Big Data Analytics: Is there reality behind the hype?
Abstract:
There is a lot of marketing & hype over big data and there is a general perception that we are at the peak of this hype cycle.
In this talk, I will attempt to peel away some of the layers behind this hype and describe some emerging big data use cases that provide an opportunity for real innovation, interesting science, and a variety of interesting research challenges. Using examples drawn from a number of industry verticals, I will show you how enterprises in retail, media & entertainment, financial services, telecom, healthcare, and energy, are applying large-scale analytics to a variety of new data types and data sources. In the process, I will highlight some of the research projects at IBM Research that are looking at platforms and analytic engines to support these emerging big data applications.
A Multi-label Collective Classification in a Multi-attribute Multi-relational Network Data: ML-M2AR
Abstract:
Classical Machine Learning techniques assume the data to be i.i.d, but the real world data is inherently relational and can generally be represented using graphs or some variants of them. The importance of modelling structured data is evident from its increasing presence: Telecom networks, WWW, social networks, organizational networks, images, protein sequences, etc. This field has recently been receiving a lot of attention in various communities under different themes depending on the problem addressed and the nature of solution proposed. Despite that, most of the works handle single attribute, single relational networks, and specifically single label classification tasks only. There is not much attention paid towards multi-attribute multi-relational (M2AR) networks under multi-label scenarios. In M2AR networks, nodes can be represented using multiple types of attributes and there are multiple link between the nodes.
For example, in Twitter, users can be represented using their tweets, urls shared, hash-tags and list memberships. And different Twitter users can be connected using followers, followed by and re-tweet networks. Motivated by the fact that this can be found in many networks, we propose learning and inference techniques, by extending the existing relational approaches, for multi-label classification using multi-attribute multi-relational collective classifier which captures both the inter/intra attribute and relation type label correlations on such networks.
Low Power Computing Systems Design Considerations for Big-Data
Abstract:
Power optimization is very crucial to increase power-to-performance and also to maintain thermal control in high-performance systems used in large data centers. The power consumed during testing a chip is significantly larger than the power consumed during normal functional operation. This talk presents details on techniques for power optimization when the system is working in normal functional mode as well as that during test mode. Interesting details on how delay and power vary with temperature and process parameters are discussed. The talk also presents details on how thermal and process awareness could be built into digital designs so as to maintain iso-performance across varying thermal profiles and process strengths.
Speaker Profile:
V. Kamakoti is a Professor at Department of Computer Science and Engineering, IIT Madras. His areas of interests include Computer Architecture, Secure Systems Engineering and VLSI Design. He completed BE in Computer Science and Engineering from University of Madras followed by MS (By research) and PhD at Department of Computer Science and Engineering, Indian Institute of Technology Madras. Prior to joining his alma mater as a faculty, he completed two post doctoral assignments at Supercomputer Education and Research Center, Indian Institute of Science Bengaluru and Institute of Mathematical Sciences, Chennai.
Scalable Graph Clustering via Sparsification
Abstract:
Many real world problems (biological, social, web) can be effectively modeled as networks or graphs where nodes represent entities of interest and edges mimic the interactions or relationships among them. The study of such complex relationship networks, recently referred to as "network science", can provide insight into their structure, properties and emergent behavior. Of particular interest here are rigorous methods for uncovering and understanding important network structures and motifs (communities) at multiple topological and temporal scales. Achieving this objective is challenging due to the presence of noise (false or missing interactions), topological(scale-free)) properties and scalability. Given the importance of the graph clustering problem, a number of solutions ranging from hierarchical methods to spectral methods have been designed and developed.
In this talk I will discuss a novel approach to sparsifying or sampling the edges of a graph while retaining the relevant content and structure important to motif and community detection. Empirical results demonstrate both qualitative as well as quantitive improvements over existing approaches on a wide range of datasets drawn from socio- technological- and biological- domains. Time permitting, I will also illustrate the value of such an approach from the perspective of visually teasing out relevant structure from large scale graphs and networks.
