Title:  Incorporating Streaming and Approximate Techniques in Graph Processing  

Date: 27/07/2018

Time: 14h00

Location: Sala 0.17, Pavilhão Informática II, IST, Alameda 

Advisors: Professor Luís Veiga / Professor Alexandre Francisco

Abstract: This work highlights recent contributions in distributed systems addressing graph processing, targeting (but not restricted to) algorithms such as community detection and vertex centrality. The study is being conducted as part of a larger context of big data, dynamic graphs and the emerging roles of approximate and incremental computing. Experiments are being conducted with the aim of analyzing the performance of distributed graph processing approaches in light of different algorithms in this context.  The systems used in empirical evaluation were the result of a two-phase study: 1) a broad qualitative analysis of several graph processing systems and databases and a choice based on desirable features; 2) a more in-depth analysis of the APIs and features of some systems. Given the increasing dimension of data sets represented as graphs in the advent of big data, how far can one relax from exact to approximate computing? How may we reliably define and ensure controlled error bounds given the volatility of graphs such as those representing social networks? And to what extent can performance be gained with a trade-off in accuracy in distributed system graph-processing scenarios? Employing distributed systems, how may one scale-out methods of graph clustering (as an example of a relevant graph algorithm) to process large graphs while harnessing the computational power of distributed computing infrastructures (potentially in line with statistical and machine learning techniques)? How easy is it to adopt incremental solutions for the processing of dynamic graphs? Could we update existing graph processing results based solely on previous results and a given change, or set of changes, in the graph?Establishing a basis for solving these challenges would pave the way for greater performance and resource-efficiency in the analysis of many graph-based big data scenarios. 

Title: Manipulation of 3D Objects in Immersive Virtual Environments

Date: 24/07/2018

Time: 14:30

Location: Anfiteatro PA-3 (Piso -1, Pavilhão de Matemática) IST, Alameda

Advisors: Professor Alfredo Manuel dos Santos Ferreira Júnior / Professor Joaquim Armando Pires Jorge

Abstract: Object manipulation is a key feature in most virtual environments (VE), including immersive VE. The spatial input typically associated with these immersive VE can offer natural metaphors, allowing users to directly grab, move and rotate objects in a similar way to how it is done in the physical world, making manipulations feel more natural. However, mid-air gestures compromisse manipulation accuracy, whether due to limitations in tracking solutions or the human dexterity itself. In this research, we aimed at reducing the impact of the lack of precision when interacting with 3D objects in mid-air within immersive VE. After surveying the state-of-the art on object manipulation following several interaction paradigms, we began by assessing the performance of existing mid-air manipulation techniques in both semi and fully-immersive VE through user evaluations. Our findings suggest that, if no restrictions exist, the best approach is to use exact mappings, within a fully-immersive environment. Focusing on increasing the precision of object manipulation in mid-air, we investigated if it can benefit from separating degrees-of-freedom (DOF), which has been proved useful in other interaction paradigms. Our results showed that single DOF control can improve precision at the cost of additional time for complex tasks. To test if it can be combined with scaled translations for increased precision, and if custom transformation axes could perform faster than using axes from fixed frames, we developed two novel manipulation techniques. We found that no significant improvements came from scaling down isolated translations, and that users favored 3-DOF manipulations above all, while keeping translation and rotation independent. The accuracy of mid-air gestures also play an important role in selecting objects outside arms’ reach. We developed a new selection technique for immersive VE, which combines cone-casting with na iterative progressive refinement strategy, teleporting users closer to intersected objects. Results of a user evaluation revealed our technique as a versatile approach to out-of-reach target acquisition, combining accurate selection with consistent times across different scenarios. In conclusion, we have validated our thesis, which states that DOF separation and iterative progressive refinement strategies can be successfully used to provide more effective mid-air interactions within immersive virtual environments.