Collaborative Research Initiatives

IBM Research and Laboratorio Nacional de Computacao Cientifica will 1) modify IBM Deep Thunder to allow the "plug-and-play" use of other open source weather codes which are customized by the Brazilian scientific community (eg, BRAMS and OLAM). Deep Thunder currently uses the WRF-ARW open source model; 2) work on the scalability of BRAMS and OLAM codes on HPC platforms such as BlueGene/P.

IBM Research and Akademia Gorniczo-Hutnicza University of Science and Technology will develop cross-cloud application deployment governance in IaaS clouds and QoS on top of OpenStack.

IBM Research and Ohio State University will collaborate on application of low-latency network and HPC techniques to transactional and application middlewares.

IBM Research and the University of California, San Diego will extend the Olive VM image library architecture with a knowledge base that contains the indexed fingerprints of software components found within those images, together with properties inferred about their behavior (e.g. propensity for defects). There are three research challenges addressed here: (a) how to leverage machine learning and data analytics technologies to mine the contents of VM images and infer properties about software configurations found within them, (b) how to compute fingerprints of large VM images that enable similarity detection and content search, and (c) how to use the information derived from (a) and (b) for applications such as security analysis.

IBM Research and Carnegie Mellon University are developing the architecture and component technologies for an open virtual machine image library to provide a public domain virtual machine image collection. This will include services for publishing, searching, securing, and maintaining images, and instantiating those images on clouds, servers or desktops that support the execution of standard virtual machine image formats.

IBM Research and University of Fort Hare will evaluate the use of people as a source of information for cities concerning city affairs ('crowdsensing'). The vision is that cities and states will collect information from citizens to support operational and planning needs.

IBM Research and University of Minnesota will investigate and develop a middleware platform for mobile crowdsensing applications to enable the efficient collection, analysis and access of sensor data from mobile devices and to support the human participants in crowdsensing. The parties see to demonstrate the validity of the middleware platform by exploiting it to build innovative smarter planet applications.

IBM Research and Koc University will develop methods for detection of deviations from transportation plans in supply networks due to supply shortages, unexpected demand conditions, product promotions and unexpected problems in the transport chain using proactive event monitoring and business analytics.

IBM Research and National Taiwan University will develop open source software and documentation to enable effective and efficient use of Power systems accelerated with reconfigurable logic with a special focus on systems where the reconfigurable logic is dynamically reconfigured. The intent is to drive the proposed work and demonstrate benefit with workloads of mutual interest.

IBM Research and Carnegie Mellon University will develop powerful analytics for genome-wide association studies in complex diseases; providing efficient and scalable implementations using BigData technologies; developing an infrastructure to setup a pipeline for the integrated execution of these algorithms together with other systems.

IBM Research and Technion will perform exploratory activities for improved performance, reliability and manageability in cloud computing.

IBM Research and University of Moratuwa will collaborate on the topic of workload optimization for hybrid architectures. Hybrid systems are based on a wide architecture spectrum from heterogenous cores of network-attached system elements/accelerators/FPGAs. The research will include the symbiotic interaction of software and hybrid system architectures. The research would involve systematic approaches to innovate offline and online learning mechanisms to characterize workloads for acceleration and to analyze and predicate performance effects.

IBM Research and Massachusetts Institute of Technology will collaborate on the topic of workload optimization for hybrid architectures. Hybrid systems are based on a wide architecture spectrum from heterogenous cores of network-attached system elements/accelerators/FPGAs. The research will include the symbiotic interaction of software and hybrid system architectures. The research would involve systematic approaches to innovate offline and online learning mechanisms to characterize workloads for acceleration and to analyze and predicate performance effects.

IBM Research and National Taiwan University will develop open source software and documentation to enable socially-engaged personal wellness management. The scope of work includes model design, system development and a demonstration of project achievements.

IBM Research and the University of Massachusetts will collaborate to develop cloud computing technologies targeting smarter planet mechanisms and conduct an open, transparent evaluation of these technologies to facilitate the rapid advancement of the cloud computing field.

