The following scholars are invited to give the keynote during UKCI 2019.


Prof Yaochu Jin, University of Surrey

Title: Communication Cost Sensitive Federated Learning


Federated learning is a new learning paradigm for decentralized privacy-preserving machine learning. Instead of communicating data from the end devices such as mobile phones to the cloud, local models are trained on the local devices and a global model is generated by aggregating the local models communicated from the end devices. In federated learning, communication efficiency as well as the performance of the global model is of great importance. In this talk, I will present two recently proposed algorithms that can reduce the communication costs without degrading the performance of the global model. The first algorithm simultaneously minimizes the communication costs and classification error by optimizing the architecture of the local models using a multi-objective evolutionary algorithm. The second algorithm introduces an asynchronous update strategy together with a temporally weighted aggregation. Empirical results on commonly used datasets demonstrate the effectiveness of the proposed algorithms.


Prof Ahmad Lotfi, Nottingham Trent University

Title: Ambient Computational Intelligence


Computing will become transparent to humans and naturally involved in our everyday living. Ambient Computing refers to a digital environment that proactively supports people in their daily lives. It is an emerging discipline that brings intelligence to our living environments, makes those environments sensitive to us, and adapting according to the user’s needs. By enriching an environment with appropriate sensors and interconnected devices, the environment would be able to sense changes and support decisions that benefit the users of that environment.  Such smart environments could help to reduce energy consumption and thus the cost of facilities, improve safety and security, while at the same time increase user’s comfort. One specific area of interest is the application of ambient computing in Assisted Living, where the home environment provides assistance with daily living activities for people with different cognitive and physical disabilities. To enhance the intelligence of the environment, Computational Intelligence techniques as a set of nature-inspired computational methodologies are available to address such complex problems for which traditional approaches are ineffective. This talk will provide a review of the technologies and environments that comprise ambient computing.


Prof Trevor Martin, University of Bristol

Title: Approximately Right is Better Than Precisely Wrong


In recent years, the data-driven model of AI has underpinned an expansion of intelligent systems into a wide range of applications, and many autonomous AI systems perform tasks without significant human input. Examples include product recommendation, game-playing, control of appliances and driverless vehicles. In contrast, collaborative intelligent systems aim to use the complementary strengths of humans and computers in partnership for tasks such as assisted driving, computer-aided diagnosis and complex data analysis tasks such as enhancing the security of information systems, networks and devices.

In this talk, I will emphasise the need for cybersecurity systems to be based on human-computer collaboration, rather than on computer autonomy. Human analysts can provide insight and interpretation, while machines perform data collection, repetitive processing and visualisation.

An important aspect of collaborative intelligence is the common definition of terms used by humans and machines to identify and categorise the data. In this talk we will argue that graded concepts (based on fuzzy set theory) are a natural framework for the interaction and exchange of information between analysts and machines. We will describe a new approach to approximate (fuzzy) categorisation, and outline examples where this assists collaborative intelligence.


Prof Alexandra Cristea, Durham University

Title: Learning Analytics – where is it now and where is it going?


Computational Intelligence is currently a thriving area and is applied to many fields. Within this, an emerging important area is that of learner analytics, a new field enabled by the recent advances in analytical and visualisation tools for big data, and the improved data formats and advances in computing technology [according to the Higher Education Academic in UK]. Learner Analytics refers to the measuring, collecting and analysis of data about learners and their environments, with the aim of improving teaching and learning practices, usually in online environments. This talk uncovers this exciting new field of research, discussing challenges, success stories, specific aims and goals in the context of education, and future developments. Some recent results on MOOCs are also presented.


Prof Gabriela Ochoa, University of Stirling

Title: Artificial Evolution and Complexity: A visual Perspective


The intricacy and beauty of natural behaviours and shapes have inspired computational methods to understand and predict them, while Natural Evolution has inspired powerful computational problem-solving methods. We illustrate these two fields, Complexity and Evolutionary Computation, with a series of visual examples and metaphors; from the artificial evolution of plant-like structures to the automated design of transportation routes, educational timetables, and medical treatments.