Search

Containers are a mainstream virtualization technique commonly used to run stateful workloads over persistent storage. In multi-tenant hosts with high utilization, resource contention at the system kernel often leads to inefficient handling of the container I/O. Assuming a distributed storage architecture for scalability, resource sharing is particularly problematic at the client hosts serving the applications of competing tenants. Although increasing the scalability of a system kernel can improve resource efficiency, it is highly challenging to refactor the kernel for fair access to system services. As a realistic alternative, we isolate the storage I/O paths of different tenants by serving them with distinct clients running at user level. We introduce the Danaus client architecture to let each tenant access the container root and application filesystems over a private host path. We developed a Danaus prototype that integrates a union filesystem with a Ceph distributed filesystem client and a configurable shared cache. Across different host configurations, workloads and systems, Danaus achieves improved performance stability because it handles I/O with reserved per-tenant resources and avoids intensive kernel locking. Danaus offers up to 14.4x higher throughput than a popular kernel-based client under conditions of I/O contention. In comparison to a FUSE-based user-level client, Danaus also reduces by 14.2x the time to start 256 high-performance webservers. Based on our extensive experience from building and evaluating Danaus, we share several valuable lessons that we learned about resource contention, file management, service separation and performance stability.

Function-as-a-Service (FaaS) has raised a growing interest in how to "tame" serverless to enable domain-specific use cases such as data-intensive applications and machine learning (ML), to name a few. Recently, several systems have been implemented for training ML models. Certainly, these research articles are significant steps in the correct direction. However, they do not completely answer the nagging question of when serverless ML training can be more cost-effective compared to traditional "serverful" computing. To help in this task, we propose MLLess, a FaaS-based ML training prototype built atop IBM Cloud Functions. To boost cost-efficiency, MLLess implements two key optimizations: a significance filter and a scale-in auto-tuner, and leverages them to specialize model training to the FaaS model. Our results certify that MLLess can be 15X faster than serverful ML systems [24] at a lower cost for ML models (such as sparse logistic regression and matrix factorization) that exhibit fast convergence.

This companion paper is to support the replication of paper "Campus3D: A Photogrammetry Point Cloud Benchmark for Outdoor Scene Hierarchical Understanding", which was presented at ACM Multimedia 2020. The supported paper's main purpose was to provide a photogrammetry point cloud-based dataset with hierarchical multilabels to facilitate the area of 3D deep learning. Based on this provided dataset and source code, in this work, we build a complete package to reimplement the proposed methods and experiments (i.e., the hierarchical learning framework and the benchmarks of the hierarchical semantic segmentation task). Specifically, this paper contains the technical details of the package, including file structure, dataset preparation, installation package, and the conduction of the experiment. We also present the replicated experiment results and indicate our contributions to the original implementation.

In this companion paper, we provide the details of the reproducibility artifacts of the paper "Visual Relation of Interest Detection" presented at MM'20. Visual Relation of Interest Detection (VROID) aims to detect visual relations that are important for conveying the main content of an image. In this paper, we explain the file structure of the source code and publish the details of our ViROI dataset, which can be used to retrain the model with custom parameters. We also detail the scripts for component analysis and comparison with other methods and list the parameters that can be modified for custom training and inference.

In this paper we reproduce experimental results presented in our earlier work titled "Describing Subjective Experiment Consistency by $p$-Value P-P Plot" that was presented in the course of the 28th ACM International Conference on Multimedia. The paper aims at verifying the soundness of our prior results and helping others understand our software framework. We present artifacts that help reproduce tables, figures and all the data derived from raw subjective responses that were included in our earlier work. Using the artifacts we show that our results are reproducible. We invite everyone to use our software framework for subjective responses analyses going beyond reproducibility efforts., Comment: Please refer to the original publication: https://dl.acm.org/doi/10.1145/3474085.3477935 Related paper: https://dl.acm.org/doi/abs/10.1145/3394171.3413749 or arXiv:2009.13372

This chapter focuses on cost control by analyzing the intelligent control system of waste paper deinking production line, and finally designs a solution. In the aspect of control index, machine learning has designed a solution that pays less attention to the benefit of a single control loop and focuses on maximizing the overall benefit of the production line. In the aspect of control demand, according to the complexity of waste paper deinking production line, a distributed control solution is designed Finally, according to the dynamic fuzziness in machine learning system, the basic concept of dynamic fuzzy machine learning is constructed by using dynamic fuzzy set The dynamic fuzzy machine learning model is described and the related learning algorithm is given. Finally, a simulation example is given, and satisfactory results are obtained by comparing with the elastic BP algorithm RPROP of neural network.