IBM Research and the University of Haifa will develop a crowd computing platform which may be used to develop crowd computing applications. Such a platform would allow defining the required task, the destination crowd and the allowed incentives as input and will provide the output of the required task by leveraging the crowd given the incentives.

IBM Research and Columbia University will collaborate using the DeepQA technology developed by IBM which has already started showing strong potential. This collaboration is meant to create a prototype system that leverages large components of the DeepQA system, Natural Language Processing ("NL") parsing capabilities adapted for the medical domain and associated content, and that uses a variety of medical knowledge sources (textbooks, medical reports, case studies) to provide relevance-ranked passages that answer questions posed by a doctor.

IBM Research and University of Trento will collaborate using the DeepQA technology developed by IBM which has already started showing strong potential. This collaboration is meant to create a prototype system that leverages large components of the DeepQA system, Natural Language Processing ("NL") parsing capabilities adapted for the medical domain and associated content, and that uses a variety of medical knowledge sources (textbooks, medical reports, case studies) to provide relevance-ranked passages that answer questions posed by a doctor.

IBM Research and University of British Columbia (Canada) will develop software for design in inverse problems with emphasis on large-scale problems.

IBM Research and National Taiwan University will develop open source software and documentation to enable effective and efficient use of Power Systems accelerated with reconfigurable logic with a special focus on systems where the reconfigurable logic is dynamically reconfigured. The intent is to drive the proposed work and demonstrate benefit with workloads of mutual interest.

IBM Research and Renmin University of China will collaborate with a unique approach to develop a large scale cloud database system that helps to achieve large data management and large data processing in the Cloud environment. The project is based on the key technology of massive data organization, management and service etc. used in the Cloud Computing platform.

IBM Researchand University of Oulu (Finland) will explore enabling intelligent emergency response coordination during natural disasters and large events (e.g., Super Bowl) within an effected geographic area by drawing upon collective expertise through extraction of community-generated content (e.g., Twitter feeds, Facebook blogs).

IBM Research, Portland State University, and the Vrije Universiteit Brussel (Belgium) will explore the ease of programming and run time performance of the Renaissance manycore programming model. This exploration will be done by jointly developing and testing computer programs written in the Ly programming language, or it's variants.

IBM Research—Zurich and ETHZ - Federal Institute of Technology Zurich (Switzerland) will develop innovative methods and algorithms for the creation and refinement of system implementations with emphasis on security certification.

IBM Research, Ghana Telecom University College, and University of Dar es Salaam (Tanzania) will collaborate to port the current iLab reference implementation to commonly available open source software technology such as Linux, Apache, MySQL and PHP and the prototype of the XLab in Holography and Optics.

IBM Research and Zurich University of Applied Sciences (Switzerland) will develop a modular, open-source framework that is able to simulate the operation of an entire railway network. The framework's architecture should be modular and define and support open standards, in order to interact with existing components (e.g. established commercial packages for physical simulation of trains and the signaling system) as well as new components based on advanced analytics and optimization (such as operations management, scheduling, and performance measurement modules).

IBM Research, Carnegie Mellon University, and North Carolina State University are developing the architecture and component technologies for an open virtual machine image library to provide a public domain virtual machine image collection. This will include services for publishing, searching, securing, and maintaining images, and instantiating those images on clouds, servers or desktops that support the execution of standard virtual machine image formats.

IBM Research and the King Abdullah University of Science and Technology (KAUST) are developing a scalable data management layer that is capable of managing petabytes of data efficiently and supporting real-time decision making. The application needs of large-scale scientific and web applications will be studied, and mechanisms that can provide the required data manipulation functionalities will be developed with a focus on the efficiency and accuracy of data movement and data analysis. This collaboration is critical to enable scientific clouds as large data needs to be shared and analyzed across customer sites and cloud providers.

