Google Tech Talk
August 22, 2012

Presented by Lee Worden

ABSTRACT

I will present a modeling approach to theorizing about processes and protocols for collective deliberation among humans — that is, ways for a group efficiently and effectively to find a mutually agreeable solution to a problem when participants may disagree on what is acceptable. Seen from this perspective, the problem becomes one of collective search in a complex problem space. I will draw connections to work in evolutionary dynamics, political science, and distributed problem solving in computers. I will be very interested in hearing feedback from Googlers about the search and distributed information processing aspects of the project.

SPEAKER INFO

Lee Worden is an applied mathematician specializing in dynamics of collective behavior, coexistence, communities, and adaptation. He has studied transitions from competitive to mutualistic interactions in ecological evolution and game theory, and emergence of self-regulation in model planetary ecosystems. His current research includes threshold models for abrupt social change, as well as developing a modeling approach to collective deliberation. He is currently working at UC Berkeley and (virtually) at McMaster University, and teaching at the SF Art Institute.

Read more about this video…

Lecture by Sasha Rush for the Programming Paradigms (CS107) course in the Stanford University Computer Science department. Sasha lectures on the Haskell programming language, providing several code examples while comparing it to other common programming languages.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain lectures on the Python programming language in relation to XML processing and the internet.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses the Python dictionary and illustrates a small program containing the imperative, object-oriented, and functional paradigms.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain introduces Python, a scripting language, by providing a basic overview of the language and some examples.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses memory models in Scheme programming language by giving several detailed examples.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain elaborates upon Scheme programming language by giving several detailed code examples.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain continues discussing the functional program and the Scheme programming language by focusing upon function pointers.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. Prof. Cain continues lecturing on the functional paradigm and the Scheme programming language by using the flatten function as an example.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain introduces a new programming paradigm, the functional paradigm, and begins lecturing on the Scheme programming language.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses the synchronization directive by using the ice cream store simulation as an example.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain continues discussing semaphores, and moves on to more practical applications of threading in relation to C and C++ programming.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain continues his lecture on thread libraries, discussing semaphores in greater depth, in the context of the C and C++ programming languages.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses thread libraries in the context of the C and C++ programming languages.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses struct codes and wraps up the implementation of C and C++ code.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses how linking and compilations work together in the context of C++ and C language programming.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses pre-processing, compilations, and linking in the context of C++ and C language programming.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain continues his lecture on computer architecture by discussing function call and return in further depth as well as introducing the swap function in relation to C++ code.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain continues his discussion on computer architecture by talking about function call and returns in relation to C code.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain introduces computer architecture and the Assembly programming language while referencing C code.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses heap segments and their use in the C programming language.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain continues his lecture on the C programming language and generic stacks.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses C language programming by focusing on different forms of stack.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain focuses on linear search and stack within the C programming language.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses C and C++ programming, including bit patterns, memory copy, and linear search.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses C programming, focusing upon string duplicates, string copy, and memory diagrams.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. In this lecture, Prof. Cain discusses C and C++ programming codes, as well as binary addition and subtraction.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Lecture by Professor Jerry Cain for Programming Paradigms (CS107) in the Stanford University Computer Science department. Professor Cain provides an overview of the course.

Programming Paradigms (CS107) introduces several programming languages, including C, Assembly, C++, Concurrent Programming, Scheme, and Python. The class aims to teach students how to write code for each of these individual languages and to understand the programming paradigms behind these languages.

CS 107 Course Website:
http://www.CS107.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Read more about this video…

Introduction
Project mission statement, history, internal organization, partners, CGAL in numbers.

What's in CGAL
A survey on available data structures and algorithms, as well as examples how and by whom they are used. Topics include Triangulations, Voronoi diagrams, Boolean operations on polygons and polyhedra, arrangements of curves and their applications, Mesh generation, Geometry processing, Alpha shapes, Convex hull algorithms, Operations on polygons, Search structures, Interpolation, Shape analysis, fitting, and distances, Kinetic data structures...

