English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Outlier detection (OD) is a key learning task for finding rare and deviantdata samples, with many time-critical applications such as fraud detection andintrusion detection. In this work, we propose TOD, the first tensor-basedsystem for efficient and scalable outlier detection on distributed multi-GPUmachines. A key idea behind TOD is decomposing complex OD applications into asmall collection of basic tensor algebra operators. This decomposition enablesTOD to accelerate OD computations by leveraging recent advances in deeplearning infrastructure in both hardware and software. Moreover, to deploymemory-intensive OD applications on modern GPUs with limited on-device memory,we introduce two key techniques. First, provable quantization speeds up ODcomputations and reduces its memory footprint by automatically performingspecific floating-point operations in lower precision while provablyguaranteeing no accuracy loss. Second, to exploit the aggregated computeresources and memory capacity of multiple GPUs, we introduce automaticbatching, which decomposes OD computations into small batches for bothsequential execution on a single GPU and parallel execution on multiple GPUs.  TOD supports a diverse set of OD algorithms. Extensive evaluation on 11 realand 3 synthetic OD datasets shows that TOD is on average 10.9x faster than theleading CPU-based OD system PyOD (with a maximum speedup of 38.9x), and canhandle much larger datasets than existing GPU-based OD systems. In addition,TOD allows easy integration of new OD operators, enabling fast prototyping ofemerging and yet-to-discovered OD algorithms.

TOD: GPU-accelerated Outlier Detection via Tensor Operations