Institute of High Performance Computing

navigation

Powerding Discoveries!

Research

----

Computing Science (CS) Department

>>Department Members

Researchers in the CS Department develop techniques that draw out the efficiency, insight and intelligence in computing to power scientific discovery and technological advances.

The CS Department consists of seven Capability Groups that oversees the several key research areas. In addition, five horizontal programmes tackle application areas through various cross-cutting technologies derived from the different capability groups.

Figure 1: The Structure of the Computing Science Department.



High Performance Computing

The High Performance Computing (HPC) group aims to design and develop a generic programming and system framework that efficiently utilises heterogeneous architectures and accelerators to solve computational problems faster.

Research Directions / Challenges
Our R&D in productivity-enhancing tools enables HPC developers to perform source code transformations:

  • Tool to visualise code structure as a design aid to the developer
  • Tools for integrated code transformation as an alternative to manual editing
  • Computer-aided source code transformation for migrating legacy codes to HPC systems
  • Memory-hierarchy optimisation techniques for GPUs
  • Runtime adaptive tuning framework for new generation of HPC systems
We also provide consultancy and training to enable HPC developers to make better use of next-generation computing platforms:
  • Collaborate with researchers and industry partners to optimise and scale promising applications with significant performance gain
  • Establish programming know-how for multicore and GPUs
  • Train and develop talent pool of HPC application engineers
  • Provide a knowledge-base on various parallel programming models such as MPI, OpenMP, CUDA, etc.

Figure 2: Visualising code structure as a design aid to the developer. Above is an example showing data dependencies of a simple set of nested loops performing Gaussian elimination on a matrix.


Distributed Computing

The Distributed Computing group aims to establish new paradigms in the way computing technologies and platforms are being used. The research areas involve:

  • Design and implementation of end-to-end web-based solutions to power scientific discovery and to enhance efficiency in solving industry problems
  • New service-oriented paradigms and techniques in cloud computing such as software-as-a-service, MapReduce programming, etc.
  • Development of collaborative technologies to enable researchers and engineers to cooperate with each other remotely and effectively

Research Directions / Challenges

With the increasing prevalence of HPC-as-a-Service (HaaS), the group undertakes research to investigate how we can transform traditional HPC usage into pervasive, scalable and commercially-viable services. In particular, we are concerned about how HPC services can be established, valued, managed and used in a scalable and adaptive manner. Research activities explore:

  • New ways in which parallel and distributed HPC services can be mapped into a cloud-based service marketplace
  • New paradigms in negotiating service contracts and pricing services in a marketplace
  • Effective techniques in managing, adapting and scheduling services in a multi-provider ecosystem that help to maximise service quality and utility to both clients and providers
  • Reliable and cost-effective methods of automating the execution of data-aware workflow in hybrid clouds
  • Energy-efficient green computing technologies including adaptive virtual resource allocation, power management, and dynamic thermal-aware workload distribution in future data centers

Figure 3: FusionCloud for high performance computing as services.


Figure 4: Workflow management in hybrid clouds


Cross-disciplinary Data-intensive Analytics

Over the past few decades, we have observed (and still witnessing) an exponential growth of data in multitude of fields such as in life sciences and enterprises. Whilst the amount of data has grown through better acquisition devices, the methodologies to analyse and derive insight from this large volume of data, which is often of different types and varieties, has not kept up, resulting in a data deluge.

The Cross-disciplinary Data-intensive Analytics (CDA) group researches on analytics methodologies that reduces time- and space-complexities of computations, establishes frameworks and integrate tools to generate insight quickly and accurately from large-scale data and apply them to real problems.

The CDA group has the capabilities to handle the challenges brought about by big data. In-line with the research frontier of increasingly focusing on interfaces of different disciplines, we have multi-disciplinary capabilities that handle cross-domain analytics. This is crucial for today’s companies, who are keen to extract and derive insight from these data in order to gain competitive advantage and move up the value chain.

