header

トップ   一覧 単語検索 最終更新   ヘルプ   最終更新のRSS

***** 専攻談話会(セミナー) [#t67d9871]

原則的に水曜日の13:30から15:00におこなっています。
客員教授、新任教員以外にも講演者を募集しています。
集中講義などでいらっしゃる先生など、お心あたりのある方は専攻幹事 [[鈴木:http://www.is.titech.ac.jp/~msuzuki4/]] (masahiro at is.titech.ac.jp) までお知らせください。
集中講義などでいらっしゃる先生など、お心あたりのある方は専攻幹事 青谷 (aotani at is.titech.ac.jp) までお知らせください。

2011年度談話会は[[CompViewセミナー:http://compview.titech.ac.jp/]]を兼ねます。

なお、談話会を開くときには、日時、会場、講演者、題目の情報を専攻幹事までメールにてお送りください。
**2015年度第6回 [#u80d28ed]

**2011年度第5回 [#sc26d268]
日時:''12月24日, 15:20-16:20''

日時: ''2011年7月6日(水)14:15-''
場所:''東工大大岡山キャンパス西八号館W809''

講演者:''Marvin K. Nakayama (New Jersey Institute of Technology)''

Title: ''Efficient Monte Carlo Methods for Risk Estimation and its Error, With Applications to Safety Analyses of Nuclear Power Plants''

Abstract: Analysts often measure risk through a quantile or a failure probability. The p-quantile of a continuous random variable is the constant for which p of the mass of its distribution lies below the quantile; e.g., the median is the 0.5-quantile. In project management, one may want to determine a time T such that the project has a 95% chance of completing by T, which is the 0.95-quantile. In finance, where a quantile is known as a value-at-risk, analysts 
frequently measure risk with the 0.99-quantile of a portfolio’s loss. For complex stochastic models, analytically computing a quantile is usually not possible, so Monte Carlo simulation is often employed. In addition to providing a point estimate for a quantile, we also want to measure the simulation estimate's statistical error, typically through a confidence interval (CI) for the quantile. Indeed, the U.S. Nuclear Regulatory Commission requires that licensees of nuclear power plants demonstrate compliance using a “95/95 criterion,” which entails ensuring (with 95% confidence) that a 0.95-quantile is less than a mandated limit. 

In this talk we present some methods for constructing CIs for a quantile estimated via Monte Carlo simulation. Unfortunately, simple random sampling can produce unusably wide CIs, so analysts may apply variance-reduction techniques (VRTs), such as importance sampling or control variates, in simulations to decrease the statistical error. We first discuss forming a quantile CI using a finite difference, and the second approach employs a procedure known as sectioning, which is closely related to batching. The asymptotic validity of both CIs follows from a so-called Bahadur representation, which shows that a quantile estimator can be approximated by a linear transformation of a probability estimator. We have established Bahadur representations for a broad class of VRTs, including antithetic variates, control variates, Latin hypercube sampling, and importance sampling. We present some numerical results comparing the different CIs.

This work is supported, in part, by the National Science Foundation through grants CMMI-0926949, CMMI-1200065, DMS-1331010, CMMI-1537322.

Speaker Bio: Marvin K. Nakayama is a professor of computer science at the New Jersey Institute of Technology. He received a B.A. in mathematics-computer science from U.C. San Diego, and an M.S. and Ph.D. in operations research from Stanford University. He won second prize in the George E. Nicholson Student Paper Competition sponsored by INFORMS, received a CAREER Award from the National Science Foundation, and won the Best Theoretical Paper Award at the 2014 Winter Simulation Conference. He is the Simulation Area Editor for the INFORMS Journal on Computing and an associate editor for the ACM Transactions on Modeling and Computer Simulation. His research interests include simulation, applied probability, statistics, risk analysis, reliability, and modeling of cascading failures.


**2015年度第5回 [#qc83d5d9]
日時:''11月16日15:00--16:00''

場所:''東工大大岡山キャンパス西八号館W904''

講演者:''Kostis Sagonas (Uppsala University)''

Title: ''Compiling Erlang to Native Code''

Abstract: For about fifteen years now, we been developing,
maintaining, and extending the native code compiler for the
concurrent functional programming language Erlang, to keep
up-to-date with additions at the language level, the runtime
system of Erlang/OTP, compiler technology advances, and new
interesting architectures.

In this talk, we will review the internals of HiPE (the
High-Performance Erlang) native code compiler and its ErLLVM
backend, which targets the LLVM compiler infrastructure.  Besides
presenting the overall architecture of HiPE and ErLLVM and their
integration in Erlang/OTP, we will try to critically discuss
technical challenges and decisions we took, and present a
detailed performance evaluation.