Stacked Temporal Relational Classifiers for Prediction in Attribute Evolving Networks
Abstract:
Classical Machine Learning techniques assume the data to be i.i.d, but the real world data is inherently relational and can generally be represented using graphs or some variants of them. The importance of modelling structured data is evident from its increasing presence :  Telecom networks, WWW, social networks, organizational networks, images, protein sequences, etc. This ï¬eld has recently been receiving a lot of attention in the community under different themes depending on the problem addressed and the nature of solution proposed. Despite that, most of the works handle static networks only, and there is not much attention paid towards temporally dynamic networks. In this work, we extend the existing relational approaches to attribute evolving networks. We propose learning and inference techniques for a stacked temporal relational classifier which captures both the temporal and relational dependencies on such networks. We evaluate the proposed method on alarm prediction task in Telecom Base Station Sub-system (BSS). Telecom BSS is responsible for handling traffic between a user equipment (UE) and network switching sub-system in a cellular telephone network.
Fraud Detection Using Data Analytics in the Finance Sector
Abstract:
Data analysis technology enables auditors and fraud examiners to analyze an organization's business data to gain insight into how well internal controls are operating and to identify transactions that indicate fraudulent activity or the heightened risk of fraud. Data analysis can be applied to just about anywhere in an organization where electronic transactions are recorded and stored. In this talk, I highlight instances of fraudulent transactions discovered solely through data analytics.
Speaker Profile:
Mr. N. Krishnan is a graduated from the College of Engineering, Trivandrum securing the First Rank . He is an authority in the the area of Instrumentation and Process Automation. He has held various positions in the industry. He was a Senior Director at CDAC-Trivandrum and later moved to NeSt as the Senior Vice President- New Initiatives. He was the Director of CERT, Kerala and the Director General of STPI and is the current CIO of Muthoot Fincorp Ltd and CEO of Muthoot Pappachan Technologies.
EU-India Fostering COOperation in Computing Systems (EUINCOOP) - Next Generation Computing Systems: Challenges and Opportunities.
Abstract:
There are number of projects active in Computing Systems research, working towards realizing key technologies across all computing segments: embedded, mobile, desktops, servers, datacentres, clouds and applications. Because of its overarching nature of computing systems, the challenges from different computing disciplines across computing system layers and across international market segment have to be identified. Hence, the objective of the session is to address the new approach for addressing the challenges ahead with international cooperation involving joint research activities.
One of the major challenges is to develop Software to support the advanced computing systems. Software development has not evolved as fast as hardware capability and network capacity. The session would address improvements and automation in software and Hardware development for the next generation computing systems addressing societal applications and challenges reinforcing industrial competitiveness enabled by advanced computing systems. The session will have invited experts discussing potential research topics for Euro-India joint research projects.
EURO-India Cooperation
Computing systems are universal. All aspects of public, private and commercial life are affected by computing systems. The dominance by desktops, laptops, and server PCs is waning and being replaced by smart embedded systems, mobile devices, and large-scale data centers. The European computing industry has a strong embedded ecosystem spanning from low power VLSI technologies to consumer products. However, the advanced computing systems performance depends on coordinated functionalities across Software and Hardware. EUINCOOP has studied this problem and has identified international cooperation in developing such a balanced system and getting access to International expertise in SW development to complement HW expertise available. Thus this session should be of high interest to Researchers in advance computing systems and industry partners of India.
Expected outcome
The traditional computing systems market is being replaced by a new market of smart embedded systems, mobile devices, and large-scale data centers, all converging to support global-scale applications that gather data from embedded systems and users, process it in large data centers, and control our environment or provide customized, timely information to millions of users through their mobile. With this background EUINCOOP project has produced research roadmap of advanced computing systems for the Horizon2020 framework. International cooperation through joint projects for the future competitive next generation computing systems will be discussed in the session, which can be of high value for research consortium to identify the partners with SW and HW expertise at the international level, and develop networks for mutual benefits. The session would identify key topics and challenges to be addressed for joint Euro-India research projects in the framework Horizon2020.
