Restricted Area  |

2013 Workshop on the Brazilian Data Mining and Machine Learning Initiatives

January, 17th, 2013 – 10:00 – 18:00

Room 2077 – ICEx

Department of Computer Science

Universidade Federal de Minas Gerais

Description

This workshop aims to congregate researchers and practitioners on the areas of data mining and machine learning, as well as their applications in areas such as social media and e-health. The agenda will comprise discussions about research agendas, and the planning of initiatives in terms of joint research projects, events, technology development and transfer, and reach out initiatives. The participants include researchers and students from several universities (RPI, UFMG, UFU, UFF, USP, and UFSJ, among others) as well as practitioners from industry (SERPRO, IBM, UOL, and EMC).

Organization

InWeb – National Institute of Science and Technology for the Web

PPGCC – Graduate Program in Computer Science from UFMG

DCC – Department of Computer Science – UFMG

Speakers

  • Alexandre Plastino – UFF
  • André Carvalho – USP-SC
  • Mohammed Zaki – RPI
  • Sandra de Amo – UFU
  • Wagner Meira Jr. – InWeb – UFMG

Schedule

9:00 – 10:00 – Welcome coffee

10:00 – 10:30 – Random Musings on Pattern Discovery

Mohammed Zaki – RPI

Pattern mining is one of the fundamental techniques in data mining. As one increases the expressive power of the pattern types, from subsets to subsequences, subtrees, subgraphs, and beyond, one discovers potentially more informative patterns. Due to the complexity of real-world data, the traditional complete search methods have given over to random sampling approaches to discover interesting patterns. I will outline the current trends and applications of this new paradigm, and will offer thoughts on directions for the future.

10:30 – 11:00 – Characterizing and understanding the dynamics of online social networks

Wagner Meira Jr. – InWeb /UFMG

The internet has been evolving from a communication media to an environment where users talk about the most diverse topics, reflecting the dynamics of the society at broad. Characterizing and understanding how the internet data may be used for assessing real events becomes a key component of many Internet-based applications and demands the development of new data mining models and techniques. Data mining in such scenarios is challenging because the data is intrinsically uncertain and multiscale, the patterns to be mined are complex and evolve through time, and there is a huge amount of information that need to be processed in real time. In this talk we present a framework for the research and development of data mining models, algorithms and systems that target these challenging scenarios. We also present the Web Observatory, a platform for collecting, analyzing and presenting, at real time, information mined from social networks and the web, as well as some of its instances that focused on sports, politics, and health.

11:00 – 11:30 – Hybridization of Metaheuristic with Data Mining Techniques

Alexandre Plastino – UFF

Recent research has shown that the hybridization of metaheuristics is a powerful mechanism to develop more robust and efficient methods to solve hard optimization problems. The combination of different techniques and concepts behind metaheuristics, if well designed, has the potential to exploit their advantages while diminishing their drawbacks, which results in methods suited to a more diverse set of real problems. The DM-GRASP heuristic is one such hybrid method that has achieved promising results. It is a hybrid version of the GRASP metaheuristic which incorporates a data mining process. Data mining refers to the automatic recognition and extraction of knowledge from datasets in terms of patterns and rules. The basic idea of the proposed hybridization is that patterns mined from good quality solutions could be used to guide the search, leading to a more effective exploration of the solution space. In this seminar, we will review how this hybridization was designed and survey the results of its practical applications evaluated until now.

11:30 – 12:00 – Contextual Preference  Mining

Sandra de Amo – UFU

Elicitation of preferences consists basically in providing the user a way to inform his/her preferences on objects belonging to a dataset, with a minimal effort for him/her. It can be achieved by following different strategies: (a) by using a query interface where users are asked to express their preferences or (b) by capturing implicit user’s choices and applying preference mining techniques in order to infer the user preferences. The first alternative is not efficient since the users in general are not able to express their preferences in an exact and consistent way. In this seminar we focus on the second alternative for preference elicitation and discuss several important research issues related to this topic: (1) Preference Modeling: How to formalize the concept of preference ? (2) The preference mining task: Does preference mining task reduce to a classification task ? (3) Preference Mining Techniques: (a) Which are the input data ?  a ranking of objects  ? a small pairwise sampling  of objects ? (b) What is the final goal:  to extract  a  partial order from the input data  ? to extract a ranking function ? (c) How the existing techniques for pattern discovery and classification can be adapted for the preference mining task ? (4) Recommendation Systems: How preference mining techniques can be used in  recommendation systems ?

12:00 – 12:30 – Learning to select Machine Learning Algorithms

André de Carvalho – USP-SC

Learning algorithms have been successfully applied over the years to different application domains. This has resulted in the development of a large number of algorithms and their variations. However, according to empirical and theoretical results there is no single algorithm that outperforms the others in every problem. Meta-Learning provides a general framework for the selection of the most suitable algorithm for a new problem. In this talk I will show how metalearning can be used for algorithm selection.

12:30 – 14:00 – Lunch

14:00 – 16:00 – Working Groups

WG1 – Projects

This working group will discuss research and technological initiatives in data mining and machine learning to be pursued by both academia and industry. Besides specific projects that leverage on the experience and expertise of the various groups and organizations, we should discuss initiatives such as data and software repositories as well as collective services that will benefit the researchers and practitioners.

WG2 – Outreach initiatives

This working group will discuss initiatives that include events and other activities to foster critical mass in the workshop areas. We should discuss at least three types of events: academic, industrial, and schools. Further, we should also discuss the organization of international events in Brazil, in particular SIGKDD and ICML. Regarding other activities, identification of other research groups and attraction of students, in particular from Latin America seem to be relevant initiatives.

16:00 – 16:30 – Coffee break

16:30 – 18:00 – Plenary

Presentation of the results of the working groups and further actions.

Sneaks by https://essay4today.com jordana was funded by our zell founders fund
Copyright © 2013 InWeb - Instituto Nacional de Ciência e Tecnologia para a Web - All rights reserved.
XHTML 1.1 OKXHTML 1.1 CSS 2.1 OKCSS 2.1 razz