Generic Programming Paradigm
CGAL data structures are C++ template classes and functions, usually taking several template parameters (with default values for ease of use). This gives developers an incredible flexibility to adapt the data structures to their needs, which is important internally for code reuse, and important for end users, as they typically integrate CGAL in already existing applications. Parts of CGAL are also interfaced with languages and software like Python, Java, Scilab, Qt and the Ipe drawing editor.

Exact Geometric Computing Paradigm
We present how to make geometric algorithms correct, robust, and nevertheless fast, by combining floating point arithmetic with exact arithmetic, and clever filtering mechanisms to switch between these two modes. These mechanisms can be used for geometric predicates, as well as for geometric constructions, which instead of a discrete return value generate new geometric entities.

Conclusion and Outlook
A wrapup, and a sneak preview on algorithms that might make it into future releases of CGAL.

Speaker: Andreas Fabri, PhD, GeometryFactory
As member of the initial development team of the CGAL project, Andreas is one of the architects of the CGAL software. For several years he chaired the CGAL Editorial Board. In 2003, Andreas founded the GeometryFactory as spin-off of the CGAL project, offering licenses, service and support to commercial users. Andreas received his PhD in 1994 from the Ecole des Mines de Paris, while working on geometric algorithms for parallel machines at INRIA.

Speaker: Sylvain Pion, PhD, INRIA Sophia-Antipolis
Sylvain got involved in the CGAL project during his PhD, which he received in 1999 at INRIA. He worked then on providing generic solutions to numerical robustness issues arising in geometric algorithms. Later on he worked on the efficiency of some fundamental geometric algorithms such as 3D Delaunay triangulations. He is now also involved in C++ standardization, and is working on parallel geometric algorithms. He is employed as researcher at INRIA, and is the current chair of the CGAL Editorial Board.

Google Tech Talks
March, 3 2008

Read more about this video…

Lecture slides:

• Mining Large Graphs
• Networks –Social and Technological
• Examples of Networks
• Networks of the Real-world
• Mining Social Network Data
• Networks as Phenomena
• Models and Laws of Networks
• Networks: Rich Data
• Networks: A Matter of Scale
• Networks: Scale Matters
• Structure vs. Process
• Structure of Networks
• Diffusion in Networks
• Tutorial outline
• Mining Large GraphsPart 1: Structure and models of networks
• Part 1: Outline
• Part 1.1: Structural properties
• Traditional approach
• Motivation: New approach
• Graphs and networks
• Small-world effect
• Degree distributions
• Poisson vs. Scale-free network
• Network resilience
• Community structure
• Spectral properties
• What about evolving graphs?
• Networks over time: Densification
• Densification & degree distribution
• Shrinking diameters
• Properties hold in many graphs
• Part 1.2: Models
• 1.2 Models: Outline
• (Erdos-Renyi) Random graph
• Properties of random graphs
• Evolution of a random graph
• Subgraphs in random graphs
• Random graphs: conclusion
• Exponential random graphs (p* models)
• Exponential random graphs
• Small-world model
• Preferential attachment
• Edge copying model
• Community guided attachment
• Forest Fire Model
• Forest Fire: Phase transitions
• Kronecker graphs
• Idea: Recursive graph generation
• Kronecker product: Graph
• Kronecker product: Definition
• Kronecker graphs
• Stochastic Kronecker graphs
• Kronecker graphs: Intuition
• Properties of Kronecker graphs
• 1.3: Fitting the models to real graphs
• The problem
• Model estimation: approach
• Fitting Kronecker graphs
• Challenges
• Challenge 1: Node correspondence
• Challenge 2: calculating P(G|Θ,σ)
• Model estimation: solution
• Solution 1: Node correspondence
• Sampling node correspondences
• Solution 2: Calculating P(G|Θ,σ)
• Experiments: Synthetic data
• Convergence of properties
• Experiments: real networks
• AS graph (N=6500, E=26500)
• AS: comparing graph properties
• Epinions graph (N=76k, E=510k)
• Scalability
• Conclusion
• Why should we care?
• Reflections
• References
• Coming up next…