Big Data, Small Testing
Abstract: As evidenced by the theme of ADCOM 2013, Big Data has become the buzzword of choice in recent times, especially in the software industry. The acompanying hoopla has spawned frenetic claims foretelling the development of great and wondrous solutions to Big Data challenges. However, there is very little said about the testing of such systems, an essential pre-requisite for deployment. In this talk, we will discuss the research challenges involved in the testing process, especially from the database perspective. We will also present CODD, a graphical tool that takes a first step towards the effective testing of Big Data deployments through the metaphor of "data-less databases". CODD is currently in use at several industrial and academic institutions worldwide.
Automatic Mining of Association Rules from Text
Abstract: We describe the Extracting Association Rules from Text system for discovery of association rules from text. It discovers association’s -patterns of co-occurrence amongst keywords labeling the items in a collection of textual documents. It depends on keyword features for discovering association rules amongst keywords. System uses a query-centered view of knowledge discovery, in which a discovery request is viewed as a query over the implicit set of possible results supported by a collection of documents. The EART system treats texts only not images or figures. It consists of four phases: Text Collection Phase, Preprocessing Phase, Text Mining i.e. Rule Mining Phase and Visualization Phase. The System is domain independent; we can apply it on any domain. Here we apply it on MARKET-BASKET ANALYSIS domain and visualize results.
Out-of-Core Tridiagonal Reduction - Sraban Kumar Mohanty
Abstract: Reduction to tridiagonal form is a major performance bottleneck in the computation of the eigenvalues of a symmetric matrix; it takes O(N^3) flops. All the known blocked and unblocked direct tridiagonal reduction algorithms have an I/O complexity of O(N^3/B). To improve the I/O performance by incorporating matrix-matrix operations in the computation, usually the tridiagonal reduction is computed in two steps: the first reducing the matrix to a symmetric banded form, and the second further reducing it to tridiagonal form. We propose and analyse both these steps on the external memory model and also propose an improved algorithm that is amenable to multicore architecture. We also show that by suitably choosing the bandwidth for varying values of N and M (the size of the internal memory), the two steps of the reduction can be performed in O(N^3/\sqrt{M}B) I/Os for very large N and in O(N^3/\sqrt{N}B) I/Os for moderate N. Our algorithm has small constant factors. Hence it runs well in practice too.
Enhance (m,k)-firm constraint on the Real Time Streams Applied to AODV Protocol
Abstract: Mobile Ad hoc Networks (MANET) are wireless networks consisting of a collection of mobile nodes with no fixed infrastructure. Due to their decentralized, self-configuring and dynamic nature, MANETs offer many advantages and are easy to install. But many modern network applications, such as transmission of multimedia data require QoS which has raised a number of challenging technical issues for routing. To support multimedia applications, it is necessary for MANETs to have an efficient routing and QoS requirements. However, the rapid growth in number and diversity of real-time network applications has made it imperative to consider the impact of end-to-end delay requirements of traffic on network. In this article, we analyze and evaluate the performance of the network under different conditions. We study the conventional AODV protocol and its combination with the QoS, then we apply the scheduling policy using (m,k)-firm constraint in ad hoc on demand distance vector (AODV) to ensure graceful degradation of QoS for real-time streams. The results obtained show that the constraint (m,k)-firm applied to AODV protocol for real-time streams, performs better than conventional AODV.
Wide area measurement systems for electrical transmission networks
Abstract: Phasor measurement units (PMU) are used to transmit synchrophasors to electrical control centres at a rate which could be as high as 50 Samples per second. In this presentation, we will discuss analytics which are being planned to process this data. They include, vulnerability analysis of distance relays, estimation of line parameters, linear state estimation etc.
A WSN based electronic fence for human intrusion detection
Abstract: Of the various applications of Wireless Sensor Network (WSN), electronic surveillance for detecting human intrusion into an area or zone protected by an electronic fence is an important one that has attracted good amount of attention from researchers. One of the problems in building of WSN based electronic fence is the requirement for the sensing nodes to be randomly deployed within the area or zone encircled by the electronic fence. In such a scenario, the geographic locations of the sensing nodes have to be derived from either an onboard GPS receiver or from a set of beacon nodes whose geographic location is known a priori. The former option tends to be expensive while the later tends to be inaccurate and/or unreliable due to variability in the characteristics of the radio transmission media. In this paper we propose a scheme that provides a feasible and low cost solution for implementing an electronic fence where the sensing nodes are arbitrarily deployed in the preplanned subregions of the area or the zone encircled by electronic fence. The proposed scheme has been tested with sensing nodes comprising of TelosB motes and Passive Infrared (PIR) Sensors.