Time permitting, we may briefly introduce software tools for
Erlang
(Dialyzer, TypEr, tidier, PropEr, Concuerror and CutEr) built by the
speaker's group.


**2015年度第4回 [#fcc850d2]
日時:''11月24日15:00--16:30''

場所:''東工大大岡山キャンパス西八号館E1001''

講演者:''Grady Booch (IBM Research)''

Title: ''The Future of Software Engineering''

Abstract: No matter what future we may envision, it relies on
software that has not yet been written. Even now,
software-intensive systems have woven themselves into the
interstitial spaces of civilization, and we as individuals and as
a species have slowly surrendered ourselves to computing. Looking
back, we can identify several major and distinct styles whereby
we have built such systems. We have come a long way, and even
today, we certainly can name a number of best practices for
software development that yield systems of quality. However, by
no means can we stand still: the nature of the systems we build
continues to change, and as they collectively weave themselves
into our live, we must attend not only to the technical elements
of software development, we must also attend to human needs. In
this presentation we will look at the history of software
engineering and offer some grand challenges for the future.

Biography: Grady Booch is Chief Scientist for Software Engineering at
IBM Research where he is currently engaged in the architecting of
cognitive systems encompassing both IBM's Watson as well as non-von
Neumann platforms. Grady is also developing a major transmedia
documentary for public broadcast on the intersection of computing and
the human experience. Booch is an IBM Fellow, an ACM Fellow, an IEEE
Fellow, and on behalf of the BCS has been awarded the Lovelace Medal
and given the Turing Lecture. Author of six books, Grady wrote the
long-running column "On Architecture" and now writes "On Computing"
for IEEE Software; over his career he has published hundreds of
technical articles and has lectured and consulted around the
world. Co-author of the Unified Modeling Language, Grady has helped
with the architecture of complex software-intensive systems in most
every domain imaginable.

**2015年度第3回 [#jbd67aaa]
日時:''8月28日11時から''

場所:''東工大大岡山キャンパス西八号館W811''

講演者:''Professor Daniel Keren (University of Haifa)''

Title: ''Classification Using a Background Prior''

Abstract: A canonical problem in machine learning is category
classification (e.g. find all instances of human faces, cars
etc., in an image). Typically, the input for training a
classifier is a relatively small sample of positive examples, and
a much larger sample of negative examples, which in current
applications can consist of images from thousands of
categories. The difficulty of the problem sharply increases with
the dimension and size of the negative example set. In this talk
I will describe an efficient and easy to apply classification
algorithm, which replaces the negative samples by a prior and
then constructs a "hybrid" classifier which separates the
positive samples from this prior. The resulting classifier
achieves an identical or better classification rate than SVM,
while requiring far smaller memory and lower computational
complexity to train and apply.

**2015年度第2回 [#s5f210db]

日時: ''2015年7月29日(水) 14:00--16:00''
会場: ''西8号館W棟9階 W911号室''

講演者: ''[[Walter Binder:http://www.inf.usi.ch/faculty/binder/]] (Università della Svizzera italiana (USI), Switzerland)''

題目: ''Accurate Profiling in the Presence of Dynamic Compilation''

要旨: Many profilers based on bytecode instrumentation yield
wrong results in the presence of an optimizing dynamic compiler,
either due to not being aware of optimizations such as stack
allocation and method inlining, or due to the inserted code
disrupting such optimizations. To avoid such perturbations, we
present a novel technique to make any profiler implemented at the
bytecode level aware of optimizations performed by the dynamic
compiler. We implement our approach in a state-of-the-art Java
virtual machine and demonstrate its significance with several
concrete profilers. We quantify the impact of escape analysis on
allocation-hotspot profiling and on object-lifetime analysis, as
well as the impact of method inlining on callsite profiling.  We
illustrate how our approach enables new kinds of profilers, such
as a profiler for non-inlined callsites and a testing framework
for locating performance bugs in dynamic compiler
implementations.  Joint work with Yudi Zheng and Lubomir Bulej.

略歴: Walter Binder is an associate professor in the Faculty of
Informatics, Università della Svizzera italiana (USI),
Switzerland. He holds an MSc, a PhD, and a Venia Docendi from
Vienna University of Technology, Austria.  His main research
interests are in the areas of program analysis, virtual machines,
parallel programming, and Cloud computing.

