Series

School Seminar Series: Statistically Consistent Estimation and Efficient Inference for Natural Language Parsing

Statistically Consistent Estimation and Efficient Inference for
Natural Language ParsingBy Shay Cohen, University of Edinburgh.

Abstract:
In the past few years, there has been an increased interest in the machinel earning community in spectral algorithms for estimating models with latent variables. Examples include algorithms for estimating mixture of Gaussians or for estimating the parameters of a hidden Markov model.

The EM algorithm has been the mainstay for estimation with latent variables, but because it is guaranteed to converge to a local maximum of the likelihood, it is not a consistent estimator. Spectral algorithms, on the other hand, are often shown to be consistent. They are often more computationally efficient than EM.

In this talk, I am interested in presenting two types for spectral algorithms for latent-variable PCFGs, a model widely used in the NLP community for parsing. One algorithm is for consistent estimation of L-PCFGs, and the other is for efficient inference with L-PCFGs (or PCFGs). Both algorithms are based on linear-algebraic formulation of L-PCFGs and PCFGs.

BIO:
Shay Cohen is a Chancellor’s fellow (assistant professor) at the University of Edinburgh (School of Informatics). Before that, he was a postdoctoral research scientist in the Department of Computer Science at Columbia University, and held an NSF/CRA Computing Innovation Fellowship. He received his B.Sc. and M.Sc. from Tel Aviv University in 2000 and 2004, and his Ph.D. from Carnegie Mellon University in 2011. His research interests span a range of topics in natural language processing and machine learning, with a focus on structured prediction. He is especially interested in developing efficient and scalable parsing algorithms as well as learning algorithms for probabilistic grammars.

Distinguished Lecture Series 2014: Luca Cardelli

The 2014 Distinguished Lecture Series took place on Tuesday in Lower College Hall. This year’s speaker Prof Luca Cardelli of Microsoft Research and the University of Oxford, delivered three lectures involving Morphisms of Reaction Networks that Couple Structure to Function.

Slides from the lectures are now available: http://lucacardelli.name/indexTalks.html

Luca pictured in Lower College Hall on Tuesday

Luca pictured in Lower College Hall on Tuesday

Abstract
The mechanisms underlying complex biological systems are routinely represented as networks. Network kinetics is widely studied, and so is the connection between network structure and behavior. But it is the relationships between network structures that can reveal similarity of mechanism.

We define morphisms (mappings) between reaction networks that establish structural connections between them. Some morphisms imply kinetic similarity, and yet their properties can be checked statically on the structure of the networks. In particular we can determine statically that a complex network will emulate a simpler network: it will reproduce its kinetics for all corresponding choices of reaction rates and initial conditions. We use this property to relate the kinetics of many common biological networks of different sizes, also relating them to a fundamental population algorithm. Thus, structural similarity between reaction networks can be revealed by network morphisms, elucidating mechanistic and functional aspects of complex networks in terms of simpler networks.

Tuesday’s Programme:
09:15-09:30 Introduction by Prof Simon Dobson

09:39-10:30 Lecture 1 – Molecular Programming

11:00-12:00 Lecture 2 – The Cell Cycle Switch Computes Approximate Majority

13:30-14:30 Lecture 3 – Morphisms of Chemical Reaction Networks

14:30-15:30 Q & A Session

Image courtesy of Prof Saleem Bhatti

School Seminar: Cloud Platform in Financial Services – Allan Beck, J.P. Morgan

Title: Cloud Platform in Financial Services

Presenter: Allan Beck, Cloud Platform and Strategy Lead from JPMorgan Chase

Abstract: Cloud Computing is revolutionising the delivery of compute services and driving the next generation of web-scale application design. This presents enormous opportunities but also challenges, particularly in heavily regulated sectors such as Financial Services.
Allan Beck, Cloud Platform and Strategy Lead from JPMorgan Chase, will discuss the current approach and challenges to Cloud in Financial Services. This will include an overview of available Cloud services and capabilities, the specific challenges to Cloud in Financial Services (private and public Cloud) and an overview of the next-generation Cloud platform and developer experience at JPMorgan Chase.

School Seminar: Complex Networks and Complex Processes

Simon Dobson, School of Computer Science, University of St Andrews

Abstract:

Complex networks provide a way of modelling systems with lots of
dependent elements, such as traffic networks or social networks. By
running processes over these networks we can explore how the topology of
the network affects the way the process evolves, and potentially
identify factors that accelerate or impede it. This opens-up
possibilities both for study (science) and control (engineering).

This talk will briefly introduce the mechanics of complex networks and
the processes that run on them, review some recent results we have
obtained, and look to future research programme where we will combine
simulation with sensing to give us new ways of looking at the world.

