EVENT | National Institute of Informatics

Lecture Series: Machine Translation (MT) and Computer-Aided Translation (CAT)

Melody — 2009-04-14T14:00:33+09:00

Lecture Series:Machine Translation (MT) and Computer-Aided Translation (CAT)
Lecturer: Prof. Christian Boitet
Place: National Institute of Informatics, 20F, Lecture Room
Date: April 14, 16, 21, 23, 28, 30
Time: 2pm - 4pm
Fee: Free
Registration: Not required

Overall schedule:
Lecture 1: Linguistic, computational and operational architectures of MT systems (Apr. 14)
Lecture 2: Linguistic architectures of MT systems (Apr. 16)
Lecture 3: Computational architectures of MT systems (Apr. 21)
Lecture 4: Engineering of MT and CAT systems (CAT = MT + TA, TA = translation aids) (Apr. 23)
Lecture 5: Evaluation of MT and CAT systems for various operational architectures (Apr. 28)
Lecture 6: Corpora for hybrid MT/CAT systems (Apr. 30)

Biography
Christian Boitet is full professor of computer science and NLP at Université Joseph Fourier (Grenoble, France). He learned MT with Pr. Bernard Vauquois, a famous pioneer and inventor in this field. He has been director of GETA (Study Group on MT) since 1985. His research activity concerns all theoretical, methodological and operational aspects of MT (Machine Translation), enlarged to other parts of NLP. He has advised 38 PhD and State Doctoral Thesis students to their defense, and advises or coadvises 11 PhD students at the moment, 2 of them currently at NII, and 1 at Kyodai.
Although his main training was in mathematics and computer science, he has a degree in Russian and has studied many languages, in particular German, English, Spanish, Italian, Hungarian, Malay and Japanese, to attain proficiency or simply to understand their specificities and be able to work on them with linguists or translators. He has tried to study all MT systems and approaches from the very beginning of the field, down to technical details of tools, techniques, and resource building. He is one of the authors of Ariane-G5, GETA's generic environment for developing "rule-based" fully automatic MT systems. He participated and led research on "static grammars" (Vauquois & Chappuy 1983, 1985), a formalism for semi-formal specification of string-tree correspondences, that brought lingware engineering closer to software engineering. He has also done or led research on several other MT paradigms, such as interactive MT (DBMT, LIDIA project), translation memories, example-based MT (in particular, S-SSTC model of Tang, USM), analogical MT (with Y. Lepage at ATR), and probabilistic MT (for speech and more recently for small sublanguages such as classified ads sent by SMS).
He has presented communications in many national and international conferences and published in various journals and books (about 200). He has also edited a volume presenting Prof. Vauquois' scientific work and the proceedings of DBMT-90, COLING-92, and MIDDIM-96. He has organized workshops on Dialog-Based MT and multimodal interactive disambiguation, and co-organized the COLING-92 international conference. In 1998, he chaired the program committee of COLING-ACL'98, with P. Whitelock as ACL co-chair. He is frequently asked to serve as a reviewer (about 30 conference papers and 2-3 journal papers each year).
He has been principal investigator for several industrial research contracts (about 150 contract reports). At the international level, he has also participated in or led GETA's involvement in several cooperative research efforts. He has been invited or visiting researcher in several laboratories, notably TAUM (Montréal, 1 year), MGPIIA (Moscow), SFB-100 (Saarbrücken), UTMK (Penang, 6 months in total), KDD (Tokyo), NII (Tokyo), and ATR (almost 2 years in total), where he learned about speech translation before starting it in France in 1995.
His current research interests include multilingual interpersonal communication over networks (UNL project); machine-assisted human translation; speech translation; personal dialog-based MT for monolingual authors; portable and readable encodings for multilingual documents; computational tools for post-editors and occasional translators; contributive multilingual lexical data bases; specialized languages and environments for linguistic engineering and research; contributive construction of NLP resources; participative translation; and task-related MT evaluation. He works now on Ariane-Y, a new "MT factory" able to integrate expert (rule-based) as well as empirical (SMT, EBMT) phases.

Objectives of the lectures
Learn about essential aspects of automatic and automated translation. The first lecture will introduce several fundamental facts about MT, MT systems, and their variety. In particular, it will show that it is far too simplistic to divide MT systems into "statistical" and "rule-based", and that the popular belief among researchers that all operational MT systems are now statistical is totally erroneous. The next three lectures will present a deeper analysis of the linguistic, computational and operational (engineering) architectures of MT systems. The last lectures will go deeper in two important questions: how to evaluate MT "in operation", and how to model and handle "parallel corpora for modern MT", which (will) have to be multilingual, multi-annotated, and multimedia.

Detailed description:
Lecture 1: Linguistic, computational and operational architectures of MT systems
Apr. 14, 2pm - 4pm
This lecture will center on the following points.
. The various tasks of MT and their difficulty.
. The CxAxQ "MT theorem" (Coverage x Automaticity x Quality << 100% while 2 factors can approach 100%) and the intrinsic difficulty of HQ translation.
. The independence of linguistic and computational architectures of MT systems. That claim will be supported by numerous examples of MT systems adopting 1 of the 13 existing linguistic architectures.
. Some aspects of the operational architecture of MT systems and their influence on the choice of linguistic and computational architectures.
. The current evolution towards "deeper" linguistic architectures and hybrid computational architectures, coupled with user involvement.

Lecture 2: Linguistic architectures of MT systems
Apr. 16, 2pm - 4pm
This lecture will center on the following points.
. Characteristics of various possible intermediate representations.
. Monolevel and multilevel structures.
. Units of translation: segments, infrasegments, supersegments, whole documents?
. Different sorts of "deep pivots" (hybrid, semantico-linguistic, semantico-pragmatic).
. Pros and cons of various linguistic architectures w.r.t. "translational situation" (a part of the operational context).

