The Symposium is now fully booked and registration is closed.

PROGRAMME

13.15-14.00 Registration (Old Library)

13.30-14.00 Poster session (Old Library)

14.00-15.00 Language dynamics: a neurocognitive approach to incremental interpretation Prof. Lorraine Tyler (Dept. of Psychology)

Understanding spoken language involves a complex set of processes that transform the auditory input into a meaningful interpretation. Our percept is not of acoustic-phonetic detail but of the speaker’s intended meaning. This transition occurs on millisecond timescales, with remarkable speed and accuracy, and with no awareness of the complex computations on which it depends. How is this achieved? What are the processes and representations that support the transition from sound to meaning, and what are the neurobiological systems in which they are instantiated? Surprisingly little is known about the specific spatio-temporal patterning and the specific neuro-computational properties of this complex dynamic system. In current research we address these issues by combining advanced techniques from neuroimaging, multivariate statistics and computational linguistics to probe the dynamic patterns of neural activity that are elicited by spoken words and the incremental processes that combine them into syntactically and semantically coherent sentences. Computational linguistic analyses of language corpora enable us to build quantifiable models of different dimensions of language interpretation – from phonetics and phonology to argument structure and semantic integration - and we test for their presence using multivariate methods on combined electro- and magneto encephalography (EMEG) data, as the utterance unfolds in real time. In this talk, I will present the novel account of speech comprehension that is emerging from this research.

15.00-15.30 Natural Language Processing and online health reports (or OMG U got flu?) Dr Nigel Collier (Language Technology Lab, Dept. of Theoretical and Applied Linguistics)

Online health information is widely reported by individuals in social media, chat rooms, discussion boards and also by the news media. These informal sources of evidence about our individual health, attitudes and behaviours are driving the development of new techniques for studying human health in areas ranging from real-time disease surveillance, to understanding mental illness, to providing evidence for new applications of drugs. Informal patient data on the Web is increasing, accessible, low cost and seems likely to cover a greatly expanded population compared to traditional survey methods. However in order to understand and integrate this data researchers in Natural Language Processing (NLP) must grapple with theoretical, practical and ethical challenges. For example: How can machines achieve fine-grained analysis and understanding of laymen's health language? How does online health report data complement traditional survey data? How can we integrate online health data with other data sources such as databases and ontologies? What benefit could there be in a longitudinal analysis of an individual's online health reports over a period of time? My talk will illustrate answers to these questions in the light of our recent attempts to harness social and news media as a new type of signal for understanding human health.

15.30-16.00 Group presentation - poster presenters

16.00-16.30 Refreshments; poster session (Old Library)

16.30-17.00 Does natural language understanding have anything to do with understanding natural language? Prof. Ann Copestake (Computer Laboratory)

Natural language understanding (NLU) and natural language generation (NLG) are the main general goals of computational work on human language. Does such work have anything to tell us about the scientific/linguistic goal of understanding how natural languages behave? Can better understanding of linguistics help the development of practical computational techniques? It is far from obvious that the answer to these questions is yes, since successful computational modelling does not necessarily imply any real understanding. For instance, trajectories of physical objects can be modelled without understanding the underlying physics. Fred Jelinek, in a talk given at an award ceremony in 2004, admitted making the notorious comment 'every time I fire a linguist, our system performance improves' but argued that the goals of practical speech recognition simply did not coincide with the interests of the linguists. However, in this talk, I will suggest some more positive answers to these questions. I will describe DELPH-IN, a long-standing international collaboration involving researchers who are explicitly addressing both linguistic and computational goals, and discuss some of its successes. I'll outline how the development of precision grammars for various languages allows us to investigate language scientifically, and also to build practical systems for end users. I'll conclude by speculating how this sort of work can further progress in the brave new computational world of deep learning.

17.00-18.00 A molecular genetic perspective on speech and language

Prof. Simon Fisher (Max Planck Institute for Psycholinguistics, Nijmegen)

The rise of molecular technologies has yielded exciting new routes for studying the biological foundations of language. In particular, researchers have begun to identify genes implicated in neurodevelopmental disorders that disrupt speech and language skills. My talk will illustrate how such work can provide powerful entry points into critical neural pathways, using FOXP2 as an example. Rare mutations of this gene cause problems with learning to sequence mouth movements during speech, accompanied by wide-ranging deficits in language production and comprehension. FOXP2 encodes a regulatory protein, a hub in a network of other genes, several of which have also been associated with language-related impairments. Versions of FOXP2 are found in similar form in many vertebrate species; indeed, studies of animals and birds suggest it has conserved roles in the development and plasticity of certain sets of neural circuits. Thus, the contributions of this gene to human speech and language involve modifications of evolutionarily ancient functions. Searches for additional language-related genes are underway, taking advantage of dramatic advances in genomic methods. Overall, the FOXP2 story illustrates the value of an interdisciplinary approach for unravelling the complicated connections between genes, neurons, brain circuits and language.

18.00-19.30 Reception (Old Library)

Supported by Cambridge University Press

shcup txtblack cmyk

Registrations are closed

Cambridge Language Sciences Symposium 2016