Abstract: Modern biological research hinges on technologies that are able to generate very large and complex datasets. For example, recent advances in DNA sequencing technologies have led to global collections in the multi-petabyte range that are doubling every five months. These data require organising in a form that allows interpretation by a very large and diverse user community that are interested in everything from human health and disease, through crop and animal breeding to the understanding of ecosystems. In this talk I will first give an overview of core molecular biology concepts and some of the different types of data that are currently collected, I will then focus on work from my group in visualisation and analysis of sequence alignment data before turning to examples of prediction of properties and features from biological data.

Continue reading

Systems software, such as an operating system or a network stack, underlies everything we do on a computer, whether that computer is a desktop machine, a server, a mobile phone, or any embedded device. It is therefore vital that such software operates correctly in all situations. In recent years, dependent types have emerged as a promising approach to ensuring program correctness using languages and verification tools such as Agda and Coq. However, these tools operate at a high level of abstraction and so it can be difficult to map these verified programs to efficient low level code, working with bit-level operations and interacting directly with system services.

In this talk I will describe Idris, a dependently typed programming language implemented with systems programming in mind. I will show how it may be used to implement programs which interact safely with the operating system, in particular how to give precise APIs for verifiable systems programming with external C libraries.

Bio: Edwin Brady is a SICSA Advanced Research Fellow at the University of St Andrews
(http://www.cs.st-andrews.ac.uk/~eb)

Many ‘failures’ in software-intensive organisational systems result from human actions or inaction so, to reduce the possibility of such ‘failures’ and so improve system dependability, it is important to adopt a holistic view where system designers consider the environment in which a system will be used as well as the technical characteristics of the system itself.

In this talk, I will introduce the notion of responsibility modelling, which aims to represent the responsibilities of human and automated agents in a complex system, where the system may be created by integrating different systems from different agencies. The motivation for the work is that many ‘system failures’ are actually failures of agents in the system to fulfil their expected responsibilities and so responsibility modelling offers the opportunity to analyse how responsibilities are distributed and to identify responsibility vulnerabilities, before these lead to system failure.

Bio

 Ian Sommerville has been Professor of Computer Science at St Andrews University since 2006. For more details see http://www.software-engin.com.

In this talk we will introduce how new mobile base station architectures are evolving not only to meet demand but also to become “greener” since at current rates of deployment, mobile networks are becoming very large CO2 contributors. These new base station architectures are also becoming enablers for new Radio Access Networks (RANs) where the same infrastructure can support multiple radio technologies simultaneously by backhauling their baseband leading towards distributed antenna systems and distributed base station architectures. Some would argue that this represents a paradigm shift towards “cloud” enabled mobile networks. In the talk we will cover the reasons why things are progressing this way, how this is being enabled through technology innovation, and what Agilent has done to meet some of the test and measurement challenges in this evolved 4G wireless communications space.