| M | T | W | T | F | S | S |
|---|---|---|---|---|---|---|
| « Dec | ||||||
| 1 | ||||||
| 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| 9 | 10 | 11 | 12 | 13 | 14 | 15 |
| 16 | 17 | 18 | 19 | 20 | 21 | 22 |
| 23 | 24 | 25 | 26 | 27 | 28 | 29 |
| 30 | 31 | |||||
These are the photos from the 26th IST seminar on December 8, 2011.
The number of participants: approximately 17 (including 2 presenters).
(1) presentation by Dr. Shuichi Miyazaki
(2) presentation by Ms. Mayumi Kamada
(3) discussion
Time/Date: 13:00-15:00, 08 Dec
Location: Room 304, Research Building #5 (総合研究５号館).
http://www.media.kyoto-u.ac.jp/ja/access
First slot:
Talk: “Approximation algorithms for the stable matching problem” (in Japanese)
Speaker: Associate Prof. Shuichi Miyazaki, (Okabe Lab)
Abstract:
The stable matching problem is the problem of finding a “stable” matching for given preference lists of participants. This problem is used in many situations, such as assigning residents to hospitals, university students to laboratories, and so on. In this talk, I give some basic introduction of the stable matching problem, as well as the results on the approximation algorithms for it.
Second slot:
Talk: “Prediction of protein residue contacts and structure-function analysis for time-series data of protein structure.” (in Japanese).
Speaker: PhD candidate, Mayumi Kamada, (Akutsu Lab.).
Abstract:
Proteins have an important role in life activity. They can serve their functions only by forming their structures and interacting with other molecules. Since interaction between protein residues is a clue for understanding these mechanisms, many methods have been developed to predict protein residue contact from amino acid sequence. In this talk, I will introduce our prediction method for predicting protein residue contacts based on co-evolution information using random field model. Proteins are dynamic molecules within the cell, and the dynamics is thought to play a crucial role for structure change and interaction between molecules. For correlation analysis between structure motion and protein function, we extract motion features from time-series data of protein structure using wavelet transform. I will show some analyses using the features.
These are the photos from the 25th IST seminar on November 10, 2011.
The number of participants: approximately 11 (including 2 presenters).
(1) presentation by Ms. Natalia Efremova
(2) presentation by Mr. Yamato Okamoto
(3) discussion
Time/Date: 13:00-15:00, 10 Nov
Location: Room 304, Research Building #5 (総合研究５号館).
http://www.media.kyoto-u.ac.jp/ja/access
First slot:
Talk: “The network for view-invariant object recognition and classification” (in English)
Speaker: PhD student Natalia Efremova, (Inui Lab)
Abstract:
It has been reported that the primate visual system is able to perform classification and categorization of objects based on their shape,
and that humans rely heavily on shape similarity among objects for object categorization and identification. The goal of our research is to propose a neurophysiologically plausible model of three-dimensional object recognition and classification, based on recent psychophysical and neurophysiological findings. We propose a neural network of ventral visual pathway for recognition and classification of 3D real-world objects. The model consists of hierarchy of modules, representing the visual areas V1-V4 and Inferotemporal cortex (IT). The architecture of the model represents the self-organized map (SOM) of functional radial-basis function (RBF) modules. We show that the proposed architecture is capable of recognizing and categorizing 3D real-world objects in a way consistent with recent fMRI findings.
Second slot:
Talk: “Human re-identification between non-overlapping cameras” (in Japanese).
Speaker: OKAMOTO Yamato, M1, (Minoh Lab).
Abstract:
Recently, there are many surveillance cameras in our daily space.
It is useful to obtain people’s trajectories from videos of such
surveillance cameras automatically. For example, we can realize automatic surveillance system. Since a field of view in a camera is limited, matching persons across non-overlapping cameras has been studied. In this talk, I’m going to introduce such researches.
These are the photos from the International Seminar on Primordial Knowledge Model on October 25, 2011.
The number of participants: approximately 13 (including 1 presenter).
(1) Introduction by Dr. Toyoaki Nishida
(2) presentation by Dr. Jordi Vallverdú