Lecture 3: Computational architectures of MT systems
Apr. 21, 2pm – 4pm
This lecture will center on the following points.
. Taxonomy of computational architectures (empirical vs. expert)
. Taxonomy of algorithmic methods (how to fight non-determinism, fuzziness and noise of correspondences between successive levels of description) .
. Empirical computational architectures (statistical, example-based with and without annotations).
. Expert computational architectures ("procedural" methods, "rule-based" methods).
. Rules of well-formedness (grammars), transition rules (automata), rewriting rules (on strings or trees).
. Examples of SLLPs (Specialized Languages for Linguistic Programming) of 3 types (creation, addition, substitution).
. Ensuring and relaxing decidability of SLLPs. Examples: ATNs, Q-systems, ATEF, ROBRA, GRADE.

Lecture 4: Engineering of MT and CAT systems (CAT = MT + TA, TA = translation aids)
Apr. 23, 2pm – 4pm
This lecture will center on the following points.
. From homogeneous MT systems to heterogeneous CAT systems.
. Convergence of evolutions since 1980.
. Environments for Developing Lingware (EDL): EDL specific to 1 MT system (MTS), Meta-EDL for heterogeneous MTS, towards Integrating EDL.
. Good practices in implementing specialized languages (SLLPs).
. Design of Lexical Data Bases for MT systems.
. PIVAX for heterogeneous MTS sharing a common leical pivot.

Lecture 5: Evaluation of MT and CAT systems for various operational architectures
Apr. 28, 2pm – 4pm
The topic of MT evaluation began to be studied in the early 50's and has taken new turns since the advent of empirical (most notably probabilistic) methods in MT. Ch. Boitet has published last year an article with Hervé Blanchon on MT evaluation in the TAL journal. This lecture will draw from that article and center on the following points:
. similarities and differences in text and speech MT evaluation.
. arguments and proposals for task-related measures.
. integrating evaluation as a "no-cost" measure in actual (needed) translation and post-edition tasks.

Lecture 6: Corpora for hybrid MT/CAT systems
Apr. 30, 2pm – 4pm
The author has published last year in RFLA, Revue Française de Linguistique Appliquée (French Journal of Applied Linguistics) a paper on that topic, which is currently studied in depth by C.P. Huynh (predoctoral fellow at NII for 5 months) for his PhD. A main idea is that corpora for MT cannot be reduced to collections of "bi-segments", or source-target translation pairs.
This lecture will center on the following points.
. The notion of "segment" varies with systems, so that corpora should be "multi-segmented".
. Translation units may be segments as well as infra-segments and super-segments.
. Segments may recursively contain subdocuments (e.g., text in balloons).
. Context is important, and has several aspects (linguistic, situational, dialogic), all crucial to solve some important problems (anaphora, ellipsis, tense agreement, gender of addressee, politeness expression, etc.).
. Varied and complex annotations have to be attached to segments and possibly to higher hierarchical levels (paragraphs, sections, etc.). They can concern 1 language at a time (e.g. POS or linguistic trees), or 2 languages (various alignments), or all languages (semantic or pragmatic representations).
. In the case of speech translation, sound files and various transcriptions have to be handled.
. All that justifies the development of systems, programmable at different specialization levels, to operate on translation corpora. This will be illustrated by the SECTra_w system built by Huynh C. P., and its evolution.

Lecture Series: Knowledge Representation and Reasoning for Systems Biology

Melody — 2009-03-23T14:00:05+09:00

Lecture Series:Knowledge Representation and Reasoning for Systems Biology
Lecturer: Assoc. Prof. Andrei Doncescu
Place: National Institute of Informatics,
            March 2 (Mon): Presentation Room, 19F
            March 10 (Tue): Meeting Room, 20F
            March 23, 30 April 6 (Mon): Lecture Room, 20F
Date: March 2, 10, 23, 30, April 6
Time: 2pm - 4pm
Fee: Free
Registration: Not required

Overall schedule:
Lecture 1: Fuzzy Logic for Engineering : A Tutorial
Lecture 2: 1. Plausible Reasoning
                2. Quantities Measurements: time, space, minds
Lecture 3: Knowledge Basis Discovery by Data Mining Techniques
Lecture 4: The Analytical Modelling in Medicine
Lecture 5: Logic, Knowledge and Computation for Systems Biology

Abstract:
Andrei Doncescu received his Ph.D from the Department of Computer Science & Electric Engineering University of Poitiers, France.
He is current Associate Professor in the University Paul Sabatier and Research Leader of Medical Diagnosis Team in Laboratory of Architecture and Analysis of Systems Toulouse France.
His research activity is oriented Knowledge Basis Discovery for Modeling of Complex Systems and more of 150 papers have been published in this research field.
He serves on the editorial boards of both IEEE and non-IEEE journals.
He received in 2005 and 2007 outstanding awards from the IEEE society and in 2006 the Best Presentation Awards of International Conference in Soft Computing and Intelligent System.
He and Prof. Katsumi Inoue (NII) are the Leaders of the JST-CNRS Project Knowledge-based Discovery in Systems Biology.

Detailed description:
Lecture 1: Fuzzy Logic for Engineering : A Tutorial
March 2 (Mon): Presentation Room, 19F
A fuzzy logic system (FLS) is unique in that it is able to simultaneously handle numerical data and linguistic knowledge. This tutorial provides a guided tour through, those aspects of fuzzy sets and fuzzy logic that are necessary to synthesis a fuzzy logic systems. Because the Biological Systems are for the most part causal, the causality is impose as a constraint of the development of FLS. In the end of this tutorial we show the manner to update a FLS by incorporate knowledge in a unified mathematical manner.