**2015年度第1回 [#p3fa19d5]

日時: ''2015年6月2日(火) 17:00-18:30''

会場: ''西8号館W棟9階 W911号室''

講演者: ''[[Raffi Khatchadourian:https://openlab.citytech.cuny.edu/khatchad/]] (City University of New York)'' 

題目: ''Open Problems in Automatically Refactoring Legacy Java Software to use New Features in Java 8''

要旨: Java 8 is one of the largest upgrades to the popular language and
framework in over a decade. In this talk, I will first overview
several new, key features of Java 8 that can help make programs
easier to read, write, and maintain, especially in regards to
collections. These features include Lambda Expressions, the
Stream API, and enhanced interfaces, many of which help bridge
the gap between functional and imperative programming paradigms
and allow for succinct concurrency implementations. Next, I will
discuss several open issues related to automatically migrating
(refactoring) legacy Java software to use such features
correctly, efficiently, and as completely as possible. Solving
these problems will help developers to maximally understand and
adopt these new features thus improving their software.

**2014年度第2回 [#gdaa1d28]

日時: ''2014年8月13日(水) 13:30-14:30''

会場: ''西8号館W棟8階W809号室''

講演者: ''岩田具治 氏(NTTコミュニケーション科学基礎研究所)''
講演者: ''Victor Magron (Laboratory for Analysis and Architecture of Systems (LAAS)-CNRS)''

題目: ''確率モデルに基づくデータマイニング''
題目: ''New applications of moment-SOS hierarchies''

要旨:
NTT コミュニケーション科学基礎研究所にて行ってきた確率モデルに基づくデータマイニング研究(定額制サービスのための推薦,クラス構造の可視化,ソーシャルアノテーションの解析)を紹介する.確率モデルを用いることにより,データに内在する不確実性に対応でき,また,確率論の枠組みに従い,多様な情報を統一的に扱うことができる.
要旨: Semidefinite programming is relevant to a wide range of mathematic fields, including combinatorial optimization, control theory, matrix completion. In 2001, Lasserre introduced a hierarchy of semidefinite relaxations for particular polynomial instances of the Generalized Moment Problem (GMP). My talk emphasizes new applications of this moment-SOS hierarchy, investigated during my PhD and Postdoc research.

定額制サービスのための推薦:定額制サービスを提供しているオンラインストアが収益を上げるためには,ユーザの契約期間をできるだけ延ばすことが必要である.従来推薦法では,購入される確率を最大化するためにユーザの嗜好に合致する商品を提示する.しかしながら,従来法により必ずしも契約期間が延びるとは限らない.そこで,定額制サービスを想定し,契約期間が延びる確率を最大にする推薦法を提案する.提案法では,まず,生存時間解析を応用し,契約期間の長いユーザに特徴的な購買パターンを抽出する.そして,抽出されたパターンと同じような購買行動になるように商品を推薦する.また,効果的な推薦にするため,最大エントロピーモデルを用いてユーザの嗜好を推定する.携帯電話用漫画配信サイトのログを用いた実験を紹介する.
In the context of formal proofs for nonlinear optimization, one can combine the moment-SOS hierarchy with maxplus approximation of semiconvex functions. Such a framework is mandatory for formal certification of nonlinear inequalities, occurring by thousands in the proof of Kepler Conjecture by Hales.

クラス構造の可視化:データをそのクラス構造とともに低次元の可視化空間へ埋め込む非線形次元圧縮手法を紹介する.提案法は,可視化空間において混合正規分布を仮定し,原空間における事後確率と可視化空間における事後確率のKullback-Leibler ダイバージェンスができるだけ小さくなるようにデータとクラスを埋め込む.データ間距離(非類似度) を直接計算しないため,従来法に比べ計算効率がよい.分類済Web ページを用いたクラスラベル付きデータの可視化,手書き文字を用いた分類モデルの可視化,単語群を用いたラベルなしデータの可視化を行い,提案法の有効性を示す.
I also present how to approximate, as closely as desired, the Pareto curve associated with bicriteria polynomial optimization problems or the image of semialgebraic sets under polynomial maps. For each problem, one builds a hierarchy of semidefinite programs, so that the sequence of bounds converges in L1 norm.

ソーシャルアノテーションの解析:ソーシャルアノテーションサービスでは,ユーザが任意のタグを付与できるため,しばしば内容に関連しないタグが含まれる.内容に関連するタグを自動的に抽出することにより,情報検索の性能向上や,文書分類や画像認識などの機械学習タスクの精度向上が期待できる.そこで,潜在トピックモデルに基づく内容に関連するタグの抽出法を提案する.提案法では,内容とタグが生成される過程をモデル化し,確率的EMアルゴリズムを用いてモデルを推定することにより,関連するタグを自動的に抽出する.人工データ,および,Web ページと画像を対象とするソーシャルアノテーションサービスのデータを用いて提案法の有効性を示す.
Finally, this hierarchy allows to analyze programs containing loop invariants with polynomial assignments.


