, Canada

"Linkage Discovery: Bidirectional and Multidirectional Associative MemoriesIn Data Analysis" 

Associative memories are representative examples of associative structures, which have been studied intensively in the literature and have resulted in a plethora of applications in areas of control, classification, and data analysis. The underlying idea is to realize associative mapping so that the recall processes (both one-directional and bidirectional) are characterized by a minimal recall error.

We carefully revisit and augment the concept of associative memories by proposing some new design directions. We focus on the essence of structural dependencies in the data and make the corresponding associative mappings spanned over a related collection of landmarks (prototypes). We show that a construction of such landmarks is supported by mechanisms of collaborative fuzzy clustering. A logic-based characterization of the developed associations established in the framework of relational computing is discussed as well.

Structural augmentations of the discussed architectures to multisource and multi-directional memories involving associative mappings among various data spaces are proposed and their design is discussed.

Furthermore we generalize associative mappings into their granular counterparts in which the originally formed numeric prototypes are made granular so that the quality of the associative recall can be quantified. Several scenarios of allocation of information granularity aimed at the optimization of the characteristics of recalled results (information granules) quantified in terms of coverage and specificity criteria are proposed.

See more about Prof. Pedrycz here.

 

 

 

 

Ronald Yager

Professor of Informarion Systems and Director of the Machine Intelligence Institute at Iona College.

 

See more about Prof. Yager here.

 

 

 

Further speakers will be confirmed.

 

 

BAFI Conference