SGAI

AI-2018 Thirty-eighth SGAI International Conference on Artificial Intelligence. CAMBRIDGE, ENGLAND 11-13 DECEMBER 2018

home | schedule | technical stream | application stream | poster sessions
workshops | proceedings | exhibition | registration | sponsors | organisers
enquiries | social | visa info | venue | accommodation | panel session | special session
ai open mic | information for speakers | previous conferences | letter of invitation

call for papers | paper submission and info for authors | accepted papers
internet access for delegates | walking tour |

BCS

Workshops

The first day of the conference comprises a range of workshops, to be held on Tuesday 11th December. Delegates will find these events to be especially valuable where there is a current need to consider the introduction of new AI technologies into their own organisations.

There will be four half-day workshops, including the Twenty-third UK CBR Workshop. Delegates are free to choose any combination of morning and afternoon sessions to attend. The programme of workshops is shown below. Note that the morning session starts at 11 a.m. to reduce the need for delegates to stay in Cambridge on the previous night. There is a lunch break from 12.30-13.15 and there are refreshment breaks from 14.45-15.15 and from 16.45-17.00.

Workshops organiser: Professor Adrian Hopgood, University of Portsmouth, UK


Stream 1 - Morning (11.00-12.30 and 13.15-14.45 Lubbock Room)

AI for Software Engineering

Chairs:
Alistair McCormick and Dr Joost Noppen, BT Research

Software development as a discipline has come a long way from its early beginnings. Languages have become more sophisticated and expressive and development processes refined to ensure the expertise of the developer is able to shine. Best practices such as automated version control, automated testing and static code analysis have contributed to increased software quality and better insight into development processes. But for all the benefit these tools bring, they come with (at times) a steep learning curve that can prohibit their use in the workplace.

At the same time there has been a rise in software development activities in areas where this was not traditionally done. For example data scientists that distil reasoning models from big data aim to make these available as software tools, but do not necessarily have the expertise or the time to exploit these advanced, best practice software engineering techniques.

The role AI can fill in software engineering is in bridging the gap between current development practice and the accepted, envisioned best practice. This can be achieved by offering automated support for setting up testing and deployment or version managing source files. Bleeding edge software developers familiar with these practices can be assisted by analysis algorithms that can automatically identify as yet unknown bugs and propose fixes. Algorithms can analyse legacy systems and suggest parts that can be turned into Software Product Line assets. In other words, AI can complement the activities of developers to help them gain access to best practice and best-of-class software engineering principles.

In this workshop we explore the landscape of AI in Software Engineering and how close we are to achieving this notion of human-machine-based best practice in development. We feature presentations by Software Engineering researchers who are working on AI algorithms to help developers in a variety of areas, ranging from formal methods to software product line development. In addition we will explore what the infrastructure for AI for Software Engineering looks like and how this can be managed in a large organisation.

Provisional programme:

  1. Software Product Lines (EBTIC, Sid Shakya)
  2. Predictive Analytics for Software Engineering (Federica Sarro, UCL)
  3. Software Repository Mining (BTIIC, Aftab Ali)
  4. A Model for Operationalising AI (BT, Joost Noppen)

Stream 1 - Afternoon (15.15-16.45 and 17.00-18.30 Lubbock Room)

Semantic Deep Learning: enhanced word embedding with semantic knowledge

Chair:
Dr Mercedes Arguello Casteleiro, University of Manchester

The benefits and the challenges
The scalability of neural language models obtained via Deep Learning opens up opportunities for: 1) routinely scanning the large bodies of documents (unstructured text); and 2) representing the meaning of terms (one or more words/tokens) in vector spaces. Transforming neural embeddings of n-grams into an augmented dataset of normalised and interlinked concepts provides the foundation to acquire reusable knowledge. However, the validation of this knowledge requires cross-checking with ground-truths that may be unavailable in an actionable or computable form.

Deep Learning algorithms have a "black-box representation". Wide acceptance and adoption of Deep Learning for professionals in a domain (i.e. domain experts) requires confidence and trust. To gain such confidence (transparency and interpretability), this workshop illustrates plausible methods for verifying if the associations (the semantic similarity and relatedness) captured by neural embeddings are reliable for human experts.

Main objective of the workshop
This workshop reviews the current understanding of semantically related terms derived from neural language models obtained via Deep Learning, and the interconnection between word distributions and the theoretical domains of computational linguistics and semantic web.

Healthcare as a use-case
The application of AI in medicine has a long-standing tradition. One day Deep Learning systems will be deployed in healthcare applications to acquire clinically meaningful knowledge locked within the biomedical literature and clinical narratives. Towards this aim, it is essential to prove the usefulness of Deep Learning for clinical activities, such as screening, diagnosis, or treatment assignment. This workshop presents the benefits of ontologies as a way of specifying the meaning of semantically related terms derived from neural language models obtained via Deep Learning, and thus, the role of ontologies supports the sharing and reuse of reliable biomedical/clinical knowledge that is formal and usable for a wide range of querying.

Further information
Please see the research group webpage for further information.


Stream 2 - Morning (11.00-12.30 and 13.15-14.45 Peterhouse Lecture Theatre)

23rd Case-Based Reasoning Workshop - session 1

Chair:
Professor Miltos Petridis, Middlesex University

Please see the UKCBR webpage for further workshop information.

Stream 2 - Afternoon (15.15-16.45 Peterhouse Lecture Theatre)

23rd Case-Based Reasoning Workshop - session 2

Chair:
Professor Miltos Petridis, Middlesex University

Please see the UKCBR webpage for further workshop information.

Stream 2 - Afternoon (17.00-18.30 Peterhouse Lecture Theatre)

Innovation session: Roborace

Chair:
Nadia Abouayoub

Innovation sessions are an opportunity to hear from industry professionals about how AI is used in their field. Mr Bryn Balcombe, Chief Strategy Officer at Roborace, will present Roborace and the use of AI for autonomous racing cars.

Please see the Roborace website for further workshop information.


SGAI

AI-2018 Thirty-eighth SGAI International Conference on Artificial Intelligence. CAMBRIDGE, ENGLAND 11-13 DECEMBER 2018

home | schedule | technical stream | application stream | poster sessions
workshops | proceedings | exhibition | registration | sponsors | organisers
enquiries | social | visa info | venue | accommodation | panel session | special session
ai open mic | information for speakers | previous conferences | letter of invitation

call for papers | paper submission and info for authors | accepted papers
internet access for delegates | walking tour |

BCS