Lecture 2: 1. Plausible Reasoning
2.1.1. Classical Probability Theory
2.1.2. Statistical Inference
2:2. Quantities Measurements: time, space, minds
March 10 (Tue): Meeting Room, 20F
The first part of the lecture reminds the basic notions of probability theory. After this short introduction we give an overview of the utility of thermodynamics at the molecular level to understand proteins and receptor-ligand binding.

Lecture 3: Knowledge Basis Discovery by Data Mining Techniques
3.1. From Clustering to Regression
3.2. Application on Systems Biology
March 23, (Mon): Lecture Room, 20F
Computational models have been playing a significant role for the computer-based analysis of biological and biomedical data. High-throughput technologies are opening global perspectives for analyzing living organism at the molecular level. In the first part of this lecture a variety of artificial intelligence technologies and statistical tools are presented to detect significant differences in gene expression levels. Using this type of approach we can infer regulatory interactions directly from data by fitting simple network models to large scale gene expression data and to extract the most well-determined interactions in the network.

Lecture 4: The Analytical Modelling in Medicine
March 30, (Mon): Lecture Room, 20F
Taking into account the current competences in the field of the classical automatic control, namely, analysis, observation and control of dynamical systems, linear, nonlinear with or without delay and/or propagation, the objective is to propose a «progressive transfer» of our «competences» towards the domains of the life sciences, and especially towards all the domains of life sciences in which a dynamic behavior can be pointed out. Every behavior of a biological system with respect to one (or several) time scaling can be interpreted in a dynamical system context by using the associated tools.

Lecture 5: Logic, Knowledge and Computation for Systems Biology
April 6 (Mon): Lecture Room, 20F
System biology provides a new approach to studying and analyzing the biological process. Biological Pathways represent a key sub-system level of organization. The aim of the biological pathway is to map and understand the cause-effect relationship in the complex interactions of biological systems. Complex diseases such as cancer have multiple origins and are therefore difficult to understand and cure. We present the use of logic relationship to model breast cancer gene expression networks with mRNA microarray data as a challenge. Some of these challenges are discussed.

Lecture Series: Wireless Networks and Pervasive Services Technologies: Fundamentals and Recent Advances

Melody — 2009-03-17T14:00:38+09:00

Lecture Series: Wireless Networks and Pervasive Services Technologies: Fundamentals and Recent Advances
Lecturer: Dr. Kun Yang, University of Essex, UK.
Place: National Institute of Informatics, 20F, meeting room
Date: March 16, 17, 18, 19, 23, 24, 26
Time: 2pm - 4pm
Fee: Free
Registration: Not required

Overall schedule:
Lecture 1: Fundamentals on Computer Networks and Wireless Networks (March 16)
Lecture 2: A Tale of Two Technologies: WiMAX vs. LTE (March 17)
Lecture 3: An Enabler for Vehicular Networks: from IEEE 802.11 to IEEE 802.16 (March 18)
Lecture 4: Wireless Network Convergence and Fixed Mobile Convergence (March 19)
Lecture 5: Wireless Sensor Networks (March 23)
Lecture 6: Pervasive Computing – Enabling Technologies (March 24)
Lecture 7: Pervasive Services Engineering: from Outside to Inside (March 26)

Short Bio of the Lecturer
Dr. Kun Yang received his PhD from the Department of Electronic & Electrical Engineering, University College London (UCL), UK. He is currently a Reader in the School of Computer Science and Electronic Engineering, University of Essex, UK. Before joining in University of Essex at 2003, he worked at UCL on several European Union research projects, such as MANTRIP, FAIN, CONTEXT. His main research interests include wireless networks, heterogeneous wireless networks, fixed mobile convergence, IP network management and pervasive service engineering. He has published 100+ technical papers in mainstream journals and major international conferences. He manages research projects funded by various sources such as UK research funding body EPSRC, industries such as British Telecom, and European Union. He is a recipient of British Telecom Visiting Research Fellowship 2007. He serves on the editorial boards of both IEEE and non-IEEE journals (such as Wiley Wireless Communications and Mobile Computing, Springer Telecommunication Systems, alongside other four). He is a Senior Member of IEEE.

Lecture Series: Wireless Networks and Pervasive Services Technologies: Fundamentals and Recent Advances
As wireless broadband networks become increasingly mobile and ubiquitous, users naturally require the services running on top of these networks to become pervasive, i.e., to be able to run anywhere, at any time and on any devices without or with little user intervention. Apart from presenting state-of-the-art fundamentals of these two inter-connected and active research areas, namely, wireless networks and pervasive services, this lecture series will also discuss new solutions as arising from the lecturer�fs research group, aiming to giving a balanced overall view of the recent advances of the broad research area of wireless networks and pervasive services. The primary topics include: fundamentals of computer networks and wireless networks; individual wireless network technologies covering WiFi, WiMAX, cellular systems, wireless sensor networks; network convergence including heterogeneous wireless networks and the convergence of wireless (WiMAX) with fixed optical networks (PON: Passive Optical Networks); various enabling technologies for pervasive computing; and pervasive service engineering.

Detailed description:
Lecture 1: Fundamentals on Computer Networks and Wireless Networks
Monday 16^th March 2009, 2pm-4pm
This lecture firstly gives a snapshot of computer network fundamentals, including basic concepts such as protocols, services, addressing, different view of network infrastructure, and the whole procedure of end-to-end package transmission. Then, the lecture goes on to present wireless communication/network preliminaries, comprising physical layer multiplexing, signal and noise, BER (bit error rate), media access control, etc. A panorama view of the mainstream wireless network technologies will also be illustrated, covering cellular systems, WiFi (IEEE 802.11), WiMAX (IEEE 802.16) together with mobile ad hoc networks, mesh networks and wireless sensor networks. Mobility issues in wireless networks, largely based on IEEE 802.21 MIH (Media Independent Handover), will also be discussed about. Some relevant recent research areas such as cognitive radio.