Author: Jure Leskovec, Condensed Matter Physics, Jožef Stefan Institute

Read more about this video…

We consider online learning and its relationship to game theory. In an online decision-making problem, as in Singer's lecture, one typically makes a sequence of decisions and receives feedback immediately after making each decision. As far back as the 1950's, game theorists gave algorithms for these problems with strong regret guarantees. Without making statistical assumptions, these algorithms were guaranteed to perform nearly as well as the best single decision, where the best is chosen with the benefit of hindsight. We discuss applications of these algorithms to complex learning problems where one receives very little feedback. Examples include online routing, online portfolio selection, online advertizing, and online data structures. We also discuss applications to learning Nash equilibria in zero-sum games and learning correlated equilibria in general two-player games.

Author: Adam Kalai, Toyota Technological Institute

Read more about this video…

We study the problem of simultaneously estimating several densities where the datasets are organized into overlapping groups, such as a hierarchy. For this problem, we propose a maximum entropy formulation, which systematically incorporates the groups and allows us to share the strength of prediction across similar datasets. We derive general performance guarantees, and show how some previous approaches, such as hierarchical shrinkage and hierarchical priors, can be derived as special cases. We demonstrate the proposed technique on synthetic data and in a realworld application to modeling the geographic distributions of species hierarchically grouped in a taxonomy. Specifically, we model the geographic distributions of species in the Australian wet tropics and Northeast New South Wales. In these regions, small numbers of samples per species significantly hinder effective prediction. Substantial benefits are obtained by combining information across taxonomic groups.

Author: Miroslav Dudík, Princeton University

Read more about this video…

Pajek is a program (for Windows) for large network analysis and visualization. It is freely available for noncommercial use at http:vlado.fmf.uni-lj.si/pub/networks/pajek/ Besides ordinary networks Pajek supports also multi-relational and temporal networks. In large network analysis we are often interested in important parts of given network. There are several ways how to determine them. The islands approach is based on an importance measure of vertices or lines. Let (V,L,p) be a network with vertex property p : V ? R and let t be a real number. If we delete all vertices (and corresponding links) with the property value less than t, we get subnetwork called vertex-cut at level t. The number and sizes of its components depend on t. Often we consider only components of size at least k and not exceeding K. The components of size smaller than k are discarded as noninteresting, while the components of size larger than K are cut again at some higher level. Vertex-island is a connected subnetwork which vertices have greater property value than the vertices in its neighborhood. It is easy to see that the components of vertex-cuts are all vertex-islands. We developed an efficient algorithm that identifies all maximal vertex-islands of sizes in the interval k..K in a given network. For networks with weighted lines we can similarly define line-islands. The line-islands algorithm is based on line-cuts. Both algorithms are very general - they can be applied for any vertex/line importance measure. Their complexity is for sparse networks subquadratic - they can be applied to very large networks. We will illustrate them applying different importance measures on selected (large) networks. We will also present the use of pattern searching in analysis of genealogies and some approaches to analysis of (multi-relational) temporal networks.

Author:Nnataša Kejžar, FDV, Univerza v Ljubljani

Read more about this video…

The advances in kernel-based learning necessitate the study on solving a large-scale non-sparse positive definite linear system. To provide a deterministic approach, recent researches focus on designing fast matrixvector multiplication techniques coupled with a conjugate gradient method. Instead of using the conjugate gradient method, our paper proposes to use a domain decomposition approach in solving such a linear system. Its convergence property and speed can be understood within von Neumann's alternating pro jection framework. We will report significant and consistent improvements in convergence speed over the conjugate gradient method when the approach is applied to recent machine learning problems.