Logic Programming (LP) is considered to be relatively simple for non-programmers, and allows the developer to focus on developing facts and rules of a logical derivation, and not on algorithms. Secure multiparty computation (MPC) is a methodology that allows several parties to process private data collaboratively without revealing the data to any party. In this paper, we bring together the notions of MPC and LP, allowing users to write privacy-preserving applications in logic programming language.

We propose new solutions that can efficiently address the problem of malicious memory access from pluggable computer peripherals and microcontrollers embedded within a system-on-chip. This problem represents a serious emerging threat to total-system computer security. Previous work has shown that existing defenses are insufficient and poorly deployed, in part due to performance concerns. In this paper we explore the threat and its implications for system architecture. We propose a range of protection techniques, from lightweight to heavyweight, across different classes of systems. We consider how emerging capability architectures (and specifically the CHERI protection model) can enhance protection and provide a convenient bridge to describe interactions among software and hardware components. Finally, we describe how new schemes may be more efficient than existing defenses.

Video cameras are becoming ubiquitous in our daily lives. With the recent advancement of Artificial Intelligence (AI), live video analytics are enabling various useful services, including traffic monitoring and campus surveillance. However, current video analytics systems are highly limited in leveraging the enormous opportunities of the deployed cameras due to (i) centralized processing architecture (i.e., cameras are treated as dumb streaming-only sensors), (ii) hard-coded analytics capabilities from tightly coupled hardware and software, (iii) isolated and fragmented camera deployment from different service providers, and (iv) independent processing of camera streams without any collaboration. In this paper, we envision a full-fledged system for software-defined video analytics with cross-camera collaboration that overcomes the aforementioned limitations. We illustrate its detailed system architecture, carefully analyze the key system requirements with representative app scenarios, and derive potential research issues along with a summary of the status quo of existing works.

Storm and flood disasters such as typhoons and torrential rains are becoming more intense and frequent. The national government and municipalities must respond to such natural disasters as soon as possible. When the scale of damage is large; however, it takes much time to investigate the severity of damage, and the initial response can be delayed. If we could precisely and rapidly estimate the severity of damage for each city at an early stage, the national government would be able to better support the municipalities, and consequently respond quickly to help citizens. In this paper, we propose a novel approach to estimate the severity of disaster damage within a short time period after a disaster occurs by exploiting real-time population data generated from cellular networks. First, we investigate the relationship between real-time population data and the severity of damage. Then, we design a Graph Convolutional Networks for Disaster Damage Estimation, called D2E-GCN, which fully exploits the directed and weighted characteristics of human mobility graph. We conduct an offline evaluation on real-world datasets including two typhoons that hit Japan. The evaluation results show that the proposed method outperforms baseline methods which do not consider the graph structure of cities, and the proposed method can estimate the severity of damage approximately 48 hours after typhoons passed. Moreover, we find the experimental insight that the estimation performance can be significantly affected by the graph construction method for GCN models.

In federated learning (FL), model updates from mutually distrusting parties are aggregated in a centralized fusion server. The concentration of model updates simplifies FL's model building process, but might lead to unforeseeable information leakage. This problem has become acute due to recent FL attacks that can reconstruct large fractions of training data from ostensibly "sanitized" model updates. In this paper, we re-examine the current design of FL systems under the new security model of reconstruction attacks. To break down information concentration, we build TRUDA, a new cross-silo FL system, employing a trustworthy and decentralized aggregation architecture. Based on the unique computational properties of model-fusion algorithms, we disassemble all exchanged model updates at the parameter-granularity and re-stitch them to form random partitions designated for multiple hardware-protected aggregators. Thus, each aggregator only has a fragmentary and shuffled view of model updates and is oblivious to the model architecture. The deployed security mechanisms in TRUDA can effectively mitigate training data reconstruction attacks, while still preserving the accuracy of trained models and keeping performance overheads low.