Lecture 2: A Tale of Two Technologies: WiMAX vs. LTE
Tuesday 17^th March 2009, 2pm-4pm
This lecture aims to providing a deeper look into two wireless technologies: WiMAX and B3G mobile cellular systems. The physical layer, MAC (Media Access Control) layer, QoS (Quality of Service) and scheduling of WiMAX standards will be presented, in close comparison with WiFi. Mobile cellular system part will cover 3G system infrastructure (using UMTS as an example), its development into 3GPP Long Term Evolution (LTE). A comparison between WiMAX and LTE, which represent wider area wireless broadband solution from IT companies and traditional telco companies respectively, will be conducted. Some emerging technologies such as femto cell etc will also be discussed.

Lecture 3: An Enabler for Vehicular Networks: from IEEE 802.11 to IEEE 802.16
Wednesday 18^th March 2009, 2pm-4pm
This lecture tunes into the application of wireless network technologies to particular domain: vehicular communications. Much work has been conducted to provide a common platform facilitating inter-vehicle communications or intelligent transportation systems largely utilizing IEEE 802.11-based technology such as WAVE (Wireless Access for Vehicular Environment). This talk explores another means to enable vehicular networks, namely, IEEE 802.16 or WiMAX, with particular focus on another type of vehicular network service: the Internet access from mobile vehicles on highways. A description of how the system operates will be presented. A particular handover scheme in the context of the newly published IEEE 802.16j will also be delivered.

Lecture 4: Wireless Network Convergence and Fixed Mobile Convergence
Thursday 19^th March 2009, 2pm-4pm
This lecture looks into network convergence, including both the convergence of different wireless network technologies and the convergence of wireless mobile networks with fixed networks such as passive optical networks (PON). The wireless convergence part presents how IEEE 802.11-based ad hoc networks can be introduced into cellular networks to improve the overall performance of the latter both in terms of voice calls and data services. The fixed mobile convergence (FMC) part introduces how to use PONs (in particular Ethernet PONs) to backhaul wireless network traffics. Going beyond the conventional way of optical wireless convergence such as radio over fibre (RoF), this lecture is to investigate into a joint bandwidth allocation algorithm between WiMAX and EPON, which lies at the MAC layer and serves as an alternative to RoF. Some experimental results on Essex test-bed are shown.

Lecture 5: Wireless Sensor Networks
Monday 23^rd March 2009, 2pm-4pm
Following an introductory presentation of the architecture, the protocol stack and major research issues of wireless sensor networks (WSNs), this lecture focuses on topology control and power assignment problems in WSNs and presents how modern heuristics such as genetic algorithms and local search, after being flavoured with network specific flavours, can be utilized to solve these problems in an offline manner at sensor deployment stage. A TDMA-based media access control will also be presented.

Lecture 6: Pervasive Computing – Enabling Technologies
Tuesday 24^th March 2009, 2pm-4pm
This lecture starts with an illustration of the evolution of computing from mainframe through client-server (including peer-to-peer) to pervasive computing. Following discussions on research areas of pervasive computing, the lecture zooms in to some fundamental technologies to help build pervasive computing systems. These are grouped into three categories: 1) web-based such as extending web beyond HTML – content/XML, extending web beyond HTTP - processing, and web services; 2) distributed object technologies such as Sun Java RMI and OMG CORBA, and 3) active code migration technologies such as active/programmable networks, mobile agents. Instead of presenting the details of every single technology, this lecture focuses on discussions on the relationships and evolution of these technologies in the context of developing adaptive and pervasive systems.

Lecture 7: Pervasive Services Engineering: from Outside to Inside
Thursday 26^th March 2009, 2pm-3:30pm
This lecture looks into an active branch of pervasive computing, pervasive services. Here a pervasive service is a piece of adaptive service that has the ability to run anywhere, at any time and potentially on any device. It starts with introduction of some basic concepts and principles of pervasive service engineering such as context-awareness, and then it proceeds to focus on one stage of the lifecycle of pervasive service engineering: service creation, which is an area that is largely neglected when providing service adaptability due to its static nature. An alternative approach to the existing literature is discussed, which advocates a shift of focus from comprehensive and complex middleware supporting environment for pervasive services, which is outside of pervasive services themselves, to the internal logic of pervasive services. In this approach, a combination of policy-based management method and model-driven architecture (MDA) is proposed for service creation. Presentation on how this approach makes positive impact on service discovery, which is a stage necessary for run-time service composition and service adaptation, is also given. Deployment issues as to how to run pervasive services on resource-constrained mobile devices are also discussed and corresponding solutions proposed.

Lectures on Numerical Analysis by Professor Eiermann

Melody — 2009-03-03T10:00:07+09:00

You are all welcome to attend the following lectures given by
Professor Michael Eiermann
Institut fuer Numerische Mathematik und Optimierung
TU Bergakademie Freiberg, Germany

Place: Lecture room 1 (1212), 12the floor (except for Feb. 25(Wed))
   Seminar room 1 (2006), 20th floor (on Feb. 25(Wed))
   National Institute of Informatics
   http://www.nii.ac.jp/introduce/access1.shtml

Time: All lectures: 10am-12am.