Author: Wenye Li, The Chinese University of Hong Kong

Read more about this video…

In this paper we propose a quadratic programming approach to computing the edit distance of graphs. Whereas the standard edit distance is defined with respect to a minimum-cost edit path between graphs, we introduce the notion of fuzzy edit paths between graphs and provide a quadratic programming formulation for the minimization of fuzzy edit costs. Experiments on real-world graph data demonstrate that our proposed method is able to outperform the standard edit distance method in terms of recognition accuracy on two out of three data sets.

Author: Horst Bunke, University of Bern

Read more about this video…

With the improvement of genotyping technologies and the exponentially growing number of available markers, case-control genome-wide association studies promise to be a key tool for investigation of complex diseases. However new analytical methods have to be developed to face the problems induced by this data scale-up, such as statistical multiple testing, data quality control, biological interpretation and computational tractability. We present a novel method to analyze genome-wide association studies results. The algorithm is based on a Bayesian model that integrates genotyping errors and genomic structure dependencies. Probability values are assigned to genomic regions termed bins, which are defined from a gene-biased partitioning of the genome, and the false-discovery rate is estimated. We have applied this algorithm to data coming from three genome-wide association studies of Multiple Sclerosis. The method practically overcomes the scale-up problems and permits to identify new putative regions statistically associated with the disease.

Author: Nicolas Omont, Serono, Biotech and Beyond

Read more about this video…

Understanding and identifying biological complex systems at work in the cell requires to develop models able to capture the stochastic nature of biological processes as well as their dynamics. Focusing on gene regulatory networks, we propose a new quantitative model in the form of a dynamical Bayesian network that allows to represent both genes and proteins in the same framework. We start from the nonlinear differential equations of Michaelis-Menten which are the gold-standard to represent biochemical interactions and develop a discrete-time and probabilistic model from these equations. Compared to previous works such as Nachman et al [1], our model takes into account the dependency between the regulatory proteins and the genes that code for them as well as protein-protein interactions and protein degradations. In the resulting nonlinear dynamical system, the proteins concentrations are hidden while gene expressions are observed. In order to learn the model's parameters, we first construct a discrete-time probabilistic model corresponding to our continuous-time state-space model and then derive a Kalman smoother algorithm based on the unscented transformation [2] to recursively estimate the parameters and unobserved protein activities. The generality of the learning method opens the door to various adaptations of the model if required by the biology.

Numerical results on parameter and state estimation for the repressilator [3] and other several small networks are presented and show the relevance of the model.

Author: Minh Quach, University of Evry

Read more about this video…

There are many challenges in computational modeling of biological processes. Few processes such as signaling pathways operate in- dependently of others but rather involve substantial coordination and shared resources. The level of abstraction appropriate for understand- ing different processes, e.g, viewing a pathway as a filter or a molecular cascade, varies by context and the type of predictions sought.

