On the power of experience

Artificial Intelligence is mimicking cognitive abilities. All sorts of behaviour we deem intelligent are being scrutinised and implemented in some way or another. Some more successfully so than others. Experience re-use is one of the successful approaches.

Experience guides us in our learning efforts and is one of the most important assets for problem solving. Experience is everywhere. For example, a recording technician needs experience in the studio to produce a recording worth listening to. Does the recording sound full and rich or still too tinny? Does the bass section sound overwhelming? A completely different field is gold ore mining. Lots of experience with mining and extracting the valuable metal has been published and even more specific knowledge resides in experts’ heads. For both domains holds that past experience — my own or someone else’s — can help me solve a current problem, for example, in the recording studio or when planning a new mining operation. Case-based reasoning, a methodology in which experience is expressed in the form of problem-solution pairs called cases, allows transferring and applying expert knowledge where needed.

Experience re-use can also be a means of personalisation, for example, in social media applications. Someone else’s similar context might improve my own user experience. The list is seemingly endless.

Thomas Roth-Berghofer, Professor of Artificial Intelligence at the University of West London, is giving a talk on this topic at the Technology and Innovation Conference Techsylvania in Cluj-Napoca, Romania, on 8 June 2015.


Master project on maintenance of myCBR knowledge models started

Estiak Bulbul, studying towards a masters degree in Intelligent Computing Systems at UWL, took on the task to design and develop a maintenance perspective for myCBR Workbench that allows to run quality assessments of the similarity measures and the case base. Maintenance of any kind performed on a system is a consequence of the fact that systems (or components) deteriorate or fail. The main idea to solve the maintenance problem of CBR systems is to employ a control loop that observes its performance and counteracts any change that is not desired. To realise the control loop, the most widely used CBR process model of Aamodt and Plaza has been enhanced with two more steps: Review and Restore. During Review the quality of the CBR system is assessed. If necessary the Restore step repairs the CBR system to get back to a desired level of quality. Estiak is looking into maintaining specifically the knowledge of product recommender systems.

The master project is co-supervised by Thomas Roth-Berghofer and David C. Wilson,  who is currently visiting researcher at UWL. Both worked on the topic of maintenance in case-based reasoning and now look into updating their results since their phd research.

Visiting researcher: Lotta Rintala at UWL

Lotta RintalaLotta Rintala, PhD student at the School of Chemical Technology, Aalto University, Finland, is visiting the University of West London. For four months she will be working with the research group of Prof Thomas Roth-Berghofer in the School of Computing and Technology. At the ICCBR 2012 Doctoral Consortium Lotta Rintala presented her research work on using CBR for ore pretreatment selection. Goal of the research visit is to study how to formally represent cases in the above mentioned domain using myCBR Workbench, and how to effectively retrieve relevant cases.

Publication: “Building Case-based Reasoning Applications with myCBR and COLIBRI Studio”

Abstract. myCBR and COLIBRI Studio are two well-established open-source frameworks for building case-based reasoning (CBR) applications, though they follow different approaches and support different phases of the CBR application development. Where myCBR supports its users in developing a knowledge model for representing cases, it leaves the software developers alone in developing an application that uses the generated knowledge model. COLIBRI Studio, on the other hand, is focused on the development of CBR applications. As soon as you have a knowledge model it offers templates for a variety of application types and supports in generating source code. This paper explains the strengths and weaknesses of both frameworks regarding the rapid development of CBR applications. It also shows how to use both of them in conjunction.

Thomas Roth-Berghofer, Juan Antonio Recio Garcia, Christian Severin Sauer, Kerstin Bach, Klaus-Dieter Althoff, Belén Díaz-Agudo, and Pedro A. González Calero. Building case-based reasoning applications with myCBR and COLIBRI Studio. In Miltos Petridis, Thomas Roth-Berghofer, and Nirmalie Wiratunga, editors, Proceedings of the 17th UK Workshop on Case-Based Reasoning (UKCBR), pages 71–82. School of Computing, Engineering and Mathematics, University of Brighton, UK, 12/2012.

Publication: “Explanation-aware Design of Mobile myCBR-based Applications”

Abstract. This paper focuses on extending the explanation capabilities of the myCBR SDK as well as on the optimisation of the myCBR SDK in the context of Android-based mobile application development. The paper examines the available knowledge for explanation generation within context-aware CBR systems. The need for the integration of new explanation capabilities is then demonstrated by an Android-based context- and explanation-aware recommender application. Upon the experience gathered during implementation of the prototype a process for the integration of explanation capabilities into the myCBR SDK is introduced. Additionally, constraints and requirements for the integration of explanation capabilities into myCBR are introduced. Within this process we distinguish domain dependent and domain independent knowledge. We do this with regard to the different requirements for the integration of explanation capabilities into myCBR for the two types of knowledge. The paper further details on our on-going effort to adapt the myCBR SDK for use on the Android platform.

Christian Severin Sauer, Alexander Hundt, and Thomas Roth-Berghofer. Explanation-aware design of mobile mycbr-based applications. In Belén Díaz-Agudo and Ian Watson, editors. Case-Based Reasoning Research and Development – 20th International Conference, ICCBR 2012, Lyon, France, September 3-6, 2012. Proceedings, volume 7466 of LNCS, pages 399–413. Springer, 2012.

Publication: “Similarity Knowledge Formalisation for Audio Engineering”

Abstract. This paper describes an approach to externalisation of the tacit knowledge used by experienced audio engineers to affectively describe emotions evoked by a sound or piece of music. We formalised the adjectives describing the timbre of a sound as well as their relationships. The main problems are the vagueness of emotions and the variation in the emotions the same single percept can trigger in different people. We demonstrate how similarity knowledge can be used to process fuzzy and incomplete queries to emulate the vagueness and differentiation associated with the emotions triggered by a sound percept. We capture the experience of audio engineers by mapping the formalised vocabulary of timbre-describing adjectives to their workflows, which describe the actions to change the spectral shaping of a sound and its emotional effect.

Christian Severin Sauer, Thomas Roth-Berghofer, Nino Auricchio, and Sam Proctor. Similarity knowledge formalisation for audio engineering. In Miltos Petridis, Thomas Roth-Berghofer, and Nirmalie Wiratunga, editors, Proceedings of the seventeenth UK workshop on Case-Based Reasoning UKCBR 2012, Cambridge, UK, 2012.

Tutorial: myCBR and Colibri Studio at AI 2012 Conference

From CBR researchers and students we learned that Colibri Studio and myCBR are perceived as competitors. But nothing is further from the truth. Colibri Sudio and myCBR complement each other. At ICCBR 2012 we gave tutorials on our open source tools (announced here) together and decided in discussions afterwards to further promote their interoperability at other venues.

We are happy to showcase myCBR and Colibri Studio at AI 2012, the thirty-second Annual International Conference of the British Computer Society’s Specialist Group on Artificial Intelligence (SGAI), which will be held in the attractive surroundings of Peterhouse College in Cambridge. The introductory tutorial will be given as part of the main conference by Christian Sauer, University of West London, and Dr Juan Antonio Recio García, Complutense University of Madrid.

Additionally, there will be a talk about myCBR and Colibri Studio at the 17th UK Workshop on Case-Based Reasoning (UKCBR 2012).

You can register online here.