I. Computational Aspects of the Stochastic Finite Element Method
   1.Feb. 12(Thurs)
   2.        13(Fri)
   3.        16(Mon)
   4.        17(Tues)

II. Krylov Subspace Methods for the Evaluation of Matrix Functions
   ---Applications and Algorithms
   1.Feb. 23(Mon)
   2.        24(Tues)
   3.        25(Wed)
   4.        26(Thurs)
   5.        27(Fri)
   6.March 2(Mon)
   7.        3(Tues)
   8.        4(Wed)

Abstract: see attached
I(PDF)
II(PDF)

Contact: Ken Hayami, National Institute of Informatics ( hayami@nii.ac.jp)

Lecture:A Class of Preconditioners Based on Splitting for Nonsymmetric System of Linear Equations

Melody — 2009-02-12T14:00:29+09:00

Time: February 12(Thursday) 2:00-3:00pm

Place: Lecture room 2 (1213), 12th floor, National Institute of Informatics
http://www.nii.ac.jp/introduce/access1.shtml

Speaker: Dr. Jun-Feng Yin
Associate Professor, Department of Mathematics,
Tongji University, Shanghai, China

Title: A Class of Preconditioners Based on Splitting for Nonsymmetric System of Linear Equations

Abstract:
Given a matrix splitting $A=M+N$ where $M^{-1}$ is easily computed, a factorization method is given by rank-one updating from $M^{-1}$ step by step using the Sherman-Morrison formula. Theoretical analysis shows that under some special choices, the factorization can give a $LDU$ decomposition of $A$. Combined with dropping rules, the factorization leads to a class of preconditioner in terms of factorized unit upper triangular matrices, diagonal matrix and $M$. Numerical experiments from discrete convection diffusion equation and practical problems show that the new preconditioner is efficient, and comparable to exist preconditioners from storage requirement and computational cost point of view.

Contact: Ken Hayami (National Institute of Informatics), hayami@nii.ac.jp

International Symposium on Asian Speech Resources

Melody — 2008-11-28T10:45:02+09:00

Lecture Series: Emerging Topics in Databases

Melody — 2008-11-27T16:30:22+09:00

Lecture Series: Emerging Topics in Databases
Lecturer: Associate Prof. Vincent Oria, New Jersey Institute of Technology University
Place: National Institute of Informatics, 20F, meeting room 1 (2009)
Date: November 11, 12, 21, 25, 27
Time: 1pm - 4pm
Fee: Free
Registration: Not required

Overall schedule:
Lecture 1: Multimedia Databases (Nov. 11)
Lecture 2: Stream Data management (Nov. 12)
Lecture 3: Peer-to-Peer Data management (Nov. 21)
Lecture 4: Mobile and Ubiquitous Databases (Nov. 25)
Lecture 5: Database Security (Nov. 27)

Biography:
Vincent Oria is currently an associate professor of computer science at the New Jersey Institute of Technology. His research interests are multimedia databases, moving object databases and database security. He has held a visiting research scholar at National Institute of Informatics (Tokyo, Japan), and a visiting professor or researcher at FhG-IPSI (Darmstadt, Germany), ENST (Paris, France), University of Paris-IX Dauphine (Paris, France), INRIA (Roquencourt, France) and CNAM (Paris France). He has served on a number of conference program committees.

Objectives:
Learn about some of the new trends in databases, and compare different database architectures and query processes beneficial for certain types of applications. This series of lecture will discuss topics that are of growing importance in both the database research community and industry. The topics that will be discussed are Multimedia database, Streaming Data Management, Peer-to-peer data management, Mobile and ubiquitous data management, and Database security and privacy. Each topic will be covered in a 3 hour lecture.

Detailed description:
Lecture 1. Multimedia Databases
Nov. 11, 1pm - 4pm
Multimedia in principle means data of more than one medium. Multimedia data usually refers to data representing multiple types of medium to capture information and experiences related to objects and events. Commonly used forms of data are numbers, alphanumeric, text, images, audio, and video. A multimedia database should be a database system that can help manage data of multiple type of medium. Due to the complexity and differences of multimedia data, each type of medium is in practice handled differently. In this lecture, I will discuss issues related to building a real multimedia database system where multiple media types can coexist.

Lecture 2. Stream Data management
Nov. 12, 1pm - 4pm
Stream data management refers to novel application needs where a large amount of data has to be processed and analyzed in real time. It differs from business activity monitoring in that the client of a stream processing application is often a program, rather than a human. Currently, stream processing is widely used in computing real-time analytics in e-trading, maintaining the state of massively multi-player Internet games, real-time risk analysis, network monitoring, and national security applications. In the future, the declining cost of sensor technology will create new markets for this technology. The lecture will discuss application needs, system architecture, query possessing and open issues.

Lecture 3. Peer-to-Peer Data management
Nov. 21, 1pm - 4pm
Peer-to-peer (P2P) computing consists of an open-ended network of distributed computational peers, where each peer can exchange data and services with a set of other peers, called acquaintances. Peers are fully autonomous in choosing their acquaintances. We can also assume that there is no global control in the form of a global registry, global services, or global resource management, nor a global schema or data repository. What are the data management issues raised by this case, knowing that each peer may have data to share with other peers. This lecture will present solutions proposed for data models, architecture for a prototype implementation, and discuss open research questions

Lecture 4. Mobile and Ubiquitous Databases
Nov. 25, 1pm - 4pm
In Ubiquitous databases, the data are physically attached to real world "objects". Imagine a health care system where every patient keeps his/her medical records for example. In this scenario, how to share these data among different organizations consistently and effectively and efficiently retrieve them? Another application is moving object databases. A moving object is essentially a time dependent geometry. One can distinguish between moving objects for which only the time dependent position is of interest and those for which also shape and extent are relevant and may change over time. Querying these data raises some issues in the processing and indexing as the data is not stable.

Lecture 5. Database Security
Nov. 27, 1pm - 4pm
The database server is certainly the most important server in a company as it stores client information, product and service information, financial information, human resource details etc. The database server contains the data that keeps a company running. When the database server is managed locally by the company, there are some solutions in place that work assuming that the database administrator can be trusted. When a company outsources its database is there a solution that can allow the company to retrieve data while keeping intruders including the database administrator from having access to the database content?