Author: Tommi Jaakkola, MIT - Massachusetts Institute of Technology

Read more about this video…

Systems biologists are interested in modelling chemical reactions in the intracellular environment, and to date much of what is done is based on the use of mass action kinetics to construct models of elementary reactions. Mass action kinetic models are based on a number assumptions which are not obviously valid in the intracellular environment. The cytoplasm is far from an ideal, isotropic wellmixed solution and often the concentrations of important chemical species are very small. Molecular crowding can have significant thermodynamic effects, but also must play an important dynamical role. An interesting approach that has been adopted to this has its roots in fractal geometry - a given molecule, depending upon its size and shape and the sizes and shapes of the molecules which surround it will find itself able to move in an environment of restricted dimension (see for example[1, 2]). Simple ideas have been suggested which give spatially homogeneous rate-like equations which attempt to account for this. It has been suggested, for example, that rate laws which depend on non-integer powers of the concentration of species might be used, and alternatively that the rate constants for elementary reactions which involve the encounter of different species (as opposed to spontaneous decomposition of individual molecules) should be time-dependent[1]. In this case the rates decay in time - the suggested form is the Zipf-Mandlebrot law which tends to a power law decay at long times, it is suggested that this power law characterises the dimension of the restricted environment of each chemical species[2]. Both of these approaches suffer from shortcomings. The use of non-integer powers of concentrations can only be justified in very limited circumstances, and has been shown to be inferior to the time-dependent rate parameter when describing certain lattice gas computer simulations of chemical reactions. However, the latter is clearly not invariant to time translation - the origin of time has a particular significance, and it is not clear as a general principle what the correct choice of time origin should be. Moreover, experimental techniques are being refined to the extent that spatio-temporal resolution of the species within a single cell is becoming possible. We might, therefore, aspire to constructing theories which describe the dynamics for spatially non-uniform distributions of active species. We have recently been working on a class of simple models of this type. These are spatio-temporal dynamical systems which model reaction and diffusion on a certain class of fractal sets. It has been known for some time now that it is possible to define random walks, and hence diffusion, on a certain class of fractals (indeed, it was this observation that motivated the work described above[1]). A simple example if this class is the Sierpinsky Gasket which has constrictions to the diffusion process in the sense that it can be disconnected by the removal of a finite set of points. The talk will focus mainly on this example, but we shall also suggest ways which could lead to more general models.

Supported by the Manchester Institute for Mathematical Science (MIMS).

Author: David Broomhead, University of Manchester

Read more about this video…

The Monte Carlo method was initially developed for scientific computing in statistical physics during the early days of the computers. Due to the rapid progress in computer technology and the need for handling large datasets and complex systems, the past two decades have witnessed a strong surge of interest in Monte Carlo methods from the scientific community. Researchers ranging from computational biologist to signal \& image processing engineers and to financial econometricians now view Monte Carlo techniques as essential tools for inference. Besides using the popular Markov chain Monte Carlo strategies and adaptive variants of it, various sequential Monte Carlo strategies have recently appeared on the scene, resulting in a wealth of novel and effective inferential and optimization tools. In this talk, we will present what we believe to be the "state-of-the art" in Monte-Carlo simulations for inference and will try to identify the next challenges.

Author: Eric Moulines, Enst Paris

Read more about this video…

In this paper, we propose a highly robust point-matching method (Graph Transformation Matching - GTM) relying on finding the consensus graph emerging from putative matches. Such method is a two- phased one in the sense that after finding the consensus graph it tries to complete it as much as possible. We successfully apply GTM to image registration in the context of finding mosaics from retinal images. Feature points are obtained after properly segmenting such images. In addition, we also introduce a novel topological descriptor for quantifying disease by characterizing the arterial/venular trees. Such descriptor relies on diffusion kernels on graphs. Our experiments have showed only statistical signifficance for the case of arterial trees, which is consistent with previous findings.

Author: M. Elena Martinez-Perez, National University of Mexico

Read more about this video…

Lecture slides:

• Things related to PageRank
• Covering projections in algebraic topology
• Covering projections in modern mathematics
• From covering projections to fibrations
• My own personal relation with fibrations
• A graph is a graph is a graph...
• Graph morphisms
• Graph fibration
• A graph fibration is...
• A basic ingredient: universal total graph
• Basic property of universal total graphs
• Minimum base
• Markov chains and graphs
• Markov chains with restart
• PageRank as a special case
• PageRank: an example
• Markov chains with restart are unichain
• Invariant distribution and limit behaviours
• Power series associated to a graph
• Power series and fibrations
• An example
• Consequences
• Power series associated to a graph
• Consequences
• Markovian spectrally distinguishable graphs
• Graph fibrations and graph isomorphism
• Experimental results

Author: Paolo Boldi, University of Milano

Read more about this video…

Lecture slides:

