Purine: A bi-graph based deep learning framework
In this paper, we introduce a novel deep learning framework, termed Purine. In Purine, a deep network is expressed as a bipartite graph (bi-graph), which is composed of interconnected operators and data tensors. With the bi-graph abstraction, networks are easily solvable with event-driven task dispatcher. We then demonstrate that different parallelism schemes over GPUs and/or CPUs on single or multiple PCs can be universally implemented by graph composition. This eases researchers from coding for various parallelization schemes, and the same dispatcher can be used for solving variant graphs. Scheduled by the task dispatcher, memory transfers are fully overlapped with other computations, which greatly reduce the communication overhead and help us achieve approximate linear acceleration.
Code (1)
Tasks
Deep LearningSimilar Papers 제목 키워드 기반
ODE and Random Boolean networks in application to modelling of 6-mercaptopurine metabolism
We consider two approaches to modelling of cell metabolism of 6-mercaptopurine, which is one of the important chemotherapy drugs used for treating of acute lymphocytic leukemia: kinetic ordinary differential equations an…
Possible traces of resonance signaling in the genome
Although theories regarding the role of sequence-specific DNA resonance in biology have abounded for over 40 years, the published evidence for it is lacking. Here, the authors reasoned that for sustained resonance signal…
Ultra-rapid, Quantitative, Label-free Antibiotic Susceptibility Testing via Optically Detected Purine Metabolites
There is an urgent need for the development of novel and truly rapid (equal or less than 1 hour) antibiotic susceptibility testing (AST) platforms in order to provide best antimicrobial prescribing practices and to help …
Development of Hydrogen Bonding Magnetic Reaction-based Gene Regulation through Cyclic Electromagnetic DNA Simulation in Double-Stranded DNA
The proton-magnetic reaction is commonly used in MRI machines with a strong magnetic field of over 1 T, while this study hypothesized that the electron magnetic reaction of hydrogen could affect the hydrogen bonds of dou…
The Electronic Behavior of Zinc-Finger Protein Binding Sites in the Context of the DNA Extended Ladder Model
The eukaryotic Cys2His2 zinc finger proteins bind to DNA ubiquitously at highly conserved domains, responsible for gene regulation and the spatial organization of DNA. To study and understand the zinc finger DNA-protein …
Specificity