International Symposium on Physics of Quantum Technology

Melody — 2008-11-25T16:12:50+09:00

Lecture:Constraint Preconditioners for Symmetric Indefinite Matrices

Melody — 2008-11-17T11:00:23+09:00

Date: November 17th (Monday) 11-12am

Place: NII 20F, Seminar Room 1 (2006)

Title: Constraint Preconditioners for Symmetric Indefinite Matrices

Speaker:
Professor Zhong-Zhi Bai
Institute of Computational Mathematics and Scientific/Engineering Computing,
Chinese Academy of Sciences

Abstract:
We study the eigenvalue bounds of block two-by-two nonsingular and symmetric indefinite matrices whose (1,1) block is symmetric positive definite and (2,2) block is symmetric indefinite. Constraint preconditioners for these matrices have been constructed by simply replacing the (1,1) block by a symmetric and positive definite approximation, and the spectral properties of the preconditioned matrices have been discussed.
In this talk, we will review a few known results and give some new ones on this topic. Numerical results show that for a suitably chosen (1,1) block matrix the constraint preconditioners outperform the block-diagonal and the block-tridiagonal ones in iteration step and computing time, when they are used to accelerate the GMRES method for solving the block two-by-two symmetric positive indefinite linear systems.

Lecture Series: Principles of Parallel Programming

Melody — 2008-10-15T16:35:16+09:00

Lecture Series: Principles of Parallel Programming
Lecturer: Prof. Lawrence Snyder, University of Washington
Place: National Institute of Informatics, 12F, conference room (1208)
Date: September 17, 24, October 1, 8, 15
Time: 1pm - 3pm
Fee: Free
Registration: Not required

Overall schedule:
Lecture 1: The New Opportunities and Challenges of Parallelism (Sep. 17)
Lecture 2: 35 Years of Research: Positive Results; Negative Results (Sep. 24)
Lecture 3: A Model Of Parallelism To Guide Thinking (Oct. 1)
Lecture 4: Parallel Languages of Today -- OpenMP to Fortress (Oct. 8)
Lecture 5: Next Parallel Languages -- Access To Parallelism For All (Oct. 15)

Detailed description:
Lecture One: The New Opportunities and Challenges of Parallelism
September 17, 1pm - 3pm
The fastest computer in the world has achieved a speed of 10¹⁵ floating point operations per second; all desktop and laptop computers sold today are parallel computers. What programming techniques can be used to effectively translate the potential parallelism in a computation to these kinds of computers? Will one language work for both situations? Should all programmers be parallel programmers? The lecture discusses answers to these questions as well as other urgent problems in parallel computation.

Lecture Two: 35 Years of Research: Positive Results; Negative Results
September 24, 1pm - 3pm
Parallel programming has been an intensively studied topic since the development of Illiac 4. But after 35 what has been learned? Researchers new to the subject think nothing from the past applies; those who have worked in the area for a long time, believe much is known. What positive results do we have -- what works? What negative results do we have -- what doesn't work? What survives to build on? How should the past inform our research agenda?

Lecture Three: A Model Of Parallelism To Guide Thinking
October 1, 1pm - 3pm
Sequential programming is different from parallel programming. In the lecture, I discuss some of the ways. Some issues include: What can we take from sequential programming and apply to the parallel programming problem? What goals should any new language have? What is to be done with the millions of lines of legacy code that must continue to run -- can it run in parallel. How can parallel architecture research help? A new model of parallel computation suitable for programming language design will be introduced.

Lecture Four: A Model Of Parallelism To Guide Thinking
October 8, 1pm - 3pm
Parallel programming language research continues to be an active topic. New languages are being implemented all of the time. We briefly touch on the languages presently in use: MPI, PVM, OpenMP. We consider the advancement provided by the Partitioned Globabl Address Space (PGAS) languages: Co-Array Fortran, UPC, Titanium. Further we consider the advancements of the new HPC languages: Chapel, X10 and Fortress. Special purpose languages like Cuda will also be touched on. What kind of language do the developers of desktop applications need?

Lecture Five: Next Parallel Languages -- Access To Parallelism For All
October 15, 1pm - 3pm
With the benefit of the preceding lectures, we consider the future directions of parallel programming. How must the languages adapt? How must our teaching of programming and computer science adapt to the future programming world. How does a 20-year veteran programmer become proficient in parallel programming, or is that necessary? Where should research be directed to answer these questions?

IVA 2008(8th International Conference on Intelligent Virtual Agents)

Melody — 2008-09-01T17:24:45+09:00

Lectures on Numerical Linear Algebra

Melody — 2008-07-24T16:57:17+09:00

Lectures on Numerical Linear Algebra

You are all welcome to attend the following lectures.

Place: National Institute of Informatics
�@�@�@�@http://www.nii.ac.jp/introduce/access1-j.shtml

Contact: Ken Hayami
(hayamiATnii.ac.jp)
------------------------------------------------------------------

1. July 3rd (Thursday) 11:00-12:00am

Room: 20F Seminar Room 1 (2006)

Speaker: Mr. Stefan Guettel
Ph.D. student
Institute of Numerical Analysis and Optimization
Technische Universitaet Bergakademie Freiberg, Germany

Title: Matrix functions and their numerical approximation in Krylov spaces

Abstract:
Matrix functions $f$ are canonical generalizations of scalar functions
with applications in many areas. For a large matrix $A$, most often it is
the action of the matrix function on a vector $b$ which is required, i.e.,
approximate $f(A)b$. For this task, Krylov methods turn out to be very
effective. After a short introduction to matrix functions, we will present
various recent results on restarted, thick-restarted and rational Krylov
methods.

