The 14th International Conference on Discovery Science (DS 2011)

Event type: 
Conference
Event time: 
2011-10-05 08:30 to 2011-10-07 18:30
Place: 
Dipoli Congress Center and Aalto University School of Science, Otakaari 24, 02150, Espoo, FI
Description: 

DS 2011 provides an open forum for intensive discussions and exchange of new ideas among researchers working in the area of Discovery Science. The scope of the conference includes the development and analysis of methods for automatic scientific knowledge discovery, machine learning, intelligent data analysis, theory of learning, as well as their application to knowledge discovery. Invited speakers of DS 2011 are Professors Ming Li (University of Waterloo, Canada), Eyke Hüllermeier (Philipps-Universität Marburg, Germany) and Johannes Fürnkranz (Technische Universität Darmstadt, Germany).

DS 2011 will be co-located with the Twenty-second International Conference on Algorithmic Learning Theory (ALT 11). The two conferences will be held in parallel, and will share their invited talks. The shared invited speaker of ALT 2011 and DS 2011 is Professor Yoshua Bengio (University of Montral , Canada).

For more information, see the DS2011 website (http://ds2011.org/)

Helsinki Privacy Experiment

Lecturer : 
Antti Oulasvirta, Jukka Perkiö
Event type: 
HIIT seminar
Event time: 
2011-11-04 10:15 to 11:00
Place: 
Kumpula Exactum B222
Description: 

 

Talk announcement
HIIT Seminar Kumpula, Friday November 4 10:15, Exactum B222

SPEAKER:
Antti Oulasvirta & Jukka Perkiö

TITLE:
Helsinki Privacy Experiment

ABSTRACT:
As of September 2011, Helsinki Privacy Experiment (HPE) had instrumented 10 ordinary 
households with a data collection system that includes videocameras and microphones, 
as well as logging of all network traffic, PC use, and smartphone sensor data. Data 
collection is continuous for one year per household. If you want to how and why we 
did this, come listen to the talk!

 

Biomine and network algorithms

Lecturer : 
Hannu Toivonen
Event type: 
HIIT seminar
Event time: 
2011-10-14 10:15 to 11:00
Place: 
Kumpula Exactum B222
Description: 
Talk announcement
HIIT Seminar Kumpula, Friday October 14 10:15, Exactum B222

SPEAKER:
Hannu Toivonen
University of Helsinki

TITLE:
Biomine and Network Algorithms

ABSTRACT:
tba

Focused Multi-task Learning Using Gaussian Processes

Lecturer : 
Jaakko Peltonen
Event type: 
HIIT seminar
Event time: 
2011-10-21 10:15 to 11:00
Place: 
Kumpula Exactum B222
Description: 
Talk announcement
HIIT Seminar Kumpula, Friday October 21 10:15, Exactum B222
(Please notice the new date!)

SPEAKER:
Jaakko Peltonen
Aalto University

TITLE:
Focused Multi-task Learning Using Gaussian Processes

*** This work by Gayle Leen, Jaakko Peltonen, and Samuel Kaski 
won the Award for Best Paper in Machine Learning at ECML PKDD 2011, 
the European Conference on Machine Learning and Principles and 
Practice of Knowledge Discovery in Databases. ***

ABSTRACT:
Given a learning task for a data set, learning it together with 
related tasks (data sets) can improve performance. Gaussian 
process models have been applied to such multi-task learning 
scenarios, based on joint priors for functions underlying the 
tasks. In previous Gaussian process approaches, all tasks have 
been assumed to be of equal importance, whereas in transfer 
learning the goal is asymmetric: to enhance performance on a 
target task given all other tasks. In both settings, transfer 
learning and joint modeling, negative transfer is a key problem: 
performance may actually decrease if the tasks are not related 
closely enough. In this paper, we propose a Gaussian process model 
for the asymmetric setting, which learns to “explain away” 
non-related variation in the additional tasks, in order to focus on 
improving performance on the target task. In experiments, our model 
improves performance compared to single-task learning, symmetric 
multi-task learning using hierarchical Dirichlet processes, and 
transfer learning based on predictive structure learning.

Symbol Elimination in Program Analysis

Lecturer : 
Laura Kovács
Event type: 
HIIT seminar
Event time: 
2011-09-23 10:15 to 11:00
Place: 
Kumpula Exactum B222
Description: 
Talk announcement
HIIT Seminar Kumpula, Friday September 23 10:15, Exactum B222

SPEAKER:
Laura Kovács
Vienna University of Technology

TITLE: 
Symbol Elimination in Program Analysis

ABSTRACT: 
Automatic understanding of the intended meaning of computer
programs is a very hard problem, requiring intelligence and reasoning. 
In this talk we present a new method for program analysis, called 
symbol elimination, that uses first-order theorem proving techniques 
to automatically discover non-trivial program properties, such as 
loop invariants and loop bounds. Moreover, symbol elimination can be 
used as an alternative to interpolation in software verification.

BIO:
Laura Kovács is a Hertha Firnberg Research Fellow of the Austrian Science
Fund at the Institute of Computer Languages of the Vienna University of 
Technology. Her research deals with the design and development of new
theories, technologies, and tools for program verification, with a
particular focus on automated assertion generation, symbolic summation,
computer algebra, and automated theorem proving. She holds an MSc from the
Western University of Timisoara, Romania, and a PhD degree from the
Research Institute for Symbolic Computation of the Johannes Kepler
University, Linz, Austria. Before joining TU Wien, she was a postdoctoral
researcher in the Models and Theory of Computation research group of Prof.
Dr. Thomas  A. Henzinger at the Swiss Federal Institute of Technology 
Lausanne (EPFL), and at the Programming Methodology research group of 
Prof. Dr. Peter Müller at the Swiss Federal Institute of Technology 
Zürich (ETH).

Pages