Blind natural video quality assessment (BVQA), also known as no-reference video quality assessment, is a highly active research topic. In our recent contribution titled "Blind Natural Video Quality Prediction via Statistical Temporal Features and Deep Spatial Features" published in ACM Multimedia 2020, we proposed a two-level video quality model employing statistical temporal features and spatial features extracted by a deep convolutional neural network (CNN) for this purpose. At the time of publishing, the proposed model (CNN-TLVQM) achieved state-of-the-art results in BVQA. In this paper, we describe the process of reproducing the published results by using CNN-TLVQM on two publicly available natural video quality datasets.

This companion paper supports the replication of the fashion trend forecasting experiments with the KERN (Knowledge Enhanced Recurrent Network) method that we presented in the ICMR 2020. We provide an artifact that allows the replication of the experiments using a Python implementation. The artifact is easy to deploy with simple installation, training and evaluation. We reproduce the experiments conducted in the original paper and obtain similar performance as previously reported. The replication results of the experiments support the main claims in the original paper.

Over 100 fatalities and more than 8000 injuries are reported on average every day in the US caused by motor vehicle accidents. In order to provide drivers a safer travel plan, we present a machine learning powered risk profiler for road segments using geo-spatial data. We built an end-to-end pipeline to extract static road features from map data and combined them with other data such as weather and traffic patterns. Our approach proposes novel methods for data pre-processing and feature engineering using statistical and clustering methods. Our model achieves significant performance improvement for risk prediction using hyper-parameter optimization (HPO) and the open source AutoGluon library to optimize the ML model. Finally, an enduser visualization interface is developed in the form of interactive maps. The results indicate 31% improvement in model performance compared to baseline when model is applied to a new geo location. We tested this approach on six major cities in the US. The findings of this research will provide users a tool to quantitatively assess accident risk at road segment level.

Trusted execution as offered by Intel's Software Guard Extensions (SGX) is considered as an enabler to protect the integrity and confidentiality of stateful workloads such as key-value stores and databases in untrusted environments. These systems are typically long running and require extension mechanisms built on top of dynamic loading as well as hot-patching to avoid downtimes and apply security updates faster. However, such essential mechanisms are currently neglected or even missing in combination with trusted execution. We present sgx-dl, a lean framework that enables dynamic loading of enclave code at the function level and hot-patching of dynamically loaded code. Additionally, sgx-dl is the first framework to utilize the new SGX version 2 features and also provides a versioning mechanism for dynamically loaded code. Our evaluation shows that sgx-dl introduces a performance overhead of less than 5% and shrinks application downtime by an order of magnitude in the case of a database system.

Our eventual goal regarding this study is to support students with developing paper-writing skill. In order to achieve this goal, we have been trying to find characteristic features of good papers through analyzing educational and other kinds of data. We take conference papers as target data and suppose full/regular papers are good because they are chosen as reviewers evaluate them more valuable to be presented in the conference than other papers. In our series of study, we have been surveying the differences of full and short papers. In this paper, we aim to investigate further differences of them by taking different analysis method. We have been using the numbers of occurrences of each word in full/short papers as the data for analysis. In this paper we use the numbers of full/short papers that contain the word instead of the total numbers of occurrences of words. We define a new index of a word which shows how likely it is used in full or short papers. We discuss its effectiveness by applying it to an experiment of distinguishing full and short papers.

Currently, the amount of information available to Web users is very large, and this situation is similar for scientific communities when searching for papers for their research. Recommender Systems (RSs) can help in this task because they combine computational techniques to select personalized items based on the users’ interests and according to the context in which users are inserted. The increase in the impact and scope of recommendations in the users’ lives, leads to the result on the ethical issues involved in the generation of recommendations and indicators for visualizing the results of the algorithms found. This paper presents a Recommender System for the Human-Computer Interaction (HCI) community, indicating papers from the Brazilian Symposium on Human Factors in Computing Systems related to the users’ profile applied to a post-processing strategy focused on fairness to balance the users’ interests. After the development of the RS and the Web environment, the results were obtained on the impact that the tool had on the community and demonstrated through the evaluation of the system.

