Dr Les Ball Intelligent Systems Division Of Computing, School of Computing, Engineering and Applied Mathematics, University of Abertay Dundee. Intelligence & Cyber. Security Social Network Analysis My research interests are centred on Intelligent Systems and their ability to enhance decision-making processes by quantitative analyses of intelligence data. Intelligence data relate predominantly to issues of national security but also concern commercial applications and health informatics. In security, applications are focused on preventative measures of what we term the ‘pre-attack’ phase (see diagram below). This involves the acquisition of biometric data and intelligence gathering from open sources. Understanding the ‘mood’ of someone also plays a part in assessing threats to security, while also finding application in customer satisfaction during purchasing transactions and in matters of health diagnosis where a patient may exhibit changes in their everyday behaviour. My work specifically addresses the questions ‘who are you? ’, ‘who do you know? ’, ‘what do you think? ’ and ‘how do you feel? ’. Did you know that your keyboard typing rhythm (dwell and flight times) is your digital equivalent to your handwritten signature? Well, we can also analyse and determine whether you are under stress when you are typing from the changes in your rhythm. Dowman M, Szymkowiak A, Coull N, Hapca S, Ball L. What Do You Think? Who Are You? Fingerprints have a long history in terms of identification. However, with improvements in technology biometric software is now capable of acquiring and analysing highresolution data for identification and verification purposes. Moreover, multimodal biometrics systems research is now coming to the fore whereby more accurate predictions are made by combining, say, face recognition with fingerprint data as below. These systems are very important in high security domains such as at airports and other border controls. http: //www. inmagine. com/ phins 104/phins 104098 photo Gesture Recognition The post-event Social Network Analysis model of the 9/11 attacks showing hijacker pilots (squared), densest section (left dashed box) incorporates 3 of the pilots and all of Mohamed Atta’s associates (circles), the chained structure of UA#175 associates (triangles), Hani Hanjour and associates (rhombi) are distinctly separated from the other hijacker pilots (right dashed box). Source: http: //orgnet. com/tnet. html The vision of the School’s Security Research Group identifies three phases on a temporal basis Keyboard Dynamics & Stress Signatures Textual data are being deposited exponentially at social networking sites such as Twitter and Facebook. This presents an opportunity to analyse these data to answer questions such as “what do people think about our new product? ” or “what is current opinion on the issue of Scottish independence? ”. Opinion mining and sentiment analysis are two synonymous techniques that use computational linguistics to solve these problems and offer rapid results that would otherwise be extremely time consuming to organise. The diagram below shows a typical architecture for a computer system to perform automatic extraction from social media to analyse political opinion. Sentiment Analysis/Opinion Mining Your facial expressions can tell us a lot about the type of emotion you are feeling by the detection of Facial Action Units. Experimental data below show that most people purse their lips when experience trouble using self-service checkouts i. e. a feeling of frustration. Related to lip press Martin C, Ball L, Archibald J. e. Health Indicators By sensing an outpatient’s data from everyday use such as movement within the house, keyboard usage or eye movements, we might detect a deviation or divergence from your normal pattern, which may in itself be an indicator of a change in health condition e. g. the onset of Alzheimer’s. Biometrics Ball L, Bradley D, Szymkowiak A, Brownsell S. References How Do You Feel? www. hitachi. com Who Do You Know? Social network analysis has been around for a long time within the social sciences. It is now, however, being applied to digital data from open sources. Its significance is that it can bring to light the emergence of a new criminal or terrorist cell in terms of its communications structure. If the key actors and substructures can be identified then intelligence can strategically target their elimination and disrupt or weaken the network. The following diagram was constructed by Valdis Krebs from open source information About the Author after the 9/11 attacks. Dr Les Ball is a lecturer within the School of Computing, Engineering and Applied Mathematics. He teaches principally how computer algorithms can be used to enhance decision-making processes in the fields of Intelligence and Security Informatics and e. Health. Poster template by Research. Posters. co. za We know that computers are great at number crunching and that they are also getting better at detecting who you are. Have you however wondered though if your computer might be able to understand how you feel? Affecting Computing is a discipline addressing this very question. We show what is called human ‘affect’ when we are under stress or in an emotional state. So, wouldn’t it be good if a computer could detect your mood by analysing your actions or expressions? The following three diagrams illustrate how keyboard dynamics, facial expression recognition and behaviour change could be used within the contexts of security, retail shopping and e. Health. Ball L, Bradley D, Brownsell S: 2011 Emotive Computing may have a Role in Telecare, Journal of Telemedicine and Telecare, 17(6): 279 -280. Ball L, Bradley D, Szymkowiak A, Brownsell S: 2011 Linking Recorded Data with Emotive and Adaptive Computing in e. Health, IEEE HISB 2011, 26 -29 July San Jose, 198 -204. Dowman M, Szymkowiak A, Coull N, Hapca S, Ball L: 2011 A Biometric System for Identifying Individuals and their emotional states from keystroke metrics, in prep for submission to PNAS. Martin C, Ball L, Archibald J: 2011 Emotion recognition in Self. Service Interactions, Third International Conference on Intelligent Human Computer Interaction, 29 -31 August, Prague. Contact Me For further information about the author use your smart phone to capture the QR code or go to http: //www. abertay. ac. uk Dr Les Ball. School of Computing, Engineering and Applied Mathematics, University of Abertay Dundee, Bell Street, Dundee, DD 1 1 HG. T : 01382 308251 E : l. ball@abertay. ac. uk
Скачать презентацию Dr Les Ball Intelligent Systems Division Of Computing
d475cc7939647efae94482f5bc686ca9.ppt