Machine Learning (ML) and Deep Learning (DL) innovations are being introduced at such a rapid pace that researchers are hard-pressed to analyze and study them. The complicated procedures for evaluating innovations, along with the lack of standard and efficient ways of specifying and provisioning ML/DL evaluation, is a major "pain point" for the community. This paper proposes MLModelScope, an open-source, framework/hardware agnostic, extensible and customizable design that enables repeatable, fair, and scalable model evaluation and benchmarking. We implement the distributed design with support for all major frameworks and hardware, and equip it with web, command-line, and library interfaces. To demonstrate MLModelScope's capabilities we perform parallel evaluation and show how subtle changes to model evaluation pipeline affects the accuracy and HW/SW stack choices affect performance.
University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；IBM T. J. Watson Research Center（IBM托马斯·J·沃森研究中心）
The current Deep Learning (DL) landscape is fast-paced and is rife with non-uniform models, hardware/software (HW/SW) stacks, but lacks a DL benchmarking platform to facilitate evaluation and comparison of DL innovations, be it models, frameworks, libraries, or hardware. Due to the lack of a benchmarking platform, the current practice of evaluating the benefits of proposed DL innovations is both arduous and error-prone - stifling the adoption of the innovations. In this work, we first identify $10$ design features which are desirable within a DL benchmarking platform. These features include: p…
Paradigm（第四范式）；Universitat de Barcelona（巴塞罗那大学）；Computer Vision Center（计算机视觉中心）；LISN/CNRS/INRIA, Univ. Paris-Saclay（巴黎萨克雷大学LISN实验室/法国国家科学研究中心/法国国家信息与自动化研究所）；ChaLearn（ChaLearn挑战赛组织）；Univ. Grenoble Alpes/CNRS（格勒诺布尔阿尔卑斯大学/法国国家科学研究中心）；Tsinghua University（清华大学）
Obtaining standardized crowdsourced benchmark of computational methods is a major issue in data science communities. Dedicated frameworks enabling fair benchmarking in a unified environment are yet to be developed. Here we introduce Codabench, an open-source, community-driven platform for benchmarking algorithms or software agents versus datasets or tasks. A public instance of Codabench (https://www.codabench.org/) is open to everyone, free of charge, and allows benchmark organizers to compare fairly submissions, under the same setting (software, hardware, data, algorithms), with custom protoc…
Training deep learning models is compute-intensive and there is an industry-wide trend towards hardware specialization to improve performance. To systematically benchmark deep learning platforms, we introduce ParaDnn, a parameterized benchmark suite for deep learning that generates end-to-end models for fully connected (FC), convolutional (CNN), and recurrent (RNN) neural networks. Along with six real-world models, we benchmark Google's Cloud TPU v2/v3, NVIDIA's V100 GPU, and an Intel Skylake CPU platform. We take a deep dive into TPU architecture, reveal its bottlenecks, and highlight valuabl…
We present CodeReef - an open platform to share all the components necessary to enable cross-platform MLOps (MLSysOps), i.e. automating the deployment of ML models across diverse systems in the most efficient way. We also introduce the CodeReef solution - a way to package and share models as non-virtualized, portable, customizable and reproducible archive files. Such ML packages include JSON meta description of models with all dependencies, Python APIs, CLI actions and portable workflows necessary to automatically build, benchmark, test and customize models across diverse platforms, AI framewo…
Deep learning has been shown as a successful machine learning method for a variety of tasks, and its popularity results in numerous open-source deep learning software tools. Training a deep network is usually a very time-consuming process. To address the computational challenge in deep learning, many tools exploit hardware features such as multi-core CPUs and many-core GPUs to shorten the training time. However, different tools exhibit different features and running performance when training different types of deep networks on different hardware platforms, which makes it difficult for end user…
Shanghai Jiao Tong University（上海交通大学）；University College London（伦敦大学学院）
We introduce Texygen, a benchmarking platform to support research on open-domain text generation models. Texygen has not only implemented a majority of text generation models, but also covered a set of metrics that evaluate the diversity, the quality and the consistency of the generated texts. The Texygen platform could help standardize the research on text generation and facilitate the sharing of fine-tuned open-source implementations among researchers for their work. As a consequence, this would help in improving the reproductivity and reliability of future research work in text generation.