**2011年度第4回 [#sc26d268]

日時: ''2011年5月27日(金)16:30-17:45''

**2014年度第1回 [#gdaa1d28]

日時: ''2014年7月1日(火) 10:45-12:15''

会場: ''西8号館W棟11階W1101号室''

講演者: ''Katerina Mitrokotsa (Department of Computer Science and Engineering, Chalmers University of Technology)''

題目: ''Authentication in constrained settings''

要旨: Access to restricted services and/or places requires authentication. However, authentication is sometimes performed in: i) noisy conditions, ii) hostile environments and iii) constrained settings. By noisy conditions, we refer to noise in the communication channel that may lead to modification of the  transmitted information. By hostile environments we mainly refer to  environments where attackers may attempt to impersonate legitimate users,
while by constrained settings we refer to environments that may include communication among wireless devices with limited resources.

Authentication is a decision making problem where we need to decide whether or not to accept the credentials of an identity-carrying entity. In the context of cryptographic authentication, we have extensively investigated the family of distance bounding protocols. These can be used as the main countermeasure against relay attacks. We analyse the security of such protocols. These authentication problems will also be briefly connected to the problem of privacy-preservation.

(講演者紹介: Katerina Mitrokotsa is an assistant professor at the department of
Computer Science and Engineering at Chalmers University of Technology. Her main research interests lie in information and network security, privacy- preservation, machine learning for security and applied cryptography. Formerly, she held positions as a Marie Curie fellow at EPFL, as professor at the University of Applied Sciences of Western Switzerland (HES-SO), as a postdoctoral researcher in TU Delft and as a visitor assistant professor at the department of Computer Science at Vrije Universiteit in Amsterdam. She has been awarded the Rubicon Research Grant by the Netherlands Organization for Scientific Research (NWO) and a Marie Curie Intra European Fellowship. Currently, among others she serves as associate editor for the IEEE Communications Letters and the Computers & Security Journal (Elsevier). She has served on the PCs of INFOCOM, ACNS, Africacrypt, Indocrypt and multiple other well-known conferences in the area on communications and information security.)




**2013年度第9回 [#gdaa1d28]

日時: ''2014年3月10日 10:00 - 12:00''

会場: ''西8号館W棟10階W1008号室''

講演者: ''Neil Immerman 氏 (University of Massachusetts Amherst)''
講演者: ''方菱 氏 (産業技術総合研究所セキュアシステム研究部門システムライフサイクル研究グループ''

題目: ''P versus NP: Approaches, Rebuttals, and Does It Matter?''
題目: ''確率モデル検査による止まらないマイコンFUJIMIの評価''

要旨:
NP contains all those problems that we could efficiently solve if
"guessing were free", i.e., all those problems whose solutions we
could recognize as correct if they were given to us in an
appropriate format and context.  We've known since 1971 that a large
class of important combinatorial problems including boolean
satisfiablity (SAT) are NP complete, i.e, hardest problems in NP.
If any of these had a solution in P, then so would all other
problems in NP.
マイクロプロセッサユニット(MPU)は、静電気放電等外部ノイズ信号を
受けた場合、システム機能がフリーズや誤動作する場合が多々ある。FUJIMIはレ
ジリエンス(resilience)技術である。この技術は定期的にCPUコアをリセット
し、フリーズや誤動作から回復させる。しかし、一般にレジリエンス技術は複雑
な相互作用因子に依存するため、評価が困難である。

How hard is SAT, or any other NP-complete problem?  It is well
believed that P is not equal to NP and that SAT requires exponential
time on some hard instances.  However, SAT solvers are now in
practical use as general problem solvers, currently working even
on instances with a million variables.
本研究は、上記のFUJIMIシステムを確率モデルによって定性と定量的な評価を
行った。定性評価は6 PCTL式により、システムを正しく回復できる性質を証明し
た。定量評価は、システム障害の低減率とADT(業界標準である年平均ダウンタ
イム)によってシステムを改善したことを具体的な数値によって示した。実験結
果によって、我々の評価方法は費用対効果が高く、信頼性があることが示された。

I will explain some of these issues, laying out what we do know,
what we do not know, and what we hope to understand once the P
versus NP question, together with many similar but less famous open
questions, are finally resolved.  At the same time, this talk will
present a survey of descriptive complexity -- the logical approach
to compexity that I have been pursuing for years.