Building Watson - the Rise of Cognitive Computing
Abstract: A computer system that can directly and precisely answer natural language questions over an open and broad range of knowledge has been envisioned by scientists and writers since the advent of computers themselves. While current computers can store and deliver a wealth of digital content created by humans, they are unable to operate over it in human terms. The quest for building a computer system that can do open-domain Question Answering is ultimately driven by a broader vision that sees computers operating more effectively in human terms rather than strictly computer terms. They should function in ways that understand complex information requirements, as people would express them, for example, in natural language questions or interactive dialogs. Computers should deliver precise, meaningful responses, and synthesize, integrate, and rapidly reason over the breadth of human knowledge as it is most rapidly and naturally produced -- in natural language text.
The Watson effort from IRL is a grand challenge in Computer Science that aims to illustrate how the wide and growing accessibility of natural language content and the integration and advancement of Natural Language Processing, Information Retrieval, Machine Learning, Knowledge Representation and Reasoning, and massively parallel computation can drive open-domain automatic Question Answering technology to a point where it clearly and consistently rivals the best human performance. A first stop along the way is the Jeopardy!
Challenge, where we are planning to build an automated system that will compete with human grand champions in the game of Jeopardy!. In this talk, we will give an overview of the DeepQA project and the Jeopardy! Challenge.
Bio: Nanda Kambhatla is a Senior Technical Staff Member and the senior manager for the Knowledge Engineering Services department at IBM Research - India (aka IRL). This department is focused on technologies for acquiring, representing, retrieving and delivering knowledge to individuals and organizations and developing new products and offerings based on these technologies. Nanda leads a global research team that is driving innovations into Watson for IBM.
Nanda Kambhatla has over two decades of research experience in the areas of Natural Language Processing (NLP), text mining, information extraction, dialog systems, and machine learning. He holds 8 U.S patents and has authored over 40 publications in books, journals, and conferences in these areas. Nanda holds a B.Tech in Computer Science and Engineering from the Institute of Technology, Benaras Hindu University, India, and a Ph.D in Computer Science and Engineering from the Oregon Graduate Institute of Science & Technology, Oregon, USA.
Acceleration of Dense Numerical Linear Algebra Kernels
Abstract: Numerical Linear Algebra (NLA) Kernels play important role in applications ranging from wireless MIMO receiver to Kalman Filter (KF). For the matrix of size n × n these kernels exhibit time-complexity of O(n3). These kernels are difficult to realize on commercially available multi- processor systems. Acceleration of these kernels has been challenging task and has been actively pursued by the researchers. In our research we focus on the NLA kernels like Matrix Multiplication (MM), LU Decomposition (LUD), Cholesky Factorization (CF) and QR Decomposition (QRD). In MM, we focus on General MM (GEMM) and Strassen’s MM while in QRD we focus on QRD by Givens Rotation (GR) and QRD by Householder Transformation (HT). We derive Generalized GR (GGR) from GR that can exploit spatial and temporal parallelism available in the algorithm. We also extend our scheme to HT, Fast Givens and Square root and Division Free Givens (SDFG). We implement these kernels on REDEFINE, a massively parallel run-time reconfigurable multi- processor system, and show performance improvement over the state-of-the-art implementations. Our practical results corroborate to our theoretical improvements of the algorithm.
SPEAKER : FARHAD MERCHANT
Bio: Farhad is currently senior PhD student in CADLab, IISc under Porf. S K Nandy.