2. July 9th (Wednesday) 11:00-12:00am

Room: 12F Lecture Room 1 (1212)

Speaker: Professor Yimin Wei
Department of Mathematics, Fudan University, P.R. of China

Title: On mixed and componentwise condition numbers for
Moore-Penrose inverse and linear least squares problems

Abstract:
Classical condition numbers are normwise: they measure the size of both
input perturbations and output errors using some norms. To take into
account the relative of each data component, and a possible data
sparseness, componentwise condition numbers have been increasingly
considered. These are mostly of two kinds: mixed and componentwise.
In this talk, we give explicit expressions, computable from the data,
for the mixed and componentwise condition numbers for the computation of
the Moore-Penrose inverse as well as for the computation of solutions and
residues of linear least squares problems. In both cases the data matrices
have full column (row) rank.

3. July 18 (Friday) 2:00-3:00pm

Room: 20F, Meeting Room 1 (2009)

Speaker: Professor Yimin Wei
Department of Mathematics, Fudan University, P.R. of China

Title: Super-Large Sparse Matrix Computations in Web Information Retrieval

Abstract:
We will discuss the following issues of PageRank:
1) Hybrid methods for speeding up the computation of PageRank;
2) Arnoldi vs. GMRES for PageRank;
3) Jordan canonical form of the Google matrix;
4) Future work.

4. July 24th (Thursday) 11:00-12:00am

Room: 12F, Lecture Room 1 (1212)

Speaker: Professor Michael Ng
Department of Mathematics, Hong Kong Baptist University

Title: A Fast Total Variation Minimization Method for Image Restoration

Abstract:
In this talk, we study a fast total variation minimization method for
image restoration. In the proposed method, we use the modified total
variation minimization scheme to denoise the deblurred image.
An alternating minimization algorithm is employed to solve the proposed
total variation minimization problem. Our experimental results show that
the quality of restored images by the proposed method is competitive
with those restored by the existing total variation restoration methods.
We show the convergence of the alternating minimization algorithm and
demonstrate that the algorithm is very efficient.

-----------------------------------------------------------------

Lecture Series: Image Processing and Pattern Recognition: Fundamentals and Applications

Melody — 2008-06-16T16:19:53+09:00

Lecture Series: Image Processing and Pattern Recognition: Fundamentals and Applications
Lecturer: Prof. Frank Y. Shih, New Jersey Institute of Technology
Place: National Institute of Informatics, 12F, conference room (1208, 1210)
Date: June 16-20, 24
Time: 2pm-4:30pm
Fee: Free
Registration: Not required

Overall schedule:
Lecture 1: Image Fundamentals and Enhancement (June 16)
Lecture 2: Mathematical Morphology (June 17)
Lecture 3: Image Segmentation and Representation (June 18)
Lecture 4: Pattern Representation and Recognition (June 19)
Lecture 5: Image Watermarking and Steganography (June 20)
Lecture 6: Face, Document, and Solar Image Applications (June 24)

Biography:
Professor Frank Shih received Ph.D. from Purdue University. He is presently a full professor in CS Department, New Jersey Institute of Technology, USA. He is the Director of Computer Vision Laboratory. He held a visiting professor position at Princeton University, Columbia University, and National Taiwan University. He is an internationally well-known scholar and serves on the Editorial Board of the International Journal of Pattern Recognition, the International Journal of Pattern Recognition Letters, the International Journal of Pattern Recognition and Artificial Intelligence, the International Journal of Recent Patents on Engineering, the International Journal of Recent Patents on Computer Science, the International Journal of Internet Protocol Technology, and the Journal of Internet Technology. He served as a steering member, committee member, and session chair for numerous professional conferences and workshops. He has received numerous grants from the National Science Foundation, Navy and Air Force, and Industry. He authored a book on Digital Watermarking and Steganography, and is writing two books on Image Processing and Pattern Recognition. He has published eight book chapters and over 190 technical papers. His current research interests include image processing, computer vision, watermarking and steganography, sensor networks, pattern recognition, bioinformatics, information security, robotics, fuzzy logic, and neural networks.

Detailed description:

Lecture Series: Image Processing and Pattern Recognition: Fundamentals
and Applications

Images are used for a variety of purposes, including entertainment, medical, business, industrial, military, civil, security, and scientific. The interests in digital image processing stem from the improvement of pictorial information for human interpretation and the processing of scene data for autonomous machine perception. Many new practical image processing, watermarking, and pattern recognition techniques will be introduced in this lecture series to illustrate the framework that provides assistance and tools in understanding and implementing the fundamental principles. The topics include image fundamentals and enhancement, mathematical morphology, image segmentation and representation, feature extraction, pattern recognition, image watermarking and steganography, face recognition, document processing, and solar image analysis.

Lecture One: Image Fundamentals and Enhancement
June 16, 2pm-4:30pm

In this lecture, I will introduce image fundamentals and mathematical preliminaries that are often used in image processing, including Laplace transform, Fourier transform, z-transform, cosine transform, and wavelet transform. I will also introduce the commonly used image enhancement techniques, including gray scale transformation, piecewise linear transformation, bit plane slicing, histogram equalization, histogram specification, enhancement by arithmetic operations, smoothing filter, sharpening filter, image blur types, and quality measures.

Lecture Two: Mathematical Morphology
June 17, 2pm-4:30pm

Mathematical morphology can extract image shape features, such as edges, fillets, holes, corners, wedges, and cracks, by operating with various shaped structuring elements. In industrial vision applications, mathematical morphology can be used to implement fast object recognition, image enhancement, segmentation, and defect inspection. In this lecture, I will introduce binary morphology, opening and closing, hit-or-miss transform, grayscale morphology, morphological edge operator, alternating sequential filters, recursive morphological operation, soft morphological operation, and general sweep morphological operation.

