On the Use and Abuse of Words In Biology, Phil Lord

In the distant past, biology was a word-centric discipline. Biologists studied life around them, and then wrote letters to each other about what they saw. Even twenty to thirty years ago, the amount of data that most biologists dealt with was small; while the letters had turned into journal articles, text was still the main form of communication.

Since that time, biology invented whole genomic sequencing and the amount of data has increased exponentially; and biology has dicovered computing and informatics. This has been an uneasy alliance in some ways: the obsession of biologists with free text goes against the computer scientists desire for a clean and co-ordinated data model; but the complexity of biology has stressed (or defeated) the computer scientists ability to describe things computationally.

Overtime, both sides have discovered a degree of communality, and are finding new ways to describe and catalogue the steadily increasing knowledge in biology. In time, this could have implications for however other discplines organise their knowledge to.

AI: Opportunities and Threats, Paul Watson

Vast amounts of data are being generated in every area of industry, and in every aspect of our everyday lives. How can we use this data to improve the economy and society for everyone's benefit? Data Science and AI are increasingly proposed as ways to achieve this, but what are they, and what are their strengths and limitations? This talk will introduce AI, giving real-world examples of the opportunities and threats that they bring.

Digital technology has greatly transformed the way in which we read literature, from the simple search functions that allow us to find an exact passage in any text to the digitization of resources that has the potential to democratize access to the most precious books in our libraries. However, no development in digital technology has been more transformative — and polemic — than the methodologies commonly known as ‘distant reading,’ or ‘computerised textual analysis’ or ‘text mining:’ essentially, a set of statistically-based ways of reading across a series of whole corpora of texts in search of patterns and trends. In a word, this type of computer analysis is currently the only digital method that is impossible (or unfeasible) for humans to repeat by hand: reading and analysing millions of texts simultaneously.

In this discussion lunch, James and Tiago will discuss some of the tools currently available for you to experiment with distant reading right now (such as Voyant Tools or Gale Digital Library), as well as discussing what these types of analysis can and can’t do. Finally, we’ll discuss whether new developments in machine learning can lead us towards a new way of computer-assisted literary research: can we teach computers to read like humans?

You can see the presentation slides here.

Nick Riches will be discussing MiMo an app he has created using the R programming language to help Speech and Language Therapists analyse linguistic data. There are a number of apps which serve this purpose, e.g. CLAN and SALT. However, these are limited by their steep learning curve. MiMo has been developed as a user-friendly alternative which provides a rich linguistic output. In particular, it

• Highlights words in particular colours according to their word classes, and thereby helps the Speech and Language Therapist to "see" the structure of sentences.
• Allows the user to rapidly identify utterances with particular linguistic features, through a search mechanism

The software also has potential for linguistics teaching. Firstly, it can help learners to acquire labels for word classes. Secondly, as it can process data in 60+ languages, it provides a tool for exploring typological differences across languages.