His area of interest are
- Numerical Linear Algebra Algorithms
- Reconfigurable architectures
- Massively Parallel systems
He was DAAD scholar from July 15, 2011 to December 15, 2011 at UMIC Research Cluster, RWTH Aachen University, Germany under Prof.Dr. -ing. Anupam Chattopadhyay.
Adaptive Asymmetric Multi-Core for Energy-Efficient High-Performance Computing
Abstract: Analysis of big data requires parallel computing with multi-cores and many cores. At the same time, power and thermal limits are driving the emergence of heterogeneous/asymmetric computing platforms consisting of cores with diverse power-performance characteristics enabling better match between the application workload and the compute engine leading to substantially improved energy-efficiency. In this talk, we will present the challenges and opportunities offered by static and adaptive asymmetric multi-cores towards low-power, high-performance computing.
For static asymmetric multi-cores, we will present a comprehensive power management framework that can provide high performance while minimizing energy consumption within the thermal design power budget. We will then describe an adaptive asymmetric multi-core architecture, called Bahurupi, that can be tailored according to the application by software. Bahurupi is designed and fabricated as a homogeneous multi-core system containing identical (simple) cores. Post-fabrication, software can configure or compose together primitive cores to create an asymmetric multi-core system that best matches the needs of the currently executing application.
SPEAKER : Thannirmalai Somu Muthukaruppan
Bio: Mr. Malai is currently a PhD candidate in the Department of Computing, School of Computing at National University of Singapore. His research interests include heterogeneous systems, system-level power managements and lifetime-reliability aware computing. He is part of the Emebedded Computing lab under Associate Professor Tulika Mitra
Cognitronics: Resource-efficient Architectures for Cognitive Systems
Abstract: Cognitronics deals with the development of integrated circuits for resource-efficient realisation of cognitive systems. In this talk we will discuss our approach toadaptive computing architectures based on Field Programmable Gate Arrays (FPGAs) and many-core systems. Currently available FPGAs can integrate complex systems and are partially reconfigurable during runtime, which make them suitablefor general purpose reconfigurable computing.Arbitrary functions can be implemented in hardware and can be downloaded to the FPGA as an exchangeable hardware task which leads us to the concept of virtual hardware. These features make FPGAs an attractive alternative for system integration as can be seen by the increasing number of embedded FPGAs in all kind of stationary as well mobile technical systems.
The recent switch to parallel microprocessors is considered as a milestone in computer engineering.Industry has laid out a roadmap for multi-core designs promising to double the number of cores on a chip with each silicon generation. This new shift toward increasing parallelism is not a result based on breakthroughs in novel software or architectures for parallelism; instead, this shift into parallelism is actually a technology push based on the progress in nanoelectronics offering 1000 cores on a chip in the near future.
SPEAKER : Prof. Ulrich Rueckert
Bio: Ulrich Rueckert received the Diplomadegree in Computer Science and aDr.-Ing. degree in Electrical Engineering from the University of Dortmund, Germany,in 1984 and 1989, respectively. From 1985 to 1994 he worked onmicroelectronic implementation of neuralnetworks at the Faculty of Electrical Engineering (University of Dortmund) and at the Technical University of Hamburg-Harburg, Germany. In 1995 he joined as a Full Professor the Heinz Nixdorf Institute at the University of Paderborn, Germany, heading the research group ‘‘System and Circuit Technology’’ and working on microelectronic systems for massive-parallel and resource-efficient information processing. Since 2009 he is Professor at Bielefeld University, Germany heading the “Cognitronics and Sensor Systems” group of the “Cluster of Excellence - Cognitive Interaction Technology”. His main research interests are now bio-inspired architectures for nanotechnologies and cognitive robo.
Adapting Precision of Computation to match Dynamic Range and Uncertainty
Abstract: Double-precision computations operating on inputs with uncertainty margins can be compiled to lower precision fixed-point datapaths with no loss in output accuracy. We observe that ideal SPICE model equations based on device physics include process parameters which must be matched with real-world measurements on specific silicon manufacturing processes through a noisy data-fitting process. We expose this uncertainty information to the open-source FX-SCORE compiler to enable automated error analysis using the Gappa++ backend and hardware circuit generation using Vivado HLS. We construct an error model based on interval analysis to statically identify sufficient fixedpoint precision in the presence of uncertainty as compared to reference double-precision design. We demonstrate 1–16 LUT count improvements, 0.5–2.4 DSP count reductions and 0.9–4 FPGA power reduction for SPICE devices such as Diode, Level-1 MOSFET and an Approximate MOSFET designs. We generate confidence in our approach using Monte-Carlo simulations with auto-generated Matlab models of the SPICE device equations.