Lecture Three: Image Segmentation and Representation
June 18, 2pm-4:30pm

In this lecture, I will introduce a number of image segmentation techniques. They include thresholding, component labeling, locating object contours by the snake model, edge detection, linking edges by adaptive mathematical morphology, automatic seeded region growing, and top-down region dividing. I will also discuss different region representation schemes, such as run-length coding, binary tree and quadtree, skeleton, and shape number, and different boundary representation schemes, such as chain code, crack code and midcrack code, fitting-line segments, and Fourier descriptors.

Lecture Four: Feature Extraction and Pattern Recognition
June 19, 2pm-4:30pm

In this lecture, I will introduce the methods of feature extraction in image processing and analysis. They include Fourier descriptor and moment invariants, shape number, corner and circle detection, Hough transform, principal component analysis, linear discriminate analysis, and feature reduction. I will describe distance transformation and shortest path planning. I will also present different pattern recognition methods, including the unsupervised clustering algorithm, support vector machine, neural networks, the adaptive-resonance-theory (ART) network, fuzzy sets, and image analysis.

Lecture Five: Image Watermarking and Steganography
June 20, 2pm-4:30pm

With the fast increasing volume of electronic commerce web sites and applications, intellectual property protection is an extremely important concern for content owners who exhibit digital representations of photographs, books, manuscripts, and original artwork on the Internet. Moreover, as available computing power continues to rise, there is an increasing interest in protecting video files from attack. The applications are widely spread on electronic publishing, advertisement, merchandise ordering and delivery, picture galleries, digital libraries, online newspapers and magazines, digital video and audio, personal communication, etc.

In this lecture, I will introduce watermarking classification, spatial domain watermarking, frequency domain watermarking, fragile watermark, robust watermark, and combinational domain digital watermarking. I will also introduce types of steganography, applications of steganography, embedding security and imperceptibility, examples of steganography software, and genetic algorithm based steganography.

Lecture Six: Face, Document, and Solar Image Applications
June 24, 2pm-4:30pm

In this lecture, I will introduce the application of image processing and pattern recognition on face images, including face and facial feature extraction using SVM, the extraction of head and face boundaries and facial features, the recognition of facial action units, and facial expression recognition. I will also introduce block segmentation and classification, rule-based character recognition subsystem, logo identification, fuzzy typographical analysis for character preclassification, and fuzzy model for character classification. I will present the applications of machine learning techniques to automated detection and classification of solar events.

HIS (11): Seminar on Sensors & Monitoring

Melody — 2008-05-28T18:11:06+09:00

Students, reasearchers and other interested people are welcome to attend the following seminar.

Dr Platon (JSPS fellow, NII) will present his research on sensor networks at the National Institute of Informatics.

Schedule: Wednesday 28 May 2008 (10:00-11:00)
Location: National Institute of Informatics (Tokyo), 19F, Room 1904
Language: English
Registration fees: None
Seminar page: http://horizons.free.fr/his/eng/seminars.htm
Need for Decentralized Security in Wireless Sensor Networks
(10:00-11:00)
by Dr PLATON Eric

Abstract:Systems involving wireless sensor networks usually have two main assets, namely the sensor data in transit over the network, and hardware nodes. Traditional security approaches protect these assets by physical means and resource-consuming software like firewall or intrusion detection sub-systems. These are no longer applicable "as-is" in wireless sensor networks, where resources are severely constrained and nodes are physically assumed vulnerable. In addition, the applications of wireless sensor network technologies refer to less than a dozen to more than a thousand nodes, depending on application requirements. This indicates a need to shift the traditional view on security from the perspective of strong individuals to one that relies on the correlation and cooperation of weak individuals.

In this presentation, I will introduce wireless network technologies, review their security issues, and present state-of-the-art achievements, along with their limitations. This description of the current landscape of security will show the potential for decentralized approaches, leveraging the number of nodes instead of their individual capabilities.
I will finally present our approach at �{�ʓc�� (Honiden laboratory), focusing on our activities related to decentralized key management and secure routing.

Speaker: Eric Platon is a visiting researcher at the National Institute of Informatics), funded by ��{�w�p�U�� (Japan Society for the Promotion of Science). He holds a joint Ph.D. in Computer Science from Paris VI and ��w�@��w (The Graduate University for Advanced Studies) for work on exception management in multi-agent systems. His research interests pertain to security, self-adaptive programs, and exception handling in mobile systems, notably wireless ad hoc networks of sensors. He is involved as a software architect in the XAC project, a government-funded research project for secure and self-adaptive applications in embedded systems.

Restrictively preconditioned conjugate gradient methods

Melody — 2008-03-04T15:18:17+09:00

Date:
Tuesday 04 March 2008 (14:00-15:00)

Place:
National Institute of Informatics 12F Lecture Room 1 (1212)

Lecturer:
Professor Zhong-Zhi Bai
Institute of Computational Mathematics and Scientific/Engineering Computing, Chinese Academy of Sciences,

Theme:
Restrictively preconditioned conjugate gradient methods

Outline:
The restrictively preconditioned conjugate gradient (RPCG) method and its practical variant (PRPCG) for solving large sparse system of linear equations of a symmetric positive definite and block two-by-two structure are studied, and their convergence speeds are estimated.
Within this setting, we present algorithmic descriptions and convergence analyses of some special RPCG and PRPCG methods that employ the block Jacobi and the block symmetric Gauss-Seidel splitting matrices as approximations to certain matrix involved in the restrictive preconditioners to the coefficient matrices. Numerical results show that these RPCG and PRPCG methods are more robust and effective thanhasegawa
the classical preconditioned conjugate gradient methods.

Inquiry:
National Institute of Informatics Hayami Ken (hayami AT nii.ac.jp)