Insilico Medicine Hong Kong Ltd；Chemistry and Chemical Biology Department, Harvard University（哈佛大学化学与化学生物学系）；Neuromation OU；Hit discovery, Discovery Sciences, Biopharmaceutics R&D, AstraZeneca Gothenburg（阿斯利康哥德堡公司发现科学部药物研发中心）；National Research University Higher School of Economics（俄罗斯国家研究型高等经济大学）；Department of Chemistry and Department of Computer Science, University of Toronto（多伦多大学化学系与计算机科学系）；National Research University Higher School of Economics, St. Petersburg（俄罗斯国家研究型高等经济大学圣彼得堡分校）
Generative models are becoming a tool of choice for exploring the molecular space. These models learn on a large training dataset and produce novel molecular structures with similar properties. Generated structures can be utilized for virtual screening or training semi-supervised predictive models in the downstream tasks. While there are plenty of generative models, it is unclear how to compare and rank them. In this work, we introduce a benchmarking platform called Molecular Sets (MOSES) to standardize training and comparison of molecular generative models. MOSES provides a training and testi…
Imperial College London（伦敦帝国理工学院）；Ryerson University（瑞尔森大学）
Path planning is a key component in mobile robotics. A wide range of path planning algorithms exist, but few attempts have been made to benchmark the algorithms holistically or unify their interface. Moreover, with the recent advances in deep neural networks, there is an urgent need to facilitate the development and benchmarking of such learning-based planning algorithms. This paper presents PathBench, a platform for developing, visualizing, training, testing, and benchmarking of existing and future, classical and learned 2D and 3D path planning algorithms, while offering support for Robot Ope…
Recent advances in state-of-the-art ultra-low power embedded devices for machine learning (ML) have permitted a new class of products whose key features enable ML capabilities on microcontrollers with less than 1 mW power consumption (TinyML). TinyML provides a unique solution by aggregating and analyzing data at the edge on low-power embedded devices. However, we have only recently been able to run ML on microcontrollers, and the field is still in its infancy, which means that hardware, software, and research are changing extremely rapidly. Consequently, many TinyML frameworks have been devel…
Southeast Missouri State University（东南密苏里州立大学）；University of Arkansas at Little Rock（阿肯色大学小石城分校）
This paper surveys benchmarking principles, machine learning devices including GPUs, FPGAs, and ASICs, and deep learning software frameworks. It also reviews these technologies with respect to benchmarking from the perspectives of a 6-metric approach to frameworks and an 11-metric approach to hardware platforms. Because MLPerf is a benchmark organization working with industry and academia, and offering deep learning benchmarks that evaluate training and inference on deep learning hardware devices, the survey also mentions MLPerf benchmark results, benchmark metrics, datasets, deep learning fra…
Geoparsing is an important task in geographic information retrieval. A geoparsing system, known as a geoparser, takes some texts as the input and outputs the recognized place mentions and their location coordinates. In June 2019, a geoparsing competition, Toponym Resolution in Scientific Papers, was held as one of the SemEval 2019 tasks. The winning teams developed neural network based geoparsers that achieved outstanding performances (over 90% precision, recall, and F1 score for toponym recognition). This exciting result brings the question "are we there yet?", namely have we achieved high en…
