Academic Year 2015

The 9th LSSE Seminar

Prof. Shuji Aou's Last Lecture
Date February 26th, 13:00-14:30
Venue Lecture Room 2
Title Brain- and life-induced smart system
Speaker Prof. Shuji Aou (Kyushu Institute of Technology)

The 8th LSSE Seminar

Date January 22nd, 17:00-18:00
Venue Lecture Room 2
Title Principle of Redundancy Reduction (PRR): Abstraction-Generalization Ability
Speaker Dr. Syozo Yasui (Kyushu Institute of Technology)
Host Prof. Tetsuo FURUKAWA
Abstract

The brain of new-born babies has random connections for its neuronal system. Every neuron is connected with any other neurons. There is no obvious synaptic rule. Such randomness keeps increasing until the brain owner becomes several years old.

The randomness (synapse density) eventually reaches a peak. Thereafter, it begins to decrease and finally settle on a plateau.

The interpretation follows. Initially, connections are prepared as many as possible. However, through everyday life and experiences, it will be learned that not all such connections are necessary. Unnecessary and redundant ones are eliminated. The remaining ones are selected neural pathways that form a pruned slimy network.

In this talk, I will try to explain that such “skeleton brain” has a high ability of abstraction and generalization for cognitive science. This is called “Principle of Redundancy Reduction (PRR). Furthermore, applications of PRR to artificial intelligence will be described, where a neural ?network algorithm automatically leans PRR adaptively, so as to deal with changing environments.


The 7th LSSE Seminar

Date November 27th, 11:15-12:30
Venue Lecture Room 2
Title The emergence of phantom agency in mechanical rhythms
Speaker Dr. Hideyuki Takahashi (Specially appointed assistant professor at Graduate School of Engineering, Osaka University)
Host Prof. Tomohiro SHIBATA
Abstract

Rhythmic synchrony between different individuals often brings various emotions in our mind. We found that we sometimes feel phantom agency, if our motor rhythm synchronize with external mechanical rhythm of non-animate objects (e.g. robot). To verify this illusion, we have performed several behavioral and fMRI experiments. In this presentation, I would like to introduce our recent findings and discuss how rhythmic synchrony is transformed to the sense of agency in our brain.


The 6th LSSE Seminar

Date November 18th, 10:30-12:00
Venue Lecture Room 2
Title Real-time execution of multi-model Spiking Neural Networks based on parallel FPGA hardware
Speaker Prof. Jordi Madrenas (Department of Electronic Engineering, Technical University of Catalunya)
Host Prof. Takashi MORIE
Abstract

The efficient emulation of complex Spiking Neural Networks (SNNs) presents two main challenges: a) Massive parallelism of neurons and synapses; b) Complex connectivity among them. In order to efficiently emulate neurons and synapses, a parallel multiprocessor architecture based on simple custom processors operating on a Single-Instruction Multiple-Data (SIMD) basis is proposed. The main properties of such architecture, namely, software-programmable algorithms, parallel emulation of neural parameters, serial emulation of synapses and local parameter memory, support any spiking model and real-time execution with excellent scalability. To cope with the interconnect complexity, a serial pipelined ring topology based on Address Event Representation (AER) is proposed. The AER transmission allows channel multiplexing while still providing real-time operation, due to the low-frequency of biological spikes. The serial ring topology presents a much better scalability than parallel or serial common buses, because it only uses local, point-to-point lines.


The 5th LSSE Seminar

Date October 16 (Fri), 10:00-12:00
Venue Lecture Room 2
Program

10:00-11:15: Talk 1

Model Selection for Latent Variable Models via Asymptotic Bayesian Inference

Dr. Kouhei Hayashi (NII, Japan)

Language: Japanese (Slides in English)

Host: Prof. Tomohiro SHIBATA

11:20-12:00: Talk 2

Seminar presented by Tokyo Steel Co, Ltd.

Language: Japanese

Host: Prof. Hirofumi TANAKA

Talk 1

Speaker

Dr. Kouhei Hayashi (Project researcher at Global Research Center for Big Data Mathematics, NII, Japan)

TiTle

Model Selection for Latent Variable Models via Asymptotic Bayesian Inference

Abstract

Latent variable models, such as mixture models and hidden Markov models, represent high-dimensional observed data by low-dimensional latent variables. The latent space captures the hidden structure of data and thus the determination of its dimensionality, i.e. model selection is an important problem. For example, if the dimensionality is misspecified, predictive performance and/or interpretability would be degraded. In this talk, from a Bayesian perspective, a few standard methods for model selection will be introduced. In addition, a recently-developed asymptotic approximation technique will be explained.

Biography

Kohei Hayashi is a project researcher at Global Research Center for Big Data Mathematics, National Institute of Informatics. He received the B.Eng degree from Ritsumeikan University in 2007, and M.Eng and Ph.D degrees from Nara Institute of Science and Technology in 2009 and 2012, respectively. Before the current position, he was a JSPS Postdoc at the University of Tokyo. He is interested in relational data analysis and Bayesian probabilistic modeling.