連絡先:渡辺治 watanabe(at)is.titech.ac.jp

**2011年度第3回 [#sc26d268]

日時: ''2011年5月18日(水)13:30-14:30''

会場: ''西8号館W棟10階W1008号室''

講演者: ''梅原雅顕 氏 (東京工業大学 数理・計算科学専攻)''
**2013年度第8回 [#gdaa1d28]

題目: ''変曲点を持たない平面閉曲線について''
日時: ''2014年1月15日 15:05 - 16:35''

会場: ''西8号館W棟8階W809号室''

講演者: ''三浦英之 氏 (東京工業大学 数理・計算科学専攻)''

題目: ''非圧縮Navier-Stokes流に対する特異点除去可能性について''

要旨:
非圧縮Navier-Stokes方程式は非圧縮粘性流体の流速場と圧力場の関係を
記述すると信じられている偏微分方程式である.その解の時間大域的な滑らかさ
の問題は, 1934年にJ.Lerayによって提案されて以来,多くの研究が行われてい
るが,現在までのところ未解決である.この問題に対する一つのアプローチとし
て,ある点で解が特異にならないための十分条件を考察する特異点除去可能性の
研究がある.この研究の最近の進展を,講演者の結果も含めて紹介したい.

平面上の正則閉曲線は,交点数が有限で高々2重点しかもたないとき,「一般的」であるという.一般的な平面正則閉曲線が与えられたとき,その閉曲線の変曲点の個数を下から評価する不変量を構成する.この不変量は,与えられた閉曲線の組み合わせ論的な情報から有限回の手段で計算可能なもので,与えられた交点数の閉曲線の中で,変曲点を持たないものを特徴づけるのにきわめて有効である.実際,この不変量を用いて,5交点以下の変曲点をもたない一般的な正則閉曲線の分類を与える.また,変曲点と2重接線との関係など,現段階で未解決の問題をいくつか紹介したい.今回話す内容は,名城大学の小沢哲也氏と,一昨年前に私が指導した現在高校教師の大野君との共同研究です.

**2011年度第2回 [#sc26d268]

日時: ''2011年5月11日(水)16:30-17:45''

会場: ''西8号館W棟10階W1008号室''
**2013年度第7回 [#gdaa1d28]

講演者: ''Valentine Kabanets 氏(Simon Fraser Univ., Computing Science)''
日時: ''2013年12月16日(月)13:20 - 14:30 その後,簡単なお茶会をします''

題目: ''Direct Product Codes: Decoding and Testing''
会場: ''西8号館W棟10F1008コラボレーションルーム''

講演者: ''Ravi Montenegro (U. Mass, Lowell)''

題目: ''The Kruskal Count: how a card trick can help us break codes''

要旨:
The Kruskal Count is a card trick popularized by Martin
Gardner. John Pollard independently used the same idea to construct a
tool that can be used to break some types of encryption. The Birthday
Paradox can explain why your ipod seems to play too many songs from the
same albums. Pollard again used similar intuition on cryptographic
problems. We discuss these curious results and explain how they are
examples of random walks, also known as Markov Chains.

※ Kabanets 先生は数理計算科学専攻の客員教授として5月〜8月まで滞在され,計算の複雑さの理論(平均時計算量)の講義を行います.(数理計算科学特別講義第三,5月16日より)

本講演はそれに先駆けて行う関連分野の(講義より少し高度な)最近の研究トピックスに関するお話です.ふるってご参加ください.
**2013年度第6回 [#gdaa1d28]

Abstract.
Given a Boolean function f, and parameter k (natural number), we define f^k to be a function mapping k-tuples of inputs x_1,...,x_k to the k-tuple of values f(x_1),...,f(x_k). If we think of the truth table of f as a message, then the truth table of f^k can be viewed as an encoding of f using the Direct-Product Code. This encoding is useful in complexity and cryptography for hardness amplification: if f is somewhat hard on average to compute by certain efficient algorithms, then f^k is much harder on average to compute by a related class of efficient algorithms.
日時: ''2013年12月4日(水) 16:30〜''

Over the past several years, together with Impagliazzo, Jaiswal, and Wigderson, we gave tight results about both decodability (actually, list-decodability) and testability of such Direct Product codes. Here
decodability is the task: given a corrupted version of f^k, recover (a small list of candidates for) a message close to the original  message f. The testability is the task: given an oracle F, decide if F is
close to being (approximately equal to) f^k, i.e., if F is a Direct-Product codeword. Somewhat surprisingly, the testability result uses the ideas developed for the decodability result, even though
there is no formal connection between these two tasks. Finally, the testability result allows us to prove a version of Parallel Repetition Theorem for certain 2-player games, basically showing that the ability
to win in k copies of a game decreases exponentially with k.
会場: ''大岡山西8号館 W棟 W1008室''