Learned Indexes are a novel approach to search in a sorted table. A model is used to predict an interval in which to search into and a Binary Search routine is used to finalize the search. They are quite effective. For the final stage, usually, the lower_bound routine of the Standard C++ library is used, although this is more of a natural choice rather than a requirement. However, recent studies, that do not use Machine Learning predictions, indicate that other implementations of Binary Search or variants, namely k-ary Search, are better suited to take advantage of the features offered by mode…
Benchmarking for Deep Uplift Modeling in Online Marketing
Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ)（广东省人工智能与数字经济实验室（深圳））；Tencent（腾讯）；Shenzhen University（深圳大学）；Renmin University of China（中国人民大学）；Shenzhen Technology University（深圳技术大学）
School of Technology Management and Engineering, NMIMS（NMIMS技术管理与工程学院）；Indian Institute of Technology, Roorkee（鲁尔基印度理工学院）；Birla Institute of Technology and Science, Pilani (BITS Pilani)（皮拉尼贝拉理工学院（BITS Pilani））
Standardized evaluation measures have aided in the progress of machine learning approaches in disciplines such as computer vision and machine translation. In this paper, we make the case that robotic learning would also benefit from benchmarking, and present the "REPLAB" platform for benchmarking vision-based manipulation tasks. REPLAB is a reproducible and self-contained hardware stack (robot arm, camera, and workspace) that costs about 2000 USD, occupies a cuboid of size 70x40x60 cm, and permits full assembly within a few hours. Through this low-cost, compact design, REPLAB aims to drive wid…
Facebook AI Research（脸书人工智能研究院）；Carnegie Mellon University（卡内基梅隆大学）
This paper introduces PyRobot, an open-source robotics framework for research and benchmarking. PyRobot is a light-weight, high-level interface on top of ROS that provides a consistent set of hardware independent mid-level APIs to control different robots. PyRobot abstracts away details about low-level controllers and inter-process communication, and allows non-robotics researchers (ML, CV researchers) to focus on building high-level AI applications. PyRobot aims to provide a research ecosystem with convenient access to robotics datasets, algorithm implementations and models that can be used t…
This work presents the Video Platform for PyTorch (ViP), a deep learning-based framework designed to handle and extend to any problem domain based on videos. ViP supports (1) a single unified interface applicable to all video problem domains, (2) quick prototyping of video models, (3) executing large-batch operations with reduced memory consumption, and (4) easy and reproducible experimental setups. ViP's core functionality is built with flexibility and modularity in mind to allow for smooth data flow between different parts of the platform and benchmarking against existing methods. In providi…
Following the increasing interest and adoption of FaaS systems, benchmarking frameworks for determining non-functional properties have also emerged. While existing (microbenchmark) frameworks only evaluate single aspects of FaaS platforms, a more holistic, application-driven approach is still missing. In this paper, we design and present BeFaaS, an extensible application-centric benchmarking framework for FaaS environments that focuses on the evaluation of FaaS platforms through realistic and typical examples of FaaS applications. BeFaaS includes a built-in e-commerce benchmark, is extensible …
Aadhaar is the world's largest biometric database with a billion records, being compiled as an identity platform to deliver social services to residents of India.Aadhaar processes streams of biometric data as residents are enrolled and updated.Besides 1 million enrolments and updates per day,up to 100 million daily biometric authentications are expected during delivery of various public services.These form critical Big Data applications,with large volumes and high velocity of data.