Automatically detecting the positions of key-points (e.g., facial key-points or finger key-points) in an image is an essential problem in many applications, such as driver's gaze detection and drowsiness detection in automated driving systems. With the recent advances of Deep Neural Networks (DNNs), Key-Points detection DNNs (KP-DNNs) have been increasingly employed for that purpose. Nevertheless, KP-DNN testing and validation have remained a challenging problem because KP-DNNs predict many independent key-points at the same time -- where each individual key-point may be critical in the targeted application -- and images can vary a great deal according to many factors. In this paper, we present an approach to automatically generate test data for KP-DNNs using many-objective search. In our experiments, focused on facial key-points detection DNNs developed for an industrial automotive application, we show that our approach can generate test suites to severely mispredict, on average, more than 93% of all key-points. In comparison, random search-based test data generation can only severely mispredict 41% of them. Many of these mispredictions, however, are not avoidable and should not therefore be considered failures. We also empirically compare state-of-the-art, many-objective search algorithms and their variants, tailored for test suite generation. Furthermore, we investigate and demonstrate how to learn specific conditions, based on image characteristics (e.g., head posture and skin color), that lead to severe mispredictions. Such conditions serve as a basis for risk analysis or DNN retraining., to appear in ISSTA 2021

Open-source hardware such as RISC-V has been gaining substantial momentum. Recently, they have begun to embrace Google's Android operating system to leverage its software ecosystem. Despite the encouraging progress, a challenging issue arises: a majority of Android applications are written in native languages and need to be recompiled to target new hardware platforms. Unfortunately, this recompilation process is not scalable because of the explosion of new hardware platforms. To address this issue, we present WasmAndroid, a high-performance cross-platform runtime for native programming languages on Android. WasmAndroid only requires developers to compile their source code to WebAssembly, an efficient and portable bytecode format that can be executed everywhere without additional reconfiguration. WasmAndroid can also trans-pile existing application binaries to WebAssembly when source code is not available. WebAssembly's language model is very different from C/C++ and this mismatch leads to many unique implementation challenges. In this paper, we provide workable solutions and conduct a preliminary system evaluation. We show that WasmAndroid provides acceptable performance to execute native applications in a cross-platform manner.

Furniture design depends on different materials. Under the social background of building an ecological and civilized society, furniture design based on paper materials makes people have a new understanding and change of paper products. Home design works exist depending on different materials. With many materials, designers can use their imagination and hands to give it different shapes and colors, and combine the author's personal consciousness and emotions to create. In various industries, traditional design ideas, design methods and design means lag behind the development of the times due to their inherent shortcomings, thus restricting the improvement of production and production efficiency. Future furniture must be inseparable from the concept of green environmental protection and recycling. This paper will take furniture products as the breakthrough point, explore the new structure and new technology of furniture product design based on computer aided design, look for the future design trend of furniture products based on the social form of energy saving and environmental protection, and analyze and explain the circulation mode of paper materials themselves.

The author claimed for the artifact associated with his paper the following ACM Reproducibility badges:(1) Artifact Available,(2) Artifact Evaluated-Functional,(3) Results Reproduced. After an in-depth review process, we agree to assign all the requested badges as we found it to meet the following requirements:i) it is uploaded on a persistent repository, accessible via a DOI; ii) it is well documented, consistent with the presented data, complete of all the necessary software sources and packages, and exercisable; iii) it is exhaustive in the reproduction of all the relevant data of the paper. Some curves in some reproduced plots are truncated, due to the computational limits imposed by the short-term deadline of the review process. Nevertheless, the overall trends are respected, and the curves are supporting the paper's claims.