Research Directions / Challenges


Figure 5: Smart automated analytics design to enable layman to perform large-scale cross-domain data analytics

  • To investigate methodologies that reduces time- and space-complexities in order to address the challenge brought about by large data sets
  • To establish simple and reusable analytics framework, that can capture the know-how of domain experts and make their job more efficient through reuse
  • To design automated and optimised analytics workflow so that non-domain experts are confident in using the analytics workflow, through appropriate prompts and advice from the workflow designer system

  • Geometrical Modelling

    The geometrical modelling group engages in R&D activities to bring the science of computational geometry, digital modelling and image processing to domains such as computational science and biomedical engineering. The aim is to build a full range of capabilities to answer cross domain-specific challenges. These include:

    • Fundamental algorithms in the pre-processing of raw multi-domain digital information and manipulation of discrete digital models, converting them into appropriate input for simulation-driven and analysis applications;
    • Computational Geometry techniques for analysis of biomedical based shapes, surface meshes, or volumetric entities which can provide domain-specific insights;
    • Image processing capabilities that utilize efficient computational techniques on large biology-based image database

    Figure 6: Geometrical Modelling capabilities in computational science and biomedical domains.

    Research Directions / Challenges

    • Develop fundamental algorithms for processing 3D/4D mesh models, especially for human anatomical model in healthcare applications.
    • Establish shape analytics methodologies for analysing mesh properties of 3D/4D anatomical digital models to extract clinical insights.
    • Utilising image processing capabilities for extraction of useful domain specific indicators on biological-related image slices.

    Computational Social Cognition

    The Programme in Computational Social Cognition (CSC) has two primary kinds of goals — strategic goals and research goals. The strategic goals include creating a Cognitive Science capability in A*STAR and making the wider scientific community in Singapore aware of our capability. The first of these goals is, at least in the first instance, to be approached by engaging in high quality research in a relatively focused area. This focused area constitutes the driving long-term research goal, namely, to build computational models of what can be best thought of as “social intelligence.” In order to do this, it is necessary to develop computationally tractable models of the processes that underlie successful and appropriate interpersonal interaction and social perception.

    Many such models have been developed by social psychologists, but rarely are these models amenable to computer implementation. Indeed, one of our motives for selecting this area, apart from its intrinsic interest, is that it is an area that has received very little attention in Cognitive Science. It is thus necessary that we sometimes develop (or modify) and test psychological models of social processes (e.g., of (first) impression formation and revision), and in some cases that we explore architectures that are capable of supporting affective and social aspects of cognition. It is our hope that in the long run some of the work that we do can be used in the design of more socially intelligent robots and other computational agents, as well as making an independent contribution to research in Cognitive Science and Social, Cognitive, and Personality Psychology.

    Our research strategy is to develop several moderate-sized process models of select socio-cognitive constructs and their interactions that we view as particularly important, rather than trying to design a monolithic super-model where the need for system integrity might force us to postpone work on important constructs because there is no apparent way to knit them together. However, some of our projects do have conceptual linkages. For example, our project on impression formation needs as input descriptions of behaviour, while our project on action explanation seeks to generate descriptions of behaviour as its output.

    The diagram below illustrates our core concerns in basic theoretical research; it also shows a number of practical application areas to which we expect our research to be able to contribute. Following the diagram, we describe each of our current projects.

    For more information, visit the CSC website at http://cogsys.ihpc.a-star.edu.sg

    Figure 7: Research in the Computational Social Cognition group.


    Complex Systems

    The Complex Systems (CxSy) group aims to develop fundamental capabilities in complex systems modeling to help understand, design, manage and evaluate systems that exhibit complex system behaviour, with applicability to urban planning and logistics. The scope of the CxSy group includes:

    • A review, verification, validation and calibration of existing laws/models/theories and subsequently create new ones appropriate for Singapore in above domains
    • An investigation of how to model interaction of multi-scale models and how to capture the dynamics of emergent behaviour of a complex system.

    The team is currently focusing on four application areas:

    Land Use
    • Modelling and quantification of the emergence of land use in various cities and prediction of various factors such as temperature and transport ridership based on the land-use data.
    Transportation
    • In-depth investigation of the public transportation system of Singapore in order to establish insight on the resilience of the transport systems and characterizing the travel patterns of commuters in Singapore
    Complex Networks
    • Analysis of the structure and behavior of various complex networks that form the underlying fabric of urban systems. (social networks, transportation networks, interaction networks)
    Housing Demand
    • Modelling the demand of public and private housing in densely populated Singapore and its dependence on various drivers such as demographic, government policies, economic conditions, etc.