The 4th LSSE Seminar

Date September 24, 13:00-14:30
Venue Lecture Room 2
Title Objective Measurement of Human's Full Body Motion and Emotion for Medical and Robotics Applications
Speaker Dr. Salvatore Sessa (International Center for Science and Engineering Programs, Faculty of Science and Engineering, Waseda University)
Host Prof. Tomohiro SHIBATA
Abstract

I believe that only if it is possible to measure human's full body Motions and Emotions then we can naturally interact with robots and understand deeply human being. In this presentation, I will briefly introduce the research conducted at Waseda University, Takanishi-laboratory. In particular, I will focus my discussion on the Waseda Bioinstrumentation systems developed on purpose for the objective evaluation of motion and emotion in medical and robotics applications. I will show several medical applications in which the objective evaluation of motion is fundamental such as surgical gesture evaluation and assessment of older adults mobility. Furthermore, I will explain how motion analysis and gesture recognition can improve and promote a natural human-robot interaction showing several examples of interactions between the anthropomorphic musical robots developed in Waseda University and their human partners. Waseda Bioinstrumentation systems has been also used for emotion measurement and detection. In this presentation, I will focus on "laugh" because is a direct and universal form of social interaction associated with positive emotion and physical well-being but sometimes it is a symptom of behavioral or neurological disorders. Objectively measure physiological signals that play the most important role in laughter and integrate several non-invasive sensors as a unique tool specifically designed for medical application and robot interaction is a big challenge.


The 3rd LSSE Seminar

Date July 10, 14:40-16:10
Venue Lecture Room 2
Title Recent Progress on Reinforcement Learning
Speaker Dr. Eiji Uchibe (Neural Computation Unit, OIST)
Host Prof. Tomohiro SHIBATA
Abstract

Reinforcement learning is a computational framework for learning control policies based on feedback from the environment and for understanding the brain's mechanisms of decision making. In this talk, I will briefly talk about the following three important progresses: (1) deep reinforcement learning, (2) design of reward, and (3) duality between control and estimation. The first topic is deep reinforcement learning that is a combination of (forward) reinforcement learning and deep learning. It has been receiving attention as a method to learn control policies from high-dimensional sensory space since Google DeepMind published their work in Nature this year. Although they achieved notable success in the Atari 2600 video game domain, they just use some classical techniques from over two decades ago, namely convolutional neural networks, experience replay, fixing a target value, and so on. In addition, it is reported that the learning time is prohibitively long and it is sensitive to the choice of meta-parameters such as learning rate. We propose a new activation function that is not a monotonically increasing function and experimental results show that our method outperforms classical deep learning. Next, I would like to discuss how we should train robots. In this talk, two approaches for designing reward are introduced. One is inverse reinforcement learning that can infer a reward function from observed behaviors which are assumed to be optimal. However, inverse reinforcement learning is an ill-posed and the reward function is not uniquely determined from behaviors. We show that inverse reinforcement learning can be simplified if the dataset sampled from a baseline policy is available as well as the dataset from the optimal policy. According to this finding, we propose a novel method using density ratio estimation. The other is so-called intrinsically motivated reinforcement learning, which provides additional intrinsic rewards calculated by the learning agent. Since the intrinsic rewards are sometimes task-irrelevant, we formulate it as a constrained optimization in which the constraints are formed by the set of extrinsic rewards. Experimental results show our approaches are very effective to design rewards. Finally, I'm going to touch on the duality between optimal control and optimal inference in the field of reinforcement learning. In general, the Hamilton-Jacobi-Bellman (HJB) equation, which is a nonlinear partial differential equation, should be solved to find an optimal policy, but the recent studies on duality shows that the HJB equation can be simplified by introducing some constraints into the reward function. As a result, some stochastic optimal control problems are converted into optimal inference problems, and it is possible to derive some state-of-the-art reinforcement learning algorithms such as KL-control, path integral reinforcement learning, and so on. I will show our approaches as long as time permits.


The 2nd LSSE Seminar

Date June 11, 14:40-15:40
Venue Lecture Room 2
Title Robotics to Rural
Speaker Prof. S.K. Saha (Naren Gupta Chair Professor, Mechatronics Lab. & Programme for Autonomous Robotics Lab., IIT Delhi, INDIA)
Host Prof. Tomohiro SHIBATA
Abstract

The presentation will cover the robotics research at IIT Delhi in two different laboratories, namely, Mechatronics Laboratory which Prof. Saha started in 2001, and the Programme for Autonomous Robotics Laboratory (a sponsored inter-disciplinary activity by three departments) started in 2010. The presentation will include a vision-based guidance of an industrial robot, an immersive environment for the tele-operation of an industrial robot, peg in tube operation using a force control algorithm, etc. The speaker will also explain the development of a truck simulator and a haptics device for medical simulation. Prof. Saha will then introduce a concept called RoCK-BEE (Robotics Competition Knowledge-based Education in Engineering). It is to encourage the students of engineering to take part in robotic competitions, e.g., ABU Robocon, to enjoy not only the competition but also the learning of the subject in a fun and effective way. It will be shown in the later part of his presentation that how the knowledge of robotics can be applied for the design optimization of many devices and processes used by the rural people, e.g., a carpet loom or a machine to clean it. This aspect is pursued by Prof. Saha under the banner called MuDRA or Multibody Dynamics for Rural Applications.


The 1st LSSE Seminar

Date June 10, 14:40-16:10
Venue Lecture Room
Title
Speaker Prof. Toyokazu Yamada (Graduate School of Advanced Integration Science, Chiba University)
Host Prof. Hirofumi TANAKA
Abstract

Page Top