IBM Research, the Indian Institute of Technology- Kharagpur and the Indian Institute of Technology-Madras will collaborate to develop dynamic state estimators using the Phasor Measurement Unit PMU) data. 1. Formulating a wide area measurement system which includes PMUs and communication systems- with constraint of minimum cost (i.e. least number of PMUs) satisfying observability condition 2. Designing the new algorithms for Dynamic State Estimation (DSE) suitable for the synchrophasor data set 3. In many systems where Supervisory Control and Data Acquisition (SCADA) system is available, Wide Area Monitoring (WAM) technology may include PMUs in phases- then a hybrid state estimator will be suitable choice which will include both SCADA data and PMU data.

IBM Research, University of Illinois at Urbana-Champaign, and University of Minnesota-Twin Cities will investigate and develop a middleware platform for mobile crowdsensing applications to enable the efficient collection, analysis and access of sensor data from mobile devices and to support the human participants in crowdsensing. The parties see to demonstrate the validity of the middleware platform by exploiting it to build innovative smarter planet applications.

IBM Research and Norwegian University of Science and Technology - NTNU will develop innovative algorithms for reservoir management and production optimization with emphasis on real world problems.

IBM Research and Technion will perform exploratory activities for improved performance, reliability and manageability in cloud computing.

IBM Research and the Tecnologico de Monterrey Mexico City Campus will collaborate in developing and testing epidemiologic models and observatory models for Influenza in the Mexico City area using IBM SYSTEM-PHIAD software applications.

IBM Research and MIT will collaborate on the topic of workload optimization for hybrid architectures. Hybrid systems are based on a wide architecture spectrum from heterogeneous cores of network-attached system elements/accelerators/FPGAs. The research will include the symbiotic interaction of software and hybrid system architectures. The research would involve systematic approaches to innovate offline and online learning mechanisms to characterize workloads for acceleration and to analyze and predicate performance effects.

IBM Research, Carnegie Mellon University, Massachusetts Institute of Technology, and the University of Virginia are exploring the use of a common set of analytic techniques to characterize key structural and behavioral properties of software architectures and the teams supporting them. They will also explore how to use the characterizations to allow business value discussions around software development. Software tools developed will be released under the Eclipse Open Source license.

IBM Research, Carnegie Mellon University, and Oregon State University are developing engineering models of user behavior for software application use and navigation. This will include developing, prototyping and validating cognitive models encoded in a cognitive architecture. that will be open sourced. The models will be integrated into software design tools that can be used to measure, compare and validate the usability of software application designs.

IBM Research and the Tel Aviv University (Israel) will make use of and contribute new building blocks to IBM's bioclinical data-mining (BDM) tool, which allows combining building blocks performing various functions into a workflow that performs a high level task, such as learning and cross validation, feature selection, and so on. The IBM BDM tool itself is not part of the scope of this collaboration.

IBM Research and the University of Massachusetts will collaborate to develop cloud computing technologies targeting smarter planet mechanisms and conduct an open, transparent evaluation of these technologies to facilitate the rapid advancement of the cloud computing field.

IBM Research and the University of Massachusetts will collaborate to develop cloud computing technologies targeting smarter planet mechanisms and conduct an open, transparent evaluation of these technologies to facilitate the rapid advancement of the cloud computing field.

IBM Research and University of Maryland are developing Open Source, Declarative Machine Learning and Graph Analytics packages built on top of IBM's Hadoop platform.

IBM Research and Columbia University will collaborate using the DeepQA technology developed by IBM which has already started showing strong potential. This collaboration is meant to create a prototype system that leverages large components of the DeepQA system, Natural Language Processing ("NL") parsing capabilities adapted for the medical domain and associated content, and that uses a variety of medical knowledge sources (textbooks, medical reports, case studies) to provide relevance-ranked passages that answer questions posed by a doctor.

IBM Research and University of British Columbia (Canada) will develop software for design in inverse problems with emphasis on large-scale problems.

IBM Research and Renmin University of China will collaborate with a unique approach to develop a large scale cloud database system that helps to achieve large data management and large data processing in the Cloud environment. The project is based on the key technology of massive data organization, management, and service, etc. used in the Cloud Computing platform.