講演者: ''石崎 一明 (IBM東京基礎研究所)''

**2011年度第1回 [#sc26d268]
題目: ''静的型付き言語用Just-In-Timeコンパイラの再利用による、動的型付き言語用コンパイラの実装と最適化''

日時: ''2011年4月27日(水)14:30-15:30''
要旨:
Web関連のソフトウェア開発において、JavaScript, PHP, Python,Rubyなどの動的型付き言語の人気が高い。これらの言語では、最初インタプリタによって実装され、その後言語の人気の高さとともに大きくなる高速化の要求に応えるためにコンパイラが開発される、という過程をたどることが多い。新たな言語のコンパイラ開発のために、言語毎にコンパイラを一から作成するのは実装にかかる労力が非常に大きい。我々は、この労力を削減し効率的な開発のために、既存の静的型付き言語Just-In-Timeコンパイラの再利用Repurpose-JIT compiler)というアプローチを用いて、Python言語のコンパイラを実装した。Repurpose-JIT compilerの実装においては、以下の2点が問題となり、高性能を得ることが難しかった。1)コンパイル対象とする言語の命令と再利用しようとするコンパイラの中間言語の命令の粒度の違い、2)高品質なコード生成にとって重要な型情報をソースコードから得ることの困難さ。本発表では、これらの問題点を解決する最適化を提案し、特に実行時情報を用いたコードの特殊化が有効であることを実験結果に基づいて紹介する。さらに、今回の経験から得られた、コンパイラ以外の性能に影響を与える要因についても紹介する。
This is joint work with Jose Castanos, David Edelsohn, Priya
Nagpurkar, Toshio Nakatani, Takeshi Ogasawara, and Peng Wu.

参考文献:

[1] Kazuaki Ishizaki, Takeshi Ogasawara, Jose Castanos, Priya
Nagpurkar, David Edelsohn, and Toshio Nakatani. Adding
dynamically-typed language support to a statically-typed language
compiler: performance evaluation, analysis, and tradeoffs, VEE 2012,
2012

[2] Jose Castanos, David Edelsohn, Kazuaki Ishizaki, Priya Nagpurkar,
Toshio Nakatani, Takeshi Ogasawara, and Peng Wu. On the Benefits and
Pitfalls of Extending a Statically Typed Language JIT Compiler for
Dynamic Scripting Languages, OOPSLA 2012, 2012

発表スライドは次のリンクからご覧になれます.
http://www.slideshare.net/ishizaki/titech20131204-public



**2013年度第5回 [#gdaa1d28]

日時: ''2013年11月7日(水) 13:20-14:50''

会場: ''大岡山西8号館W棟 W809号室''

講演者: ''鈴木大慈 氏(東京工業大学 数理・計算科学専攻)''

題目: ''マルチプルカーネル学習およびガウス過程事前分布を用いたスパース加法モデル推定''

要旨:
Multiple Kernel Learning (MKL) およびガウス過程事前分布を用いたMKLのベイズ的変種についてその統計的性質を議論する.MKLは画像認識の分野などで成功を収めているが,オリジナルのL1正則化を用いるよりもL1とL2の中間的な正則化を用いたものが多くの問題で高い精度を出すことが実験的に報告されている.そこで,L1とL2の中間的な正則化としてelasticnet型の正則化やLpノルムを考えた場合に,それら中間的な正則化がどのように汎化性能に影響を与えるかを理論的に考察する.一方,MKLのベイズ的変種を用いれば,デザインへの条件を課さないでも速い収束レートを導出することができることも紹介する.

参考文献:

[1] Taiji Suzuki, and Masashi Sugiyama: Fast learning rate of multiple
kernel learning: trade-off between sparsity and smoothness.
The Annals of Statistics, vol. 41, number 3, pp. 1381-1405, 2013.

[2] Taiji Suzuki: PAC-Bayesian Bound for Gaussian Process Regression and
Multiple Kernel Additive Model.
Conference on Learning Theory (COLT2012), 2012, JMLR Workshop and Conference
Proceedings 23: 8.1--8.20, 2012.

[3] Taiji Suzuki: Unifying Framework for Fast Learning Rate of Non-Sparse
Multiple Kernel Learning.
Advances in Neural Information Processing Systems 24 (NIPS2011).
pp.1575--1583.