The Design and Implementation of Feldspar

By: Josef Svenningsson, Chalmers University of Technology, Sweden

Feldspar is a domain specific language with the goal of raising the
level of abstraction for performance sensitive, low-level code.
Feldspar is a functional language embedded in Haskell, which offers a
high-level style of programming. The key to generating generating
efficient code from such descriptions is to use a high-level
optimisation technique called vector fusion. Feldspar achieves
vector fusion for free by employing a particular way of embedding the
language in Haskell by combining deep and shallow embeddings.

Bio: Josef Svenningsson is an Assistant Professor in the Functional
Programming group at Chalmers University of Technology. He has a broad
range of interest and has published papers on wide variety of topics,
including: program analysis, constraint solving, security, programming
language design, testing and high-performance computing.

Talk: Internship Experiences 2014

Many St Andrews CS students do internships in the summer, but we very rarely get the opportunity to learn about them.

If you are interested in what some outstanding fourth year students did this summer, including tips and hints on how to do this yourself, you cannot miss this!

Hear them talk at 2:00pm on Tuesday.

 

Details:

Andrew McCallum worked at Inclusiq on “E-learning games for diversity”

Emily Dick worked at Accenture as a “business and system analysis to help a large government client move from a paper to an online process”

Aleksejs Sazonovs worked at Microsoft Research Cambridge (Systems and Networking group) “using insights gathered from the data, to develop an effective storage and content retrieval policy for OneDrive”

Robert Dixon worked at McLaren Racing using neural networks on a tool to help the race strategy team.

 

The speaker interns at a subsequent meal with the Head of School. From left to right, Steve Linton (HOS), Aleksejs Sazonovs, Robert Dixon, and Andrew McCallum (Emily Dick could not attend the meal).

The speaker interns at a subsequent meal with the Head of School. From left to right, Steve Linton (HOS), Aleksejs Sazonovs, Robert Dixon, and Andrew McCallum (Emily Dick could not attend the meal).

Accelerating Datacenter Services with Reconfigurable Logic

by Aaron Smith, Microsoft Research

Datacenter workloads demand high computational capabilities, flexibility, power efficiency, and low cost. It is challenging to improve all of these factors simultaneously. To advance datacenter capabilities beyond what commodity server designs can provide, we have designed and built a composable, reconfigurable fabric at Microsoft to accelerate portions of large-scale software services. In this talk I will describe a medium-scale deployment of this fabric on a bed of 1,632 servers, and discuss its efficacy in accelerating the Bing web search engine along with future plans to improve the programmability of the fabric.

Bio: Aaron Smith is a member of the Computer Architecture Group at Microsoft Research. He is broadly interested in optimizing compilers, computer architecture and reconfigurable computing. Over the past 15 years he has led multiple industrial and research compiler projects at Metrowerks/Freescale Semiconductor, The University of Texas at Austin and Microsoft. He received his PhD in Computer Science from UT-Austin in 2009 and is currently serving as co-General Chair of CGO 2015.

Enterprise NoSQL in the BBC

Hear why MarkLogic was chosen as the 2012 Olympic website content store to ingest, store and deliver the data and content assets to the BBC¹s mobile app and thousands of web pages.
Speaker: Paul Preuveneers, Director, Sales Engineering, MarkLogic

Paul Preuveneers has more than 9 years of development experience with MarkLogic, with expertise in running software teams as well as spearheading the European office of MarkLogic UK. Paul Preuveneers joined MarkLogic from Elsevier Science, where he led the Agile Development Team, working on leading edge products including the many CONSULT sites and the main strategic elsevierhealth.com site. Trained in Extreme Programming and Agile Techniques, Paul has been on the forefront of many of the most innovative applications using MarkLogic in Europe. Prior to Elsevier Science, Paul held positions at Action Information Management and gained his Bsc in Computer Science at Southampton University.

Big data, the Cloud and the future of computing by Dr Kenji Takeda, Microsoft Research

Abstract: We live in an information society, with cloud computing is changing the way we live, work and play in a world of devices and services. In this talk we’ll explore what, why and how this new era of computing is changing the way we think about conceiving, developing and delivering software and services. We’ll then look at how the concept of Big Data is transforming science, and the opportunities it presents for the future.

Bio: Dr Kenji Takeda is Solutions Architect and Technical Manager in Microsoft Research. He is currently focussed on Azure for Research and Environmental Science tools and technologies. The Azure for Research programme currently supports over 300 projects worldwide, including two at the University of St Andrews – see

http://www.azure4research.com

Kenji has extensive experience in Cloud Computing, High Performance and High Productivity Computing, Data-intensive Science, Scientific Workflows, Scholarly Communication, Engineering and Educational Outreach. He has a passion for developing novel computational approaches to tackle fundamental and applied problems in science and engineering.