IBM Research and Madeira Interactive Technologies Institute of University of Maderia (Portugal) will explore enabling intelligent emergency response coordination during natural disasters and large events (e.g., Super Bowl) within an effected geographic area by drawing upon collective expertise through extraction of community-generated content (e.g., Twitter feeds, Facebook blogs).

IBM Research, Portland State University, and Vrije Universiteit Brussel (Belgium) will explore the ease of programming and run time performance of the Renaissance manycore programming model. This exploration will be done by jointly developing and testing computer programs written in the Ly programming language, or it's variants.

IBM Research—Zurich and ETHZ - Federal Institute of Technology Zurich (Switzerland) will develop innovative methods and algorithms for the creation and refinement of system implementations with emphasis on security certification.

IBM Research, Stanford University, and the Tel Aviv University (Israel) will develop software tools to enable existing Java programs to scale to multiple cores. The software tools will be designed to be interactive, and to combine dynamic and static program analysis. It is anticipated that software implementing prototype analysis algorithms will be developed which will be released in open source form.

IBM Research, Indian Institute of Technology - Bombay (India), National Institute of Design (India), and the University of Tokyo will collaborate on the topic of multimodal platform for the less literate and low-income users who can potentially use information technology (IT) but who currently do not use any or very limited IT. They will explore the current communication methods and then derive an interface on mobile devices that can be used by people on the other side of the digital divide to access information technology solutions.

IBM Research, Ghana Telecom University College, and University of Dar es Salaam (Tanzania) will collaborate to port the current iLab reference implementation to commonly available open source software technology such as Linux, Apache, MySQL and PHP and the prototype of the XLab in Holography and Optics.

IBM Research, Carnegie Mellon University, Massachusetts Institute of Technology, and the University of Massachusetts will collaborate in developing the Open Advancement of Question Answering architecture, data model and component technologies. They will also collaborate to conduct open, transparent evaluation of these technologies to facilitate the rapid advancement of the question answering field.

IBM Research, Carnegie Mellon University, and North Carolina State University are developing the architecture and component technologies for an open virtual machine image library to provide a public domain virtual machine image collection. This will include services for publishing, searching, securing, and maintaining images, and instantiating those images on clouds, servers or desktops that support the execution of standard virtual machine image formats.

IBM Research and the University of Albany are developing models and methods for specifying and detecting patterns of interest in financial data streams and are evaluating the resulting models and methods using IBM's stream processing middleware. It is expected that the techniques developed will be made available as open source under the Eclipse Public License.

IBM Research and the King Abdullah University of Science and Technology (KAUST) are developing a scalable data management layer that is capable of managing petabytes of data efficiently and supporting real-time decision making. The application needs of large-scale scientific and web applications will be studied, and mechanisms that can provide the required data manipulation functionalities will be developed with a focus on the efficiency and accuracy of data movement and data analysis. This collaboration is critical to enable scientific clouds as large data needs to be shared and analyzed across customer sites and cloud providers.

IBM Research, the Indian Institute of Technology- Kharagpur, and the Indian Institute of Technology-Madras will collaborate to develop dynamic state estimators using the Phasor Measurement Unit (PMU) data. 1. Formulating a wide area measurement system which includes PMUs and communication systems- with constraint of minimum cost (i.e. least number of PMUs) satisfying observability condition 2. Designing the new algorithms for Dynamic State Estimation (DSE) suitable for the synchrophasor data set 3. In many systems where Supervisory Control and Data Acquisition (SCADA) system is available, Wide Area Monitoring (WAM) technology may include PMUs in phases- then a hybrid state estimator will be suitable choice which will include both SCADA data and PMU data.

IBM Research, the University of California, San Diego, University of California, Santa Barbara, and the University of Rome La Spienza are performing research on foundational aspects of the artifact-centric approach to the management and execution of business processes and operations, with the focus on the relationship of ontologies to the artifact-central approach. The parties are working together to develop prototype business models and will make any software developed available as open source.

IBM Research and the Polytechnic Institute of NYU are developing techniques for determining the reputation of hosts in IP networks. A framework will be developed that determines and leverages information about known comprised hosts, as well as information about network similarity relationships to arrive at a reputation score for every host in the network.