**2013年度第4回 [#gdaa1d28]

日時: ''2013年7月24日(水) 13:20-14:50''

会場: ''大岡山西8号館W棟 W1008号室''

講演者: ''寺嶋 郁二 氏(東京工業大学 数理・計算科学専攻)''

題目: ''クラスター変換とブレイド群''

要旨:
クラスター変換の理論は,2002年に二人の数学者 S.Fomin と A.Zelevinsky によって導入されました.この講演ではクラスター変換を用いて,ブレイド群の表現を具体的かつ幾何的に構成できることを説明します.





**2013年度第3回 [#gdaa1d28]

日時: ''2013年7月18日(木) 13:20-14:50''

会場: ''大岡山西8号館W棟8F W809号室''

講演者: '' 野村 俊一 氏 (東京工業大学 数理・計算科学専攻)''

題目: ''アクチュアリーの業務、試験、課題について''

要旨:
「アクチュアリー」とは,保険・年金の業界で活躍する数理業務のスペシャリストである.
本発表では、自身の経験を踏まえてアクチュアリーに関する以下の内容を紹介する.
・アクチュアリーの役割と前職(損害保険会社)での経験
・アクチュアリー資格試験を通るまでの軌跡と勉強方法
・アクチュアリーが現在取り組んでいる主要な課題




**2013年度第2回 [#gdaa1d28]

日時: ''2013年5月29日(水) 15:30-''

会場: ''大岡山西8号館W棟10F コラボレーションルーム(W1008)''

講演者: '' Ian Piumarta氏 (Viewpoints Research Institute)''

題目: ''To trap a better mouse''

要旨:
Edsger Dijkstra once said: "The tools we use have a profound and
devious influence on our thinking habits, and therefore on our thinking
abilities." When we write programs, the single most important tool we're
using is the programming language. When we become fluent in a
programming language, we tend to think of our solutions directly in
terms of its data structures and algorithms. Unfortunately there is no
such thing as a programming language that is good for everything. In our
lives as programmers we will inevitably encounter problems for which a
natural and elegant language has not yet been invented in which to
express a solution. We then have a choice: either think poor thoughts in
a 'general purpose' programming language, or invent new languages in
which to think much better thoughts. Which option we choose can mean the
difference between a successful project and a failed project. In this
talk I will review some of the reasons why language is important, show
which parts of both natural and artificial languages tend to be
malleable, and suggest some ways of thinking about languages that might
lead to better ways of modelling and implementing the languages in which
we program -- and, therefore, think.







**2013年度第1回 [#gdaa1d28]

日時: ''2013年4月18日(木) 13:20-14:50''

会場: ''西8号館W棟8階W809号室''

講演者: '' Anne Bouillard 氏 (ENS Paris, 東京工業大学 数理・計算科学専攻)''

題目: ''Exact Worst-case Delay for FIFO-multiplexing Tandems''

要旨:
This talk is about computing the actual worst-case end-to end
delay for a flow in a tandem of network under some general assumptions
using the network calculus theory, that is originally based on the
(min,plus) algebra. Unfortunately, using those algebraic methods lead to
very pessimistic bounds. We will see that we can indeed model this
network as a mixed interger linear program and drastically improve those
bounds.





**2012年度第4回 [#j4f0fd9f]

日時: ''2013年3月21日(木) 15:00-16:30''

会場: ''西8号館E棟10階E1011号室''

講演者: '' Martin Mu"ller 氏''

題目: ''Move Quality in Monte Carlo Simulation:
A Case Study using the Fuego Go Program''

要旨:
Despite half a dozen years of intense research, designing effective
simulation policies for Monte Carlo Tree search in Go is still
considered something of a black art, and driven largely by trial and
error. Important ideas that have evolved include pattern- and
tactics-based playouts, simulation balancing, and several schemes to
dynamically modify simulation policies online. In this study, we take
an in-depth look at what happens when the Go program Fuego runs its
playouts. We develop several methods to evaluate the quality of moves
played in these simulations, and we evaluate the contribution of the
different components of Fuego's playout policy. We study the
distribution of both the number of blunders, or result-changing moves,
and the absolute loss - in terms of number of points - for many
variations of the Fuego playout policy. We use this study to identify
an improvement to the Fuego default policy.