    For more information, visit the CxSy website at http://www.ihpc.a-star.edu.sg/cxsy

    Figure 8: Real World Problems, Real World Data, Complex Systems Tools


    Intuitive Interaction Technologies

    (More content to be updated, check back later!)


    Strategic Social Systems Programme

    The Strategic Social Systems (S3) programme is an interdisciplinary collaboration which addresses public, consumer, social and economic needs for innovative insights and foresights, bringing our scientific knowledge, know-how and high-quality IPs in value co-creation with our partners.

    The programme covers three key translational project initiatives including:

    • Strategic Public Information and Communication Enhancement (SPICE) for enhancing public communication in the social era, and in particular for health domain,
    • Social Marketing Intelligence Enhancement (SMILE) for advancing consumer knowledge
    • Social Healthcare Information Enhancement (SHINE) for augmenting public health communication

    Most recently, we included another translational initiative that aims at enhancing business competitiveness leveraging capabilities in e-commerce technologies:

    • Advanced e-Commerce Technologies (ACT) for facilitating business-to-consumer transactions (e.g., e-negotiation mechanism that allows configurable automated sales

    The strongly translational, value co-creation research approach for this programme involves:

    • Social data processing and analytics. Working with our domain partners, we collect and process unstructured social data (e.g., social media content, phone logs, email archives, unprocessed point-of-sales data) and organize them into useful forms of information, including social sentiment and their prevailing emotions, key topics, top opinion influencers and consumer psychographics. This is in line with addressing the “variety” and “veracity” challenges of big data needs
    • Computational modelling and system prototyping. To scale for needs in applied scenarios, we also build computational models that augment our analytics with simulation capabilities for more systematic, holistic insights. This is in addressing the “volume” and “velocity” challenges of big data. Furthermore, we also develop methods and prototypes that enable fast “Proof-of-Concept (POC)” of key design ideas.
    • Field validation for technology adoption. We have expertise in field research that help to formalize key design ideas (e.g., strategies and tactics for decision and communication, new customer segmentation methods, new incentive mechanisms) and iteratively validate them in field, behavioural settings. The value of behavioural validation is to obtain empirical confidence of the desired effects of key design ideas from our analytics and modelling work. To facilitate successful adoption, we also assist our partners by guiding their software engineering efforts if they decide to implement our methods in their own environments.

    Figure 9: Strategic Social Systems, an interdisciplinary approach to social and economic needs


    Sustainable City Life Programme

    This programme aims at improving the way people live, study, work and play. We are focusing on two objectives:

    • Helping people plan their daily tasks
    • Personalized recommendation for products and services.

    Currently, people plan their tasks in a few steps: find location for each task, check whether they can do two or more tasks at a location, and get route information to that location. We automate this process by developing a system that returns a few such (common) locations on the map, for a set of user tasks. To achieve this objective, we have collected data on public services and businesses in Singapore and developed algorithms to minimize the number of places travelled and the distance travelled. These technologies are realized in the form of a mobile app, where users can enjoy the benefits of knowing the common locations on the map, and intuitively interacting with them for the details of the business for each task, and the fastest route to get there. We are developing more features and are open for partnership.


    The second objective is to build social recommender systems that provide accurate recommendations for products and services. The accuracy of existing systems is often limited by the amount of information on users and how they can understand and predict user behaviors. Our research would address these gaps by:

    • considering and analyzing user behaviors observed through multiple channels, such as social networks;
    • using psychographics to explain and predict behaviors and thus recommending items based on inferred user personalities and preferences; and
    • building a distributed large-scale data processing framework.

    We have participated in competitions (Yelp Challenge 2013, Expedia Contest 2013) and obtained consistently good results. Better recommendations would benefit users, reducing cognitive load and improving satisfaction. Initial discussions with companies suggest that social recommender systems have broad applications across sectors and would lead to increased sales for companies.

    Figure 10: The Sustainable City Life Programme.


    Logistics and Supply Chain Management Programme

    (More content to be updated, check back later!)


    Computationally-driven Biomedical Engineering Programme

    (More content to be updated, check back later!)


    Rolls Royce Singapore - IHPC Computational Engineering

    (More content to be updated, check back later!)


    Dr. Rick GOH Siow Mong
    Department Director



    This page is last updated at: 2-Dec-2013