Academic visitors funded through the Big Data Institute
- Dr Yang Wu (Kyoto University) & Dr Jinghao Xue (UCL)
Cascaded correlation analysis of Big Data: This cross-disciplinary project aims to initiate research collaboration on the development of a novel framework of cascaded correlation analysis of big data, between Dr Yang Wu from the Academic Center for Computing and Media Studies, Kyoto University and Dr Jing-Hao Xue from the Department of Statistical Science, University College London. Given a warehouse of pre-processed data from a variety of sources, the new framework aims to identify the correlation among the items in the warehouse and associate them with higher-level knowledge, for a fast, online and reliable recognition of targets. An application of this framework is the across-camera human tracing problem in the surveillance scenario, where the volume, velocity and variety of the data are high.
- Dr Zhanyu Ma (Beijing University of Posts and Telecommunications) & Dr Jinghao Xue (UCL)
Feature extraction of neutrally distributed big data This project aims to initiate multi-disciplinary research collaboration, between Dr Zhanyu Ma from the Pattern Recognition and Intelligent Systems Laboratory, Beijing University of Posts and Telecommunications and Dr Jing-Hao Xue from the Department of Statistical Science, University College London, on the development of a new approach to feature extraction of neutrally distributed big multimedia data. In many applications of multimedia data, there are data representing proportions. The analysis of such data remains an outstanding topic. The research objective of the project is to study the independence of non-Gaussian massive multimedia data and propose a computationally efficient strategy to extract features from such data.
- Dr Jonathan Ozik & Dr Nicholson Collier (Argonne National Laboratory and University of Chicago) & Dr Mark Altaweel (UCL)
One purpose of this visit is to continue ongoing joint efforts in developing a complex system model that looks at large urban regions and their relationship to water. This includes using text mining to create and develop agent behaviours in an agent-based model. The challenge is to create links between text mining and simulation, while validating outputs with modern urban environments and their relationship to water. This visit also offered two workshops open to UCL researchers and students in a variety of departments. One workshop was on high performance computing and computational modelling using RepastHPC, a free and open source tool developed by Collier and Ozik that runs on high performance systems, and the second workshop was on unstructured text mining using UIMA as a basis or NLTK.
- Dr Astrid Guttmann (Institute for Clinical Evaluative Sciences and University of Toronto) & Professor Ruth Gilbert (UCL)
The aim of the visit is to share and learn from the experiences of ICES in Ontario and the big data initiatives within UCL (the Big Data Institute, Children’s Policy Research Unit at UCL, the Farr Institute of Health Informatics London, the Administrative Data Research Centre- England, and the North Thames CLARHC- (Collaboration for Leadership in Applied Health Research) to inform the development of a joint research and exchange programme. The visit will focus on developing a specific project on transitions from paediatric/adolescent secondary care services to adult services for a range of chronic conditions. However, it is anticipated that there will be many other areas of common interest using linked health administrative data that could lead to joint research programmes. More fundamentally, methods underpinning the project will be relevant to many other patient groups and relevant to the development of IT systems at the interface between the service, government and academia to benefit all stakeholders and the public.
- Dr Dawn Nafus (Intel) & Dr Hannah Knox (UCL)
The main objective of this visit is to develop a case for how ethnographic methods might enrich big data analytics by providing a perspective on social data practices from the ‘bottom up’. The proposed visit will result in the publication of a position paper on this topic.
Since 2009 both Hannah Knox (UCL) and Dawn Nafus (Intel) have been involved in separate collaborative research projects that have begun to explore the specific implications of big data for anthropological understandings of social interactions and the role that ethnographic research might play in improving and/or supplementing big data analytics. The proposed research visit will provide a unique opportunity for Knox and Nafus to draw together these prior experiences with a view to publishing a position paper on the contribution of ethnographic research to discussions about the challenges of big data. An outline of a position paper will be drawn up during the research visit and a final paper will be submitted to the journal Big Data and Society within six months of the end of the visit.
- Professor M Sohel Rahman (Bangladesh University of Engineering & Technology) & Professor Richard Taylor (UCL)
The main long-term objective here is to develop a learning and predicting analytical software with stochastic capabilities for application in to groundwater science. The medium term goal would be to develop and implement a computational intelligence based algorithm, train it using the huge amount of data that is available to the research group of Prof. Taylor and finally validate the algorithm. The ultimate goal would be to develop software for the groundwater community that would be open source so that further development through similar collaboration with computer scientists is possible.
- Professor Mattias Wahde (Chalmers University of Technology) & Professor Philip Treleaven (UCL)
The overall goal of the project is to develop a software agent capable of real-time processing of large amounts of data from online sources, and (financial) decision-making based on the information thus obtained.