The emerging trend of edge computing has led several cloud providers to release their own platforms for performing computation at the 'edge' of the network. We compare two such platforms, Amazon AWS Greengrass and Microsoft Azure IoT Edge, using a new benchmark comprising a suite of performance metrics. We also compare the performance of the edge frameworks to cloud-only implementations available in their respective cloud ecosystems. Amazon AWS Greengrass and Azure IoT Edge use different underlying technologies, edge Lambda functions vs. containers, and so we also elaborate on platform feature…
Cloud computing provides on-demand access to affordable hardware (multi-core CPUs, GPUs, disks, and networking equipment) and software (databases, application servers and data processing frameworks) platforms with features such as elasticity, pay-per-use, low upfront investment and low time to market. This has led to the proliferation of business critical applications that leverage various cloud platforms. Such applications hosted on single or multiple cloud provider platforms have diverse characteristics requiring extensive monitoring and benchmarking mechanisms to ensure run-time Quality of …
Serverless computing has seen a myriad of work exploring its potential. Some systems tackle Function-as-a-Service (FaaS) properties on automatic elasticity and scale to run highly-parallel computing jobs. However, they focus on specific platforms and convey that their ideas can be extrapolated to any FaaS runtime. An important question arises: do all FaaS platforms fit parallel computations? In this paper, we argue that not all of them provide the necessary means to host highly-parallel applications. To validate our hypothesis, we create a comparative framework and categorize the architectures…
The rise in popularity of permissioned blockchain platforms in recent time is significant. Hyperledger Fabric is one such permissioned blockchain platform and one of the Hyperledger projects hosted by the Linux Foundation. The Fabric comprises various components such as smart-contracts, endorsers, committers, validators, and orderers. As the performance of blockchain platform is a major concern for enterprise applications, in this work, we perform a comprehensive empirical study to characterize the performance of Hyperledger Fabric and identify potential performance bottlenecks to gain a bette…
Internet of Things (IoT) is a technology paradigm where millions of sensors monitor, and help inform or manage, physical, envi- ronmental and human systems in real-time. The inherent closed-loop re- sponsiveness and decision making of IoT applications makes them ideal candidates for using low latency and scalable stream processing plat- forms. Distributed Stream Processing Systems (DSPS) are becoming es- sential components of any IoT stack, but the efficacy and performance of contemporary DSPS have not been rigorously studied for IoT data streams and applications. Here, we develop a benchmark …
Data-driven solutions for the investment industry require event-based backend systems to process high-volume financial data feeds with low latency, high throughput, and guaranteed delivery modes. At vwd we process an average of 18 billion incoming event notifications from 500+ data sources for 30 million symbols per day and peak rates of 1+ million notifications per second using custom-built platforms that keep audit logs of every event. We currently assess modern open source event-processing platforms such as Kafka, NATS, Redis, Flink or Storm for the use in our ticker plant to reduce the mai…
With diverse IoT workloads, placing compute and analytics close to where data is collected is becoming increasingly important. We seek to understand what is the performance and the cost implication of running analytics on IoT data at the various available platforms. These workloads can be compute-light, such as outlier detection on sensor data, or compute-intensive, such as object detection from video feeds obtained from drones. In our paper, JANUS, we profile the performance/$ and the compute versus communication cost for a compute-light IoT workload and a compute-intensive IoT workload. In a…
Along with today's data explosion and application diversification, a variety of hardware platforms for big data are emerging, attracting interests from both industry and academia. The existing hardware platforms represent a wide range of implementation approaches, and different hardware platforms have different strengths. In this paper, we conduct comprehensive evaluations on three representative big data systems: Intel Xeon, Atom (low power processors), and many-core Tilera using BigDataBench - a big data benchmark suite. Then we explore the relative performance of the three implementation ap…
Imperial College（帝国理工学院）；The Univ. of Edinburgh（爱丁堡大学）；Univ. of Manchester（曼彻斯特大学）
Real-time dense computer vision and SLAM offer great potential for a new level of scene modelling, tracking and real environmental interaction for many types of robot, but their high computational requirements mean that use on mass market embedded platforms is challenging. Meanwhile, trends in low-cost, low-power processing are towards massive parallelism and heterogeneity, making it difficult for robotics and vision researchers to implement their algorithms in a performance-portable way. In this paper we introduce SLAMBench, a publicly-available software framework which represents a starting …
In this paper, using of automotive use cases as benchmarks for real-time system design has been proposed. The use cases are described in a format supported by AMALTHEA platform, which is a model based open source development environment for automotive multi-core systems. An example of a simple Electronic Control Unit has been analysed and presented with enough details to reconstruct this system in any format. For researchers willing to use AMALTHEA file format directly, an appropriate parser has been developed and offered. An example of applying this parser and benchmark for optimising makespa…