SPEAKER : Prof. NACHIKET KAPRE,
Bio: Prof. Nachiket Kapre is an Assistant Professor at NTU, Singapore in the School of Computer Engineering since October 2012. He received his PhD from Caltech in 2010 and was an Imperial College Junior Research Fellow until 2012. He is interested in understanding the limits of modern VLSI processing systems and exploiting them to its limit. He broader research interests include parallel processing, computer architecture and domain-specific frameworks
A WSN based electronic fence for human intrusion detection
Abstract: Of the various applications of Wireless Sensor Network (WSN), electronic surveillance for detecting human intrusion into an area or zone protected by an electronic fence is an important one that has attracted good amount of attention from researchers. One of the problems in building of WSN based electronic fence is the requirement for the sensing nodes to be randomly deployed within the area or zone encircled by the electronic fence. In such a scenario, the geographic locations of the sensing nodes have to be derived from either an onboard GPS receiver or from a set of beacon nodes whose geographic location is known a priori. The former option tends to be expensive while the later tends to be inaccurate and/or unreliable due to variability in the characteristics of the radio transmission media. In this paper we propose a scheme that provides a feasible and low cost solution for implementing an electronic fence where the sensing nodes are arbitrarily deployed in the preplanned subregions of the area or the zone encircled by electronic fence. The proposed scheme has been tested with sensing nodes comprising of TelosB motes and Passive Infrared (PIR) Sensors.
Dr. Prasenjit Dey
Bio: Prasenjit Dey is a Senior Researcher at IBM Research where he works in the area of mobile HCI, mobile education, and image analysis & pattern recognition. Prior to joining IBM Research, Prasenjit was a Senior Research Scientist at Hewlett-Packard Labs working on application of crowdsourcing and machine learning to solve hard analytics problems in areas such as retail and social media. Earlier he has worked extensively in the area of multi-modal interactions using facial analysis, gesture recognition, and speech analysis to create richer media and entertainment experience in HP's personal system products. He received his Ph.D. (2004) in computer & communications engineering from Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland, and is a member of ACM and IEEE. He has 4 granted and over 25 pending patents, and has authored or co-authored over 50 publications in peer-reviewed conferences and journals.
Big Data Security, Privacy, and Veracity Aspects in the E&U Industry
Abstract: This talk will focus on privacy and security aspects in the context of the Energy & Utility (E&U) industry from a Big Data and information governance perspective, highlighting specific E&U industry imperatives and corresponding use case scenarios. The IBM security framework will be taken as a foundation to discuss data-centric aspects of cyber security and real-time prevention. We will detail out the relevance of information governance capabilities to Big Data, and how to address security, privacy but also trustworthiness and veracity in a Big Data E&U context.
Dr. Prakash Ranganathan
Bio: Dr. Prakash Ranganathan is an Assistant Professor in Department of Electrical Engineering at the University of North Dakota (UND), Grand Forks, North Dakota, USA. His area of research interests includes Software Engineering, Cyber security, Smart grid and Wireless sensor networks.
His research group is currently working on projects that improve real-time situational awareness and decision support tools that enhances system reliability in the area of Synchrophasor’s technology. His group is particularly interested in the development of analytical and computational tools for the modeling, analysis, design and optimization of large-scale integrated systems such as the smart grid.