IBM Research, McGill University, and the University of British Columbia (Canada) will (1) explore ways to teach distributed software development, (2) examine how Rational Jazz-based technology can help and (3) create ways to broker meaningful collaborations with instructors, students, and IBM.

IBM Research and the University of California, Santa Barbara are developing profiling and analysis tools for dynamic scripting languages and using these tools to investigate the behavioral and performance characteristics of modern web workloads. Performance monitoring and performance analysis toolkits will be develop and released as open source.

IBM Research, the Georgia Institute of Technology, and the University of California at Berkeley will develop an open test-bed for developing, testing, benchmarking and validating new and innovative network defense technologies for cybersecurity. The work will include exploring mechanisms for contributors to submit algorithms to be benchmarked on data that is not visible to them and to get results that contribute data and run a variety of analytics on them. The work will also include building models to profile and flag malware and disruptive attacks, and a platform that can include formulation of benign and malicious behavior models, test beds, benchmarks, and associated tools, and creation of common data sets.

IBM Research and the Tecnologico de Monterrey Mexico City Campus will collaborate in developing and testing epidemiologic models and observatory models for Influenza in the Mexico City area using IBM SYSTEM-PHIAD software applications.

IBM Research—China and Tsinghua University aim to advance research in: 1) representative identifying technical problems on wireless network cloud, 2) evaluate some typical wireless transmission technologies over Open IT technologies, especially including timing and frequency synchronization, channel estimation, MIMO detection, modulation/demodulation and channel coding/decoding, 3) design and implementation of the prototype test bed (excluding the RRH Hardware) on multicore platforms, which will become the base for wireless network cloud research, 4) any design methods, architectures and related interfaces, materials about RRH hardware are not included in the scope of this agreement.

The IBM Watson Research Center, Carnegie Mellon University, the Harvard Business School, and the University of Virginia are investigating the use of a common set of analytic techniques for exploring key structural and behavioral properties of software architectures and the teams supporting them. In addition, the parties are working to understand how these properties can be characterized to allow business value discussions around software development. New software tools developed will be open sourced under the Eclipse Open Source License.

The IBM Watson Research Center, Carnegie Mellon University and Oregon State University are developing engineering models of user behavior for software application use and navigation. This will include developing, prototyping and validating cognitive models encoded in a cognitive architecture that will be open sourced. The models will be integrated into software design tools that can be used to measure, compare and validate the usability of software application designs.

IBM Research—China and Beijing University of Posts and Telecommunications are performing research on leveraging cloud computing and other distributed computing technologies to transform service provider's platforms to support emerging mobile internet applications and mobile analytics. The initial focus is on service overlay networks for mobile widgets hosting and software for mobile social network analysis.

The IBM Watson Research Center, Carnegie Mellon University and North Carolina State University are developing the architecture and component technologies for an open virtual machine image library to provide a public domain virtual machine image collection. This will include services for publishing, searching, securing, and maintaining images, and instantiating those images on clouds, servers or desktops that support the execution of standard virtual machine image formats.

The IBM Watson Research Center and the University of Albany are developing models and methods for specifying and detecting patterns of interest in financial data streams and are evaluating the resulting models and methods using IBM's stream processing middleware. It is expected that the techniques developed will be made available as open source under the Eclipse Public License.

The IBM Watson Research Center, the University of California, Santa Barbara and the University of Rome La Spienza are performing research on foundational aspects of the artifact-centric approach to the management and execution of business processes and operations, with the focus on the relationship of ontologies to the artifact-central approach. The parties are working together to develop prototype business models and will make any software developed available as open source.

The IBM Watson Research Center and the Polytechnic Institute of NYU are developing techniques for determining the reputation of hosts in IP networks. A framework will be developed that determines and leverages information about known comprised hosts, as well as information about network similarity relationships to arrive at a reputation score for every host in the network.

The IBM Watson Research Center and the University of California, Santa Barbara are developing profiling and analysis tools for dynamic scripting languages and using these tools to investigate the behavioral and performance characteristics of modern web workloads. Performance monitoring and performance analysis toolkits will be develop and released as open source.