This is like an abstract to a paper, but it is more abstract. In fact, it is the introduction to something which is a not paper. The global Covid-19 pandemic of 2020 represented an inflection point for our post-post-modern world, a moment where our old normal was dramatically arrested. We are now in a state of comprehensive flux as ‘new normals’ emerge, begin to solidify, and may evolve into an—as yet undetermined—futures. This not paper is a facet and exploration of that flux as it relates to publication and conference culture, video conferencing systems, and how we both conduct, and share, research. You should read the whole of this abstract, but then you should take a step inside the not paper, it lives on the web over here https://designresearch.works/thisisnotapaper/

Measuring the quality of research work is an essential component of the scientific process. With the ever-growing rates of articles being submitted to top-tier conferences, and the potential consistency and bias issues in the peer review process identified by scientific community, it is thus of great necessary and challenge to automatically evaluate submissions. Existing works mainly focus on exploring relevant factors and applying machine learning models to simply be accurate at predicting the acceptance of a given academic paper, while ignoring the interpretability power which is required by a wide range of applications. In this paper, we propose a framework to construct decision sets that consist of unordered if-then rules for predicting paper acceptance. We formalize decision set learning problem via a joint objective function that simultaneously optimize accuracy and interpretability of the rules, rather than organizing them in a hierarchy. We evaluate the effectiveness of the proposed framework by applying it on a public scientific peer reviews dataset. Experimental results demonstrate that the learned interpretable decision sets by our framework performs on par with state-of-the-art classification algorithms which optimize exclusively for predictive accuracy and much more interpretable than rule-based methods.

New discoveries in science are often built upon previous knowledge. Ideally, such dependency information should be made explicit in a scientific knowledge graph. The Keystone Framework was proposed for tracking the validity dependency among papers. A keystone citation indicates that the validity of a given paper depends on a previously published paper it cites. In this paper, we propose and evaluate a strategy that repurposes rhetorical category classifiers for the novel application of extracting keystone citations that relate to research methods. Five binary rhetorical category classifiers were constructed to identify Background, Objective, Methods, Results, and Conclusions sentences in biomedical papers. The resulting classifiers were used to test the strategy against two datasets. The initial strategy assumed that only citations contained in Methods sentences were methods keystone citations, but our analysis revealed that citations contained in sentences classified as either Methods or Results had a high likelihood to be methods keystone citations. Future work will focus on fine tuning the rhetorical category classifiers, experimenting with multiclass classifiers, evaluating the revised strategy with more data, and constructing a larger gold standard citation context sentence dataset for model training.

Using machine learning (ML) services, both service customers and providers need to monitor complex contractual constraints of ML service that are strongly related to ML models and data. Therefore, establishing and monitoring comprehensive ML contracts are crucial in ML serving. This paper demonstrates a set of features and utilities of the QoA4ML framework for ML contracts.

We propose the creation of a repository which collects and makes available network-constrained trajectory data. The repository should become a central instance for researchers who want to work with network-constrained trajectory data on a large scale, allowing for efficient filtering and export of selected trajectories based on spatial, temporal and semantic attributes.

Programming languages with sufficiently expressive type systems provide users with different means of data ‘bundling’. Specifically, in dependently-typed languages such as Agda, Coq, Lean and Idris, one can choose to encode information in a record either as a parameter or a field. For example, we can speak of graphs over a particular vertex set, or speak of arbitrary graphs where the vertex set is a component. These create isomorphic types, but differ with respect to intended use. Traditionally, a library designer would make this choice (between parameters and fields); if a user wants a different variant, they are forced to build conversion utilities, as well as duplicate functionality. For a graph data type, if a library only provides a Haskell-like typeclass view of graphs over a vertex set, yet a user wishes to work with the category of graphs, they must now package a vertex set as a component in a record along with a graph over that set. We design and implement a language feature that allows both the library designer and the user to make the choice of information exposure only when necessary, and otherwise leave the distinguishing line between parameters and fields unspecified. Our language feature is currently implemented as a prototype meta-program incorporated into Agda’s Emacs ecosystem, in a way that is unobtrusive to Agda users.

PDF has long been established as the de facto format for the exchange of print-oriented documents and is known for its robust visual presentation across a variety of operating systems and platforms.However, relatively few users are familiar with the format's newer features, such as tagging, forms and security. This tutorial aims to give an overview of the most important of these features and demonstrate the benefits of creating and exchanging PDF files that make use of them.

