The roles of monotonicity in Bayesian networks
When building Bayesian networks with the help of domain experts, often properties of monotonicity arise, such as veterinarians expressing that “under more severe conditions, seeing the more severe symptoms becomes more likely”. It is well known that human decision makers will not use a network in their daily practice if such common properties of monotonicity are clearly violated, not even if the network shows overall high performance. In this talk, I will introduce some properties of monotonicity for Bayesian networks and further focus on the different roles that monotonicity has in engineering networks, both in building them by hand, in learning them from data and in fine-tuning.
Linda C. van der Gaag studied Mathematics at Delft University of Technology (the Netherlands). After some time in industry, she returned to academia and received a PhD from the University of Amsterdam (NL). She became of full professor at Utrecht University (NL) in 2000, and at the same time received a fixed-term professorship at the University of Aberdeen (Scotland). In 2007, she was elected EurAI Fellow by the European Association for Artificial Intelligence. After having served for several years as head of the Department of Information and Computing Science at Utrecht University, she joined IDSIA in June 2019. Her research interests are in the field of probabilistic graphical models, and range from foundations of probabilistic independence to methods for enhancing the practicability of using Bayesian networks in the medical and veterinary fields.
Wednesday, 30th of October 2019, 12:00-13:00
Manno, Galleria 1, 2nd floor, room G1-201
Pizza (or alternative food) and drinks will be offered at the end of the talk. If you plan to attend the lunch, please register in a timely fashion at the following link so that we will have no shortage of food.
If by any chance, after registering for a Pizza at the link above, you know you will not be able to attend the lunch, please change your registration asap into “Cancel (no) Pizza”.
Approximation Algorithms for Two-dimensional Geometric Packing Problems
The field of Geometric Packing problems has attracted the attention of
many researchers in the late years. Generally speaking, in this
setting we are given a region in the two-dimensional plane and a set
of rectangles and the goal is to pack a subset of them inside the
given region in such a way that they do not overlap and some given
objective function is optimized. In this talk we will review our
recent developments on two of these problems in the framework of
Approximation Algorithms: Strip Packing and Geometric Knapsack. In the
first problem, the goal is to pack all the rectangles into a strip of
fixed width so as to minimize the final height of the packing, while
in the second one the goal is to pack a subset of the rectangles of
maximum profit into a rectangular region of fixed size. We will also
discuss applications and open questions regarding the mentioned
problems, the talk is meant to be accessible for non-experts.
Waldo Gálvez studied Applied Math at University of Chile and recently
got his PhD at IDSIA in the Algorithms and Complexity group. Starting
in January 2020, Waldo will join the Algorithms and Complexity group
at TU Munich as a postdoc. His main area of research is the design and
analysis of approximation algorithms, specially for Packing and
Network Design problems.
Tuesday, 22nd of October 2019, 12:00-13:00
Manno, Galleria 1, 2nd floor, room G1-204
Pizza (or alternative food) and drinks will be offered at the end of
the talk. If you plan
to attend, please register in a timely fashion at the following link
so that we will have no shortage of food:
Implementation of an incremental docking method to study long-sugar chains interactions with proteins: p17-heparin case study
Glyco-bioinformatics is an emerging subfield of bioinformatics aimed at expediting research in the field of glycomics. Unfortunately, the development of sugar-based virtual structures is made difficult by some structural features of sugar such as their high charge density, conformational flexibility and the torsional angles between glycosidic bonds. As a consequence, automated prediction of the binding poses of long sugar with proteins (that is a pivotal aspect of many biological processes) has been evaded so far, also due to the solvation/desolvation, weak surface complementarity and large electrostatic interactions of sugar/protein interactions.
My PhD activity is aimed at overcoming these limits. To this aim, I have implemented a new computational method based on incremental docking that has been so far successfully applied to two important biological interactions such as that of heparin with the HIV-1 p17 matrix protein in the field of AIDS and VEGF with its VEGF receptor-2 in the field of tumor neovascularization. Perspective developments include the development of an algorithm able to automatize the developed computational methods and their application to other sugar/protein interactions of biological importance.
Giulia Paiardi has completed her MSc in Bioinformatics and medical biotechnology at University of Verona in 2016. She then moved to the University of Brescia where she worked as data analyst on the project "Neuroendocrine and behavioural aspects of experimental autism” at the section of Pharmacology of the Department of Molecular and Translational Medicine (DMMT). She is now frequenting the PhD in Technology for Health at University of Brescia, where under the supervision of Prof. Marco Rusnati, at the Macromolecular Interaction Analysis Unit of the DMMT, she is in charge of all the computational aspects of the projects currently ongoing in the Unit. To this aim, in 2018, under the supervision of Dr. Pasqualina D’Ursi, she has frequented the Bioinformatics group at the Institute for Biomedical Technologies-National Research Council (ITB-CNR) in Milan, leaded by Prof. Luciano Milanesi, with which she is still in close collaboration and that currently provides her the computing systems.
Wednesday, 16th of October 2019, 12:00-13:00
Manno, Galleria 1, 2nd floor, room G1-201
Pizza and drinks will be offered at the end of the talk. If you plan to attend, please register in a timely fashion (before the day of the talk) at the following link so that we will have no shortage of food:
You are cordially invited to attend the following talk by our Postdoc
candidate Kazuki Irie, which will take place tomorrow at 11:45 in room
222 at IDSIA. Please find additional details below.
Sjoerd van Steenkiste
*Title: Recent Advances in Neural Language Modeling for Automatic Speech
In this talk, I will shortly present two of my recent works on language
modeling for automatic speech recognition. The first work is the
application of large Transformer language models (recently popularized
by OpenAI's GPT-2 models) to automatic speech recognition. We
successfully trained deep and powerful Transformer language models and
obtained an excellent performance on the publicly available LibriSpeech
dataset (where our results currently mark the state of the art). In the
second part, I will present a large mixture of LSTM language models and
its application to real application scenarios where the data domain is
diverse, as is the case for our experiments on the YouTube speech
Kazuki Irie is a PhD student in the Human Language Technology Group at
RWTH Aachen University, Germany, under the supervision of Prof. Hermann
Ney, since May 2014. Prior to that, he received a Diplôme d'ingénieur
degree from École Centrale Paris, France, and jointly a Master degree
(Master MVA) from ENS Cachan, France, both in Applied Mathematics in
2013. His PhD research is focused on advancing language modeling for its
applications to speech recognition (and machine translation). He is
broadly interested in RNN, language, and related methods.
*Date: Tuesday, 8 October 2019, 11:45-12:30
*Location: Manno, Galleria 1, 2nd floor, room 222.