He has recently led an interdisciplinary team of researchers between North Dakota State University (NDSU), UND, and South Dakota State University (SDSU) in putting together multiple proposals on Building a Research Infrastructure in Cyber Security (BRICS) in North Dakota. He earned his Ph.D. in Software Engineering and M.S. in Electrical Engineering from North Dakota State University, Fargo, ND, USA. Dr. Ranganathan is an active senior member of IEEE and has over 25 publications in peer-reviewed conferences and journals. Ranganathan was recently awarded the North Dakota Spirit Faculty Achievement Award for the academic year 2013. The award recognizes significant contributions by faculty in teaching, research and service funded by the UND Foundation.
Dr. Rajeswari Mukesh
Bio: Dr. Rajeswari Mukesh is a Professor in the Department of Computer Science at Hindustan Uminversity. Her current Research focus includes Network Security and Big Data
Dr. V. Kamakoti
Bio: V. Kamakoti is a Professor at Department of Computer Science and Engineering, IIT Madras. His areas of interests include Computer Architecture, Secure Systems Engineering and VLSI Design. He completed BE in Computer Science and Engineering from University of Madras followed by MS (By research) and PhD at Department of Computer Science and Engineering, Indian Institute of Technology Madras. Prior to joining his alma mater as a faculty, he completed two post doctoral assignments at Supercomputer Education and Research Center, Indian Institute of Science Bengaluru and Institute of Mathematical Sciences, Chennai.
G.S. Madhusudan
Bio: G. S. Madhusudan is a Technologist/Researcher with 26+ years of global experience in Security Engineering, Systems Engineering, Consumer Electronics and Enterprise Systems. He is Currently a Researcher at IIT-Madras and leads research projects in Security Engineering (OS, Processors and Networking), Processor Design and Storage Systems. He was formerly the CEO of Pravrtti Ventures (Singapore), a technology company involved in incubating startups in Singapore/USA in the areas of Consumer Electronics, Security devices and Smart Phone Design. Prior to that, he was a Staff Engineer at Sybase and headed projects in Secure RDBMS, large cluster systems and micro-kernel architectures.
Patrons
ADCOM 2013 Programme
21st October 2013 Day 1 - Tutorials
22nd October 2013 Day 2 - Inaugural Session
Time |
Programme & Speaker |
09:00 AM |
Registration |
09:30 AM |
Welcome Address Saragur Srinidhi, President, ACCS |
09:45 AM |
Conference Chairs Prof. H. S. Jamadagni / Dr. Raghu Krishnapuram, General Chairs - ADCOM 2013, Dr. S. Ramachandran, VC Hindustan University, Chennai |
10:00 AM |
Confer ACCS - CDAC Award Dr. Jayant Haritsa Senior Professor, Department of Computer Science and Automation, Indian Institute of Science, Bangalore
ACCS-CDAC Foundation Award Winner for 2013 |
10:15 AM |
Conference Overview Dr. Balaraman Ravindran Associate Professor of Computer Science and Engineering, Indian Institute of Technology Madras Technical Program Co-Chair ADCOM 2013 |
10:30 AM |
Big Data Analytics: Is there reality behind the hype? (read abstract) Dr. Prasad M. Deshpande, IBM Research |
11:15 AM |
Tea Break |
11:30 AM |
ACCS Keynote - Big Data Security, Privacy, and Veracity Aspects in the E&U Industry (read abstract) Dr. Eberhard Hechler Executive Architect, IBM Singapore |
12:15 PM |
Distinguished Lecture - Big Data, Small Testing? (read abstract) ACCS-CDAC Foundation Awardee Dr. Jayant Haritsa, SERC, Indian Institute of Science, Bangalore |
01:00 PM |
Lunch |
02:00 PM |
IEEE CS Keynote - Networking for Big Data Dr. Raj Jain, WUSL Professor of Computer and Information Science at Washington University in St. Louis |
02:50 PM |