IBM Research—China and Tsinghua University are investigating wireless network technologies, especially software radio technologies implemented on multicore processors. The work includes evaluation of timing and frequency synchronization, channel estimation, modulation/demodulation and channel coding/decoding.

IBM Research—China and Tsinghua University are creating a set of tools for 3D content creation and processing, as well as creating 3D models using these tools. These tools will help virtual world developers and users easily create and process models for virtual worlds. Based on this work, they will propose an ideal 3D model format for virtual worlds.

IBM Research—China and Tsinghua University are developing a green supply chain that helps to achieve the optimal balance between carbon footprint, supply chain cost, service level, and product quality. A dashboard is being developed for the simulator as is software for the analysis and optimization of green supply chains.

The IBM Watson Research Center, the New York University Stern School of Business and the University of Illinois at Urbana-Champaign are developing techniques to estimate security risks arising from IT systems and to determine the impact of security investments and countermeasures to reduce security risks. They are also developing the underlying theory, models and algorithms to support risk-based decision making about the acquisition, storage, use and governance of data.

IBM Research, Almaden and Yale University are conducting research on novel algorithms for medical image and video analysis for purposes of cardiac disease analysis. They are developing algorithms for segmentation, tracking and pattern analysis of cardiac MRI and CT images to distinguish between different cardiac diseases.

IBM Research is collaborating with seven universities on various software aspects of multicore computing. The following four projects started in 2008:

• IBM Research—China and Tsinghua University are investigating multicore programming and optimization to identify programming models with high productivity and advanced compilation and runtime systems with reasonable performance for data center oriented applications on next-generation multicore systems.
• The IBM Watson Research Center and the Georgia Institute of Technology are developing new ways of managing multicore resources and virtualizing them to address the needs of future server systems and applications.
• The IBM Watson Research Center and the University of California at Berkeley are developing a system for supporting field programmable gate arrays (FPGAs) and heterogeneous systems in a transparent manner. As part of this, they are developing a single unified programming language and environment that allows parts of the system to move fluidly between software and hardware.
• The IBM Watson Research Center and the University of Texas at Austin are conducting research on benchmarks for optimistic parallelism. The benchmarks will be made available to the public as open source software.

X10: Creating the next high-computing programming language (May 22, 2008) When will we see applications for multicore systems? (Jul 6, 2008) Bridging the software-hardware gap (Jan 6, 2009)

The IBM Watson Research Center, Carnegie Mellon University, the University of Massachusetts-Amherst, the University of Southern California, and the University of Texas at Austin are developing the Open Advancement of Question Answering architecture, data model and component technologies. They are also collaborating to conduct open, transparent evaluation of these technologies to facilitate the rapid advancement of the question answering field.

The IBM Watson Research Center, the University of Dundee School of Computing (UK), and the University of Miami Miller School of Medicine completed a 2-year collaboration in mid-2009 that focused on software tools and technologies that accommodate the needs of older workers. The collaboration's empirical findings have been widely disseminated in academic circles, including a special issue of the journal ACM Transactions on Accessible Computing devoted to "Aging and Information Technology". The OCR project also laid the foundation for a UK government "Digital Inclusion" grant to Dundee University. Based on the OCR findings, tools supporting Web access and development of more usable software are in the works at IBM.

What technologies are right for older workers? (Feb 6, 2008) How can technology help older workers? (Apr 11, 2008) Can we make software for older workers? (Dec 3, 2008)

IBM Research, India and the Indian School of Business have initiated a systematic study of Service Interaction Networks, developing models that take people-centric aspects of service delivery into account. While service interaction networks are similar to social networks, several features of service delivery necessitate novel techniques in representation and algorithmic approaches. Novel algorithms have been proposed and showed to be effective in this context. A simulator of service interaction networks has been developed that incorporates techniques for simulating scenarios without access to benchmark data sets.