• Outline
• Nonlinear dynamical systems
• Nonlinear state-space models (NSSMs)
• Nonlinear state-space models (NSSMs)01
• Variational inference for the NSSM
• Discrete-time models: pros and cons
• Continuous-time NSSM
• Stochastic Differential Equations
• Continuous-time NSSM
• Approximations
• Variational continuous-time NSSM
• State inference
• Faster state inference
• Experiment: Continuous-time NSSM
• Experiment: Continuous-time NSSM01
• Experiment: Continuous-time NSSM02
• Experiment: Continuous-time NSSM03
• Experiment: State inference
• Experiment: State inference01
• Conclusion

Author: Antti Honkela, Helsinki University of Technology

Read more about this video…

In the field of structural pattern recognition graphs constitute a very common and powerful way of representing patterns. In contrast to string representations, graphs allow us to describe relational information in the patterns under consideration. One of the main drawbacks of graph representations is that the computation of standard graph similarity measures is exponential in the number of involved nodes. Hence, such computations are feasible for rather small graphs only. One of the most flexible error-tolerant graph similarity measures is based on graph edit distance. In this paper we propose an approach for the efficient compuation of edit distance based on bipartite graph matching by means of Munkres’ algorithm, sometimes referred to as the Hungarian algorithm. Our proposed algorithm runs in polynomial time, but provides only suboptimal edit distance results. The reason for its suboptimality is that implied edge operations are not considered during the process of finding the optimal node assignment. In experiments on semi-artificial and real data we demonstrate the speedup of our proposed method over a traditional tree search based algorithm for graph edit distance computation. Also we show that classification accuracy remains nearly unaffected.

Read more about this video…

Google Tech Talks
January, 25 2008

ABSTRACT

In this talk we examine how high performance computing has changed over the last 10-year and look toward the future in terms of trends. These changes have had and will continue to have a major impact on our software. A new generation of software libraries and algorithms are needed for the effective and reliable use of (wide area) dynamic, distributed and parallel environments. Some of the software and algorithm challenges have already been encountered, such as management of communication and memory hierarchies through a combination of compile--time and run--time techniques, but the increased scale of computation, depth of memory hierarchies, range of latencies, and increased run--time environment variability will make these problems much harder.

We will focus on the redesign of software to fit multicore architectures.

Speaker: Jack Dongarra
University of Tennessee
Oak Ridge National Laboratory
University of Manchester
Jack Dongarra received a Bachelor of Science in Mathematics from Chicago State University in 1972 and a Master of Science in Computer Science from the Illinois Institute of Technology in 1973. He received his Ph.D. in Applied Mathematics from the University of New Mexico in 1980. He worked at the Argonne National Laboratory until 1989, becoming a senior scientist. He now holds an appointment as University Distinguished Professor of Computer Science in the Electrical Engineering and Computer Science Department at the University of Tennessee, has the position of a Distinguished Research Staff member in the Computer Science and Mathematics Division at Oak Ridge National Laboratory (ORNL), Turing Fellow in the Computer Science and Mathematics Schools at the University of Manchester, and an Adjunct Professor in the Computer Science Department at Rice University.

He specializes in numerical algorithms in linear algebra, parallel computing, the use of advanced-computer architectures, programming methodology, and tools for parallel computers. His research includes the development, testing and documentation of high quality mathematical software. He has contributed to the design and implementation of the following open source software packages and systems: EISPACK, LINPACK, the BLAS, LAPACK, ScaLAPACK, Netlib, PVM, MPI, NetSolve, Top500, ATLAS, and PAPI. He has published approximately 200 articles, papers, reports and technical memoranda and he is coauthor of several books. He was awarded the IEEE Sid Fernbach Award in 2004 for his contributions in the application of high performance computers using innovative approaches. He is a Fellow of the AAAS, ACM, and the IEEE and a member of the National Academy of Engineering.

