• Corpus ID: 239998455

TOD: Tensor-based Outlier Detection

  title={TOD: Tensor-based Outlier Detection},
  author={Yue Zhao and George H. Chen and Zhihao Jia},
To scale outlier detection (OD) to large, high-dimensional datasets, we propose TOD, a novel system that abstracts OD algorithms into basic tensor operations for efficient GPU acceleration. To make TOD highly efficient in both time and space, we leverage recent advances in deep learning infrastructure in both hardware and software. To deploy large OD applications on GPUs with limited memory, we introduce two key techniques. First, provable quantization accelerates OD computation and reduces the… 
PyGOD: A Python Library for Graph Outlier Detection
To overcome the scalability issue in large graphs, PyGOD provides advanced functionalities for selected models, including mini-batch and sampling, and is equipped with best practices to foster code reliability and maintainability.
ECOD: Unsupervised Outlier Detection Using Empirical Cumulative Distribution Functions
A novel outlier detection method called ECOD (Empirical-Cumulative-distribution-based Outlier Detection), which is both parameter-free and easy to interpret, and an easy-to-use and scalable Python implementation for accessibility and reproducibility is proposed.


GPU Strategies for Distance-Based Outlier Detection
A family of parallel and distributed algorithms for graphic processing units (GPU) derived from two distance-based outlier detection algorithms: BruteForce and SolvingSet are proposed, which differ in the way they exploit the architecture and memory hierarchy of the GPU and guarantee significant improvements with respect to the CPU versions.
Training Deeper Models by GPU Memory Optimization on TensorFlow
A general dataflow-graph based GPU memory optimization strategy, i.e.,“swap-out/in”, to utilize host memory as a bigger memory pool to overcome the limitation of GPU memory.
Accelerating SLIDE Deep Learning on Modern CPUs: Vectorization, Quantizations, Memory Optimizations, and More
This paper argues that SLIDE’s current implementation is sub-optimal and does not exploit several opportunities available in modern CPUs, and shows how SLIDE's computations allow for a unique possibility of vectorization via AVX (Advanced Vector Extensions)-512.
Estimating GPU memory consumption of deep learning models
DNNMem employs an analytic estimation approach to systematically calculate the memory consumption of both the computation graph and the DL framework runtime, and shows that DNNMem is effective in estimating GPU memory consumption.
TensorFlow: A system for large-scale machine learning
The TensorFlow dataflow model is described and the compelling performance that Tensor Flow achieves for several real-world applications is demonstrated.
Parallel processing for distance-based outlier detection on a multi-core CPU
A new parallelization model for the parallel processing of Orca-based outlier detection on a multi-core CPU that utilizes data parallelism and a Multi-thread model and outperforms conventional parallelization models.
On-the-fly Operation Batching in Dynamic Computation Graphs
This paper presents an algorithm, and its implementation in the DyNet toolkit, for automatically batching operations, and obtains throughput similar to that obtained with manual batches, as well as comparable speedups over single-instance learning on architectures that are impractical to batch manually.
Research Issues of Outlier Detection in Trajectory Streams Using GPUs
The problem of outlier Detection in trajectory streams is presented, and the research issues that should be addressed by new outlier detection techniques for trajectory streams on GPUs are discussed.
A Distributed Approach to Detect Outliers in Very Large Data Sets
While solving the distance-based outlier detection task in the distributed scenario, the method computes an outlier Detection solving set of the overall data set ofThe same quality as that computed by the corresponding centralized method.
VS-Quant: Per-vector Scaled Quantization for Accurate Low-Precision Neural Network Inference
Per-vector scaling consistently achieves better inference accuracy at low precision compared to conventional scaling techniques for popular neural networks without requiring retraining and modification of a deep learning accelerator hardware design to study the area and energy overheads.