Proceedings of the 1st on Reproducible Quality-Efficient Systems Tournament on Co-designing Pareto-efficient Deep Learning / Luis Ceze.

Format:

Book

Author/Creator:

Ceze, Luis, author.

Series:

ACM Conferences

Language:

English

Subjects (All):

Interactive computer systems--Congresses.

Interactive computer systems.

Physical Description:

1 online resource

Other Title:

ReQuEST '18

Place of Publication:

New York, NY, USA : ACM, 2018.

Notes:

Artificial Intelligence (AI), Machine Learning (ML) and other emerging workloads demand efficient computer systems from the cloud to the edge. Systems designers, however, face numerous challenges from tackling the ever-growing space of design and optimization choices (including algorithms, models, software frameworks, libraries, hardware platforms, optimization techniques) to balancing off multiple objectives (including accuracy, speed, throughput, power, size, price). Furthermore, the lack of a common experimental framework and methodology makes it even more challenging to keep up with and build upon the latest research advances. The Reproducibly Quality-Efficient Systems Tournaments (ReQuEST) initiative is a community effort to develop a rigorous methodology, open platform and online scoreboard for co-designing the efficient and reliable software/hardware stack for emerging workloads. ReQuEST invites a multidisciplinary community to collaborate on benchmarking and optimizing workloads across diverse platforms, models, data sets, libraries and tools, while gradually adopting best practice. The community effectively creates a "marketplace" for trading Pareto-efficient implementations (code and data) as portable, customizable and reusable Collective Knowledge workflows and packages. We envision that such a community-driven and decentralized marketplace will help accelerate adoption and technology transfer of novel AI/ML techniques similar to the open-source movement. Please see the front matter for the 1st ReQuEST tournament on Co-designing Pareto-efficient Deep Learning Inference at ASPLOS'18 to learn more about the shared workflows and validated results, as well as about our next steps for the ReQuEST initiative.