Recent episodes of web overloads suggest the need to test system performance under loads that reflect extreme variations in usage patterns well outside normal anticipated ranges. These loads are sometimes expected or even scheduled. Examples of expected loads include surges in transactions or request submission when popular rock concert tickets go on sale, when the deadline for the submission of census forms approaches, and when a desperate population is attempting to sign up for a vaccination during a pandemic. Examples of unexpected loads are the surge in unemployment benefit applications in many US states with the onset of COVID19 lockdowns and repeated queries about the geographic distribution of signatories on the U.K. Parliament's petition website prior to a Brexit vote in 2019. We will consider software performance ramifications of these examples and the architectural questions they raise. We discuss how modeling and performance testing and known processes for evaluating architectures and designs can be used to identify potential performance issues that would be caused by sudden increases in load or changes in load patterns.

Keeping up with the rapid growth of Deep Learning (DL) research is a daunting task. While existing scientific literature search systems provide text search capabilities and can identify similar papers, gaining an in-depth understanding of a new approach or an application is much more complicated. Many publications leverage multiple modalities to convey their findings and spread their ideas - they include pseudocode, tables, images and diagrams in addition to text, and often make publicly accessible their implementations. It is important to be able to represent and query them as well. We utilize RDF Knowledge graphs (KGs) to represent multimodal information and enable expressive querying over modalities. In our demo we present an approach for extracting KGs from different modalities, namely text, architecture images and source code. We show how graph queries can be used to get insights into different facets (modalities) of a paper, and its associated code implementation. Our innovation lies in the multimodal nature of the KG we create. While our work is of direct interest to DL researchers and practitioners, our approaches can also be leveraged in other scientific domains.

In this companion paper, firstly, we briefly summarize the contributions of our main manuscript: Selective Deep Convolutional Features for Image Retrieval, published in ACM MultiMedia 2017. In addition, we provide detail instructions together with pre-configured MATLAB scripts which allow experiments to be executed and to reproduce the results reported in our main manuscript effortlessly. The source code is available at https://github.com/hnanhtuan/selectiveConvFeatures_ACMMM_reproducibility.

In graduate education, the quality of academic papers can reflect individual scientific research achievements. This study compared the differences in linguistic features between the papers of experts and graduate students with natural language processing. More specifically, for revealing the problems existing in the writing of graduate students, this study analyzed academic papers in a journal as experts and those of graduate students by using Chinese word segmentation. The study found that the graduate students and experts have differences in the use of various parts of speech (i.e., nouns, verbs, adjectives, pronouns, and conjunctions) and connective words (i.e., organization, comparison, summarization, etc.). The results may be helpful for further designing scaffolding to improve the writing quality of graduate students’ papers.

Named entity recognition has been extensively studied in the past decade. The state-of-the-art models, trained on general text such as Wikipedia articles and newsletters, have achieved F_1>0.90. Entity types are focused on people, location, organization, etc. However, entity recognition from domain-specific text, in particular research papers, is still challenging. In this paper, we perform a comparative study of sequence tagging (ST) methods on this task using a manually curated corpus from biomedical papers on Lyme disease. Each model we compare consists of a ST and a non-ST classification component. In this pilot study, we freeze the non-ST classifier to study how the ST component performs with variants of the conditional random field (CRF) and bidirectional long short-term memory (BiLSTM). The results shed light on the importance of pre-trained word embeddings such as ELMo and the residual unit. The attention mechanism and enriched features do not seem to boost the performance in recognizing entity mentions and their positions, which is likely to be caused by the relatively small training sample. We plan to improve the model by increasing the training corpus size and trying different combinations of features.

Linking to code repositories, such as on GitHub, in scientific papers becomes increasingly common in the field of computer science. The actual quality and usage of these repositories are, however, to a large degree unknown so far. In this paper, we present for the first time a thorough analysis of all GitHub code repositories linked in scientific papers using the Microsoft Academic Graph as a data source. We analyze the repositories and their associated papers with respect to various dimensions. We observe that the number of stars and forks, respectively, over all repositories follows a power-law distribution. In the majority of cases, only one person from the authors is contributing to the repository. The GitHub manuals are mostly kept rather short with few sentences. The source code is mostly provided in Python. The papers containing the repository URLs as well as the papers' authors are typically from the AI field.