Read more about this video…

Google Tech Talks
March, 28 2008

ABSTRACT

Bhaskara M. Marthi - Research Scientist

I will describe an algorithm for probabilistic filtering, the problem of maintaining a probability distribution over the hidden state of a dynamical system given periodic noisy observations. This problem appears in various guises in practice, such as activity monitoring, state estimation, visual tracking, and fault diagnosis. Our algorithm, known as decayed MCMC, scales better than exact methods on many problems, and is less susceptible to losing track of the mode than the popular sequential Monte Carlo or particle filtering methods. Standard Markov chain Monte-Carlo mixing time analyses are insufficient to bound the complexity of our algorithm, and so we extend them to the setting of convergence of a marginal distribution.

Speaker: Dr. Bhaskara M. Marthi - Research Scientist
Dr. Bhaskara Marthi is currently a postdoctoral research associate at MIT, working with Leslie Kaelbling and Tomas Lozano Perez on hierarchical planning and robotic manipulation. He received his PhD in 2006 from the University of California, Berkeley, working with Stuart Russell on reinforcement learning with partial programs, and its application to AI design for large real-time strategy video games. His other interests include probabilistic reasoning, relational and first-order models, and Monte Carlo algorithms.

Read more about this video…

Google Tech Talks
March, 24 2008

ABSTRACT

What constitutes a scene? Defining a meaningful vocabulary for scene discovery is a challenging problem that has important consequences for object recognition. We consider scenes to depict correlated objects and present visual similarity. We introduce a max-margin factorization model that finds a low dimensional subspace with high discriminative power for correlated annotations. We postulate this space should allow us to discover a large number of scenes in unsupervised data; we show scene discrimination results on par with supervised approaches. This model also produces state of the art word prediction results including good
annotation completion.

Speaker: Ali Farhadi
Ali Farhadi is a PhD student in the UIUC Computer Science Department working on computer vision and machine learning under David Forsyth. His interests include image segmentation, transfer learning, scene understanding and human activity recognition.

Read more about this video…

Basics of probability and statistic lectures by Mikaela Keller, IDIAP Research Institute

Read more about this video…

Google Tech Talks
September 11, 2007

ABSTRACT

The name Maple is synonymous with doing complex math on computers. Best known for its symbolic or algebraic computation abilities, Maple is one of the most important tools for the modern applied mathematician and scientist. Many of you are likely familiar with Maple from college but you've probably not kept up to date with latest developments. This presentation will present some of the latest product developments from Maplesoft. Topics include

- developments in high performance numerical computation
- recent advances in symbolic computing
- new Maple libraries including graph theory, statistics, optimization, polynomial operations, and more
- parallel and grid computing
- knowledge capture for mathematical documents
- the Maple programming language and application development
- overview of new add-on products including global optimization, and modeling and simulation

The presenter will be Mohamed Bendame, a senior engineer from Maplesoft. The presentations will include an open Q&A session.

Read more about this video…

Bayesian networks are excellent tools for modeling knowledge, especially in realistic situations where there is incomplete domain awareness. Given reasonably accurate causal relationships between variables, a Bayesian network can determine the most likely state of unobserved variables in a system. Bayesian networks are used in a wide variety of applications, including textual analysis, image processing, consumer credit scoring systems, and other decision support systems.

Until recently, open source Bayesian network libraries were not readily available, and those that have been released are, in this author's opinion, unnecessarily complex. The Simple Bayesian Networking Library (SBN) and its associated Ruby module (SBN4R) make it easy to harness the power of Bayesian networks in your application. This presentation will cover what Bayesian networks are and how they are used, as well as how to use SBN4R in your next Ruby application.

Carl Youngblood has been using Ruby since the publication of the Pickaxe book in 2000. He currently works as the CTO of Construction Capital Source, a construction lending firm with headquarters in Salt Lake City. This position is especially exciting for Carl, since it allows him to use Ruby all day at work. Carl received a bachelors degree in Portuguese from Brigham Young University and a masters degree in Computer Science from the University of Washington. He has been working professionally as a software engineer for ten years.

Read more about this video…