Search
Now showing items 31-40 of 51
A Divide and Conquer Framework For Graph Processing in Distributed Heterogeneous Systems
In many fields of science and engineering graph data structures are used to represent real-world
information. As these graphs scale in size, it becomes very inefficient to process these graphs on
a single core CPU using ...
EMF: System Design and Challenges for Disaggregated GPUs in datacenters for Efficiency, Modularity and Flexibility
With Dennard Scaling phasing out in the mid-2000s, architectural scaling and hardware specialization
take centre stage to provide performance bene fits with already stalling Moore's law. An
outcome from this hardware ...
Deep Learning for Hand-drawn Sketches: Analysis, Synthesis and Cognitive Process Models
Deep Learning-based object category understanding is an important and active area of research
in Computer Vision. Most work in this area has predominantly focused on the portion of
depiction spectrum consisting of ...
Stabilized finite element schemes for computations of viscoelastic free-surface and two-phase flows
Viscoelastic flows can be found in a wide range of industrial and commercial applications
such as enhanced oil recovery, pesticide deposition, medicinal/pharmaceutical sprays, drug
delivery, injection molding, polymer ...
Compiler controlled Task Management in Runtime Systems for Dynamic Data ow Model of Execution
For the past 40 years, relentless focus on Moore's Law transistor scaling has provided ever-increasing transistor performance and density. An ever increasing demand for large scale parallelism has driven hardware designers ...
Static Analysis and Dynamic Monitoring of Program Flow on REDEFINE Manycore Processor
Manycore heterogeneous architectures are becoming the promising choice for high-performance computing applications. Multiple parallel tasks run concurrently across different processor cores sharing the same communication ...
Methods for Improving Data-efficiency and Trustworthiness using Natural Language Supervision
Traditional strategies to build machine learning based classification systems employ discrete labels as targets. This limits the usefulness of such systems in two ways. First, the generalizability of these systems is limited ...
Minimizing latency in data acquisition, distributed processing, storage and retrieval
Achieving low latency is of utmost importance in applications demanding real-time sensing
and control as in cyber-physical systems. In this thesis, we explore three different facets of
ensuring low latency in such systems. ...
Novel Neural Architectures based on Recurrent Connections and Symmetric Filters for Visual Processing
Artificial Neural Networks (ANN) have been very successful due to their ability to extract meaningful information without any need for pre-processing raw data. First artificial neural networks were created in essence to ...
Assessing protein contribution to phenotypic change using short, coarse grained molecular dynamics simulations
Understanding the functional mapping between genotype and phenotype is an important problem that has ramifications for various diseases. Various existing computational methods can infer these disease-related functional ...