Strong interaction networks improve service delivery (Oct 6, 2008)

IBM Research and Carnegie Mellon University will develop powerful analytics for genome-wide association studies in complex diseases; providing efficient and scalable implementations using BigData technologies; developing an infrastructure to setup a pipeline for the integrated execution of these algorithms together with other systems.

IBM Research is collaborating with seven universities on various software aspects of multicore computing. The following three projects started in 2007:

• The IBM Watson Research Center and Carnegie Mellon University are studying algorithmic scheduling techniques to determine whether they can be extended to support a richer class of language constructs, and whether the techniques can be efficiently implemented in schedulers and language runtime systems for a variety of multicore architectures.
• The IBM Watson Research Center and Rice University are performing research on translation, optimization, code generation and runtime technologies for parallel programs.
• The IBM Watson Research Center and the State University of New York at Oswego are developing a common set of efficient concurrency libraries that can be used by a large class of programmers across multiple languages such as X10, Java and C++, and multiple multicore symmetric multiprocessor architectures.

IBM Research, Haifa, the Technion, and Rambam Hospital are entering the third year of a project that brings together experts from the hospital and service science researchers from IBM and the Technion to apply measurement, modeling, analysis and optimization methods to the hospital's processes in order to significantly improve the effectiveness of patient care. In May of this year, the Service Research Innovation Institute (SRII) recognized this 3-way collaboration by naming the 3 institutions the first winner of the Services Partnership Award for the best collaborative practice between industry, academia, and government. The open dissemination of the results of this research is expected to broaden its impact well beyond the institutions involved.

IBM-Technion-Rambam optimize patient-centric care (Apr 11, 2008)

The IBM Watson Research Center and Columbia University developed technologies to provide intelligent data assistance for an important daily task in healthcare, the creation and updates of daily patient progress notes in hospital intensive care units. The technologies can also be generalized to provide intelligent assistance for information processing tasks in other domains and applications. Physicians at New York Presbyterian Hospital participated in the collaboration and helped evaluate the software that was developed.

Computer Science helps patients cope with chronic disease (Sep 6, 2007) Creating a medical note-taking system for physicians (Oct 2007) Computer Science monitors chronic diseases (Nov 28, 2007)

The IBM Watson Research Center, Carnegie Mellon University, and the University of California at Davis have developed algorithms and associated software tools for solving Mixed Integer Nonlinear Programming (MINLP) problems that significantly advance the size and scope of problems that can be solved using mathematical optimization software. The software has been released in the COIN-OR open source repository as a vehicle for sharing and accelerating advances in mathematical optimization. This work has enhanced IBM's position as a leader in the application of advanced computational mathematics in business optimization. In addition, the collaboration with CMU led to CMU receiving a $1.2M NSF award with IBM as a partner.

How IBM participates in open source agreements (Feb 27, 2008) Solving tough business problems with nonlinear techniques (May 29, 2008) Confronting nonlinearity with new mathematical tools (May 29, 2008)

The IBM Watson Research Center, Carnegie Mellon University, Imperial College, and Purdue University significantly advanced the notion of a policy management framework that applies to security and privacy policies. The framework incorporates policy authoring, transformation, refinement, analysis, deployment and compliance aspects. The work has been disseminated through extensive publications. Purdue has received two government grants that expand on some of the research started during the collaboration and has helped to extend the collaboration for a third year.

How we protect privacy and ensure security (May 29, 2008) What users need from policy systems (May 29, 2008) Collaboration helps enterprises manage privacy and security policies (May 29, 2008)

The IBM Watson Research Center, Rutgers University, and the University of California at Berkeley developed a variety of software tools and techniques for identifying and remediating code problems such as regression failures, performance issues, API usability issues, bugs (functional and non-functional), user-specified code pattern bugs, design problems, and security and compliance issues. There have been or will be over a dozen open source releases, including contributions to the T. J. Watson Libraries for Analysis (WALA) open source software project on SourceForge. The results of the research have also been widely disseminated at top programming languages and software engineering conferences.

How can we find the sources of bloat in applications? (May 29, 2008) How overuse of temporary objects hampers software performance (May 29, 2008) How open source helps developers debug software (May 29, 2008)