Automatic Mining of Association Rules from Text (read abstract) Vaishali Bhujade, Chhaya Meshram [Bapurao Deshmukh College of Engineering, Sevagram] |
03:15 PM |
Out-of-Core Tridiagonal Reduction (read abstract) Sraban Kumar Mohanty [IIIT, Jabalpur]; G. Sajith [IIT Guwahati] |
03:40 PM |
Large Scale Multi-view Learning on Hadoop Cibe Hariharan, S. Shivashankar [Ericsson India] |
04:05 PM |
Tea Break |
04:30 PM |
Dr. A. Ramesh, IITM Department of Mechanical Engineering, IIT Madras |
05:15 PM |
Close |
23rd October 2013 Day 3
Time |
Programme & Speaker |
09:00 AM |
Scalable Graph Clustering via Sparsification (read abstract) Dr. Srinivasan Parthasarathy, Professor of Computer Science and Engineering and Department of Biomedical Informatics, The Ohio State University, Columbus, Ohio |
09:45 AM |
A Human Intrusion Detection System using WSN with arbitrarily deployed nodes (read abstract) Chaitra M, Reema Mathew, Nirdesh Singh and N. Rama Murthy [Center for Artificial Intelligence and Research] |
10:10 AM |
Enhance (m,k)-firm constraint on the Real Time Streams Applied to AODV Protocol (read abstract) Mohamed Tekaya [Sup Com, Tunisia ] |
10:35 AM |
Wide area measurement systems for electrical transmission networks (read abstract) Dr. Soman Shreevardhan [IIT Bombay] |
11:00 AM |
Tea Break |
11:20 AM |
Building Watson - the Rise of Cognitive Computing (read abstract) Dr. Nanda Kambhatla Senior Manager -Knowledge Engineering Department, IBM India Research Labs |
12:00 PM |
Industry Case Study Stacked Temporal Relational Classifiers for Prediction in Attribute Evolving Networks (read abstract) S. Shivashankar, Ericsson R&D |
12:30 PM |
Industry Case Study Fraud Detection Using Data Analytics in the Finance Sector (read abstract) N. Krishnan, CIO, Muthoot Fincorp |
01:00 PM |
Lunch |
02:00 PM |
Project Pakshi - A model for Open Source community for crowd sensing Dr. H. S. Jamadagni Professor, Department of Electronic Systems Engineering, Indian Institute of Science, Bangalore |
03:00 PM |
Low Power Computing Systems Design Considerations for Big-Data (read abstract) Dr. V. Kamakoti, IIT Madras Department of Computer Science and Engineering, IIT Madras |
03:30 PM |
Tea Break |
03:50 PM |
Special Session EU-India Fostering COOperation in Computing Systems (EUINCOOP) - Next Generation Computing Systems: Challenges and Opportunities (read abstract) [Chair - Dr. H. S. Jamadagni, IISc, Dr. Sathya Rao, KYOS] |
05:30 PM |
Close |
24th October 2013 Day 4 - Adaptive Computing Workshop
09:30 AM |
Acceleration of Dense Numerical Linear Algebra Kernels (read abstract) Mr. FARHAD MERCHANT, CAD Laboratory, Indian Institute of Science |
10:15 AM |
Adaptive Asymmetric Multi-Core for Energy-Efficient High-Performance Computing (read abstract) Mr. Thannirmalai Somu Muthukaruppan, Dept. of Computer Science, National University of Singapore |
11:00 AM |
Cognitronics: Resource-efficient Architectures for Cognitive Systems (read abstract) Prof. Ulrich Rueckert, Cognitronics and Sensor Systems Group, Bielefeld University |
11:45 AM |
Tea Break |
12:00 PM |
Adapting Precision of Computation to match Dynamic Range and Uncertainty (read abstract) Prof. NACHIKET KAPRE, School of Computer Engineering, NTU Singapore |
12:45 PM |
Prof. VIRENDRA SINGH, Department of Electrical Engineering, Indian Institute of Technology Bombay |
01:30 PM |
Lunch |
02:30 PM |
Adaptive Computing Panel Session Panel Members
Dr. Prasenjit Dey, Senior Research Scientist, IBM Research
Dr. Santhosh Kumar, Strand Lifesciences
Dr. Prakash Ranganathan, Assistant Professor, Dept of Electrical Engineering at the University of North Dakota
Dr. Rajeswari Mukesh, Professor, Department of CSE, Hindustan University.
Prof. Kamakoti, IIT Madras
GS Madhusudan, IIT Madras
|
04:30 PM |
Close |