News

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.

Event details

  • When: 21st January 2015 14:00 - 15:00
  • Where: Cole 1.33a
  • Series: School Seminar Series
  • Format: Talk

Research Evaluation Framework 2014

The results of the UK Research Evaluation Framework 2014 were released publicly today. The School of Computer Science was rated 14th in the UK (and a close 2nd in Scotland) in terms of research outputs, which is a measure of how our research compares to the best in the world. Over 30% of our submitted papers received the highest 4* ranking, with a further 55% receiving the 3* ranking.

“We’re delighted with this result,” said Prof Steve Linton, the Head of School. “Coming alongside our top placement in the UK for teaching Computer Science, it shows that that we’ve managed to achieve a balance between our two core activities while maintaining the quality of each.  It’s a confirmation of our high standing in the subject and leaves us excited for future developments.”

On the broader evaluation metrics that include research environment and non-academic impact, as well as research quality, the School was ranked 27th in the UK (4th in Scotland). “We submitted three impact case studies ranging from cloud computing to improving cultural engagement through virtual reality,” said Prof Simon Dobson, the School’s Director of Research. “These are all strategic areas that we’ll be keen to build on.”

Institutionally, the University of St Andrews was ranked 19th overall in the UK, and 2nd in Scotland.

FatFonts and a World Population Map

Looking for an unusual Christmas gift? Look no further, the first ever FatFonts World Population Map is now available in the Axis Maps store. All proceeds from the maps will fund more research and help us provide exciting internships for students.

fatFontPoster-01

Created by Miguel Nacenta and Uta Hinrichs, the map shows how the population of the world is distributed. It uses a typographic visualization technique – FatFonts which allows you to read the exact number of people living in a particular area with a precision within 100,000 people. View the gallery of lytro pictures for further images.

FatFonts are also available to download and use, or for the more inspired, try creating your own.

Although a quirky idea, FatFonts seem to have a bunch of usages… for example, they are convenient when you want to provide a table of numbers that is also a graphical representation. This allows the viewer (or the reader) to very quickly capture the overall distribution, but also to go in and read the specific number, which they can then use to compare to other numbers (in the FatFonts table or in their heads).

FatFonts are great in maps, and that is why Uta and I set out to create a poster that would give a picture of one of the most pressing issues of our time: world population. Thanks to SICSA (and our wonderful helpers Carson, Jed, and Michael), we got the time, money and support to develop the idea. The result is a poster that represents the population of the world using FatFonts. – Miguel Nacenta

Read more about the World Map and FatFonts on Miguel’s blog, the SACHI blog and at FatFonts.

Laidlaw Undergraduate Internship Programme

Deadline Monday 9th February 2015 at 5pm

This week, the University is launching an exciting new internship programme, ‘The Laidlaw Undergraduate Internship Programme in Research and Leadership’, thanks to the generous support from The Lord Laidlaw of Rothiemay. This is the biggest UG internship programme of its kind in Scotland and is a unique opportunity for students to gain paid professional work experience, contribute to high level research in the University and benefit from an intensive leadership training course.
Further information is available at the following webpage:-
https://www.st-andrews.ac.uk/students/involve/laidlaw/

Graduation December 2014

Congratulations to the Masters Class of 2014, and PhD students Dr Lakshitha De Silva and Dr Iain Parris who graduated today. The School also celebrated the installation of Professor Ian Miguel.
Students were invited to a reception in the school to celebrate their achievement with staff, friends and family.

CollageImage3

CollageImage5
Winter graduation coincided with a weekend of St Andrew’s Day celebrations captured in full glory by MSc student Xu Zhu.

Historic Streets, Peir, Harbour and Old Course

Historic Streets, Pier, Harbour and Old Course

R&A, Christmas Lights and Cèilidh.

R&A, Christmas Lights and Cèilidh.

Our graduates have moved on to a wide variety of interesting and challenging employment and further study opportunities, and we wish them all well with their future careers.

Funded PhD Research Studentships: Closing Date 15th December

A reminder that the closing date for applications is near. The School has funding for students to undertake PhD research in any of the general research areas in the school.

Are you a highly motivated student with an interest in these exciting research areas. A list of specific potential projects may help you decide. The studentships cover the cost of fees and an annual tax-free maintenance stipend. Exceptionally well-qualified students may be awarded an enhanced stipend.

For further information on how to apply, see our postgraduate web pages . We will make decisions on studentship allocation by February 27th 2015. Informal enquiries can be directed to pg-admin-cs@st-andrews.ac.uk or to potential supervisors.

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