(Joint work with Sumudu Fernando)

**2012年度第3回 [#j4f0fd9f]

日時: ''2013年3月8日(金) 16:00-17:00''

会場: ''西8号館W棟10階1008号室''

講演者: ''Benjamin Burton 氏(The University of Queensland)''

題目: ''Untangling knots using combinatorial optimisation''

要旨:
Unknot recognition is the algorithmic problem of determining whether a
knot (i.e., a closed loop) in 3-dimensional space can be untangled.
It is a major unsolved problem as to whether this problem has a fast
polynomial time solution.  Here we present the first conclusive
algorithm for unknot recognition which, although exponential time in
theory, exhibits a clear polynomial time behaviour in exhaustive
practical experiments.  The algorithm makes significant use of
techniques from combinatorial optimisation, and uses a
branch-and-bound framework with linear programming steps.  We also
introduce the topological software package Regina (in which the
algorithm was implemented), and discuss the relevance of our results
to other difficult problems in computational topology.

This is joint work with Melih Ozlen.

講演後には Burton 氏との懇親会を予定しています。奮ってご参加ください。

**2012年度第2回 [#j4f0fd9f]

日時: ''2012年11月12日(月) 15:00-16:30''

会場: ''西8号館E棟10階1011号室''

講演者: ''S.R.K. Branavan 氏''

題目: ''Grounding Linguistic Analysis in Control Applications''

要旨:
Natural languages are the medium in which the majority of humanity's collective knowledge is recorded and communicated.  If machines were able to automatically access and leverage this information, they could effectively perform many tasks that are currently considered to be computationally intractable, thus requiring human involvement.  Today, the only way to infuse human knowledge into computational algorithms is to have humans in the loop - i.e., to manually encode the knowledge into heuristics, through annotations, or directly into the model structure itself.

Our ultimate goal is to automate this process, so that machines can access required knowledge directly from text.  One path to this goal is to perform a semantic interpretation of text by grounding the textual information in the objects, actions and dynamics of the physical world.  From a linguistic viewpoint, this grounding of language in control applications presents a very natural notion of language semantics.  I.e., by allowing us to define semantics with respect to the control application, and avoid imposing subjective human notions of correctness.

In this talk, I will explore two aspects of the connection between language and control applications: first, how the semantic analysis of language can be driven by control performance; and second, how information from text can be leveraged to improve performance in complex control systems.  These two aspects are in fact complementary, and language analysis is central to them both.  Addressing these aspects jointly is particularly challenging.  However, as our empirical results show, connecting the semantic analysis of language to control applications allow us to leverage the synergy between the two to simultaneously learn both language analysis and application control with little or no prior knowledge.

**2012年度第1回 [#j4f0fd9f]

日時: ''2012年4月19日(木) 13:20-14:50''

会場: ''西8号館W棟8階W809号室''

講演者: ''一杉裕志 氏 (産業総合研究所産 ヒューマンライフテクノロジー研究部門)''
講演者: ''Tuan Phung-Duc 氏 (東京工業大学 数理・計算科学専攻)''

題目: ''脳とベイジアンネット''
題目: ''Markov Chain and Queue with State-dependent Transition Rates''

要旨:
In this talk, we consider continuous time Markov chains with
state-dependent transition rates arising from either retrial queues or
infinite server queues. First, for multiserver retrial queues, we show
that in the case of three and four servers, the joint stationary
distribution of the number of busy servers and the number of retrying
customers can be expressed in terms of continued fractions.
Furthermore, we formulate the general case with any number of servers
by a level-dependent QBD process for which a memory saving algorithm
based on matrix continued fractions is presented. Next, we briefly
present a topic on infinite server queues with synchronized
abandonment, where we develop a new methodology to obtain either
closed form solution or recursion for the stationary distribution of
the number of busy servers.

大脳皮質は脳の中で知能にもっとも深く関係する重要な組織である。計算論的神経科学という分野において、大脳皮質の本質的な機構がベイジアンネットであると理解する研究者が少しづつ増えている。ベイジアンネットに基づくモデルは大脳皮質の複雑で多様な振る舞いを、少ない仮定で計算論的にきれいに説明する。それだけではなく、脳がどのようなアルゴリズムやデータ構造を用い、それらをどのような神経回路で実現しているかについても、ベイジアンネットを鍵として徐々に具体的に理解されつつある。大脳皮質の情報処理原理が明らかになることで、人間のような高い知能の計算機による再現も、もはや夢ではなくなりつつある。

**過去の談話会 [#db336cc6]

[[2011年度専攻談話会]]

[[2010年度専攻談話会]]

[[2009年度専攻談話会]]

[[2008年度専攻談話会]]

[[2007年度専攻談話会]]

[[2006年度専攻談話会]]

[[2005年度専攻談話会]]

[[2004年度専攻談話会]]

[[2003年度専攻談話会]]

[[2002年度専攻談話会]]