Communication Pipeline Research Articles

A novel HPL-AI approach for FP16-only accelerator and its instantiation on Kunpeng+Ascend AI-specific platform

HPL-AI, also known as HPL-MxP, is a new benchmark program used to evaluate the upper-bound performance of AI-related tasks on a specific computing cluster. It solves a large linear equation system in FP64, preconditioned by complete LU factorization in lower precision. In this paper, we propose a new HPL-AI approach that relies on the factorization of the coefficient matrix in mixed precision: FP32 diagonals and FP16 off-diagonals. Without compromising the quality of the resultant LU preconditioner, the proposed approach only utilizes the primitive of dense matrix multiplication in FP16 on the accelerator, maximizing the FP16 throughput. Numerical analysis and experiments validate our approach, ensuring avoidance of numerical underflow or overflow during factorization. We implement the proposed approach on Kunpeng+Ascend clusters, a novel AI-specific platform with exceedingly high FP16 peak performance. By applying various optimization techniques, including 2D lookahead, HCCL-based communication pipeline, and SYCL-based tasks overlapping, we achieve 975 TFlops on a single node and nearly 100 PFlops on a cluster of 128 nodes, with a weak scalability of 79.8%.

Morphology of the submerged Ferdinandea Island, the ‘Neverland’ of the Sicily Channel (central Mediterranean Sea)

ABSTRACT We present the bathy-morphological map at a scale of 1: 50,000 of the area around the submerged Ferdinandea Island, the ‘Neverland’ of the Sicily Channel (central Mediterranean Sea). We investigate an area of 100 km2, between 10 and 350 m, which is part of a triangular morphological high, 360 km2 wide, representing the SE-wards prolongation of the Adventure Bank. The study is based on the morphometric analysis based on high resolution multibeam, and sub-bottom CHIRP profiles collected in 2015. The area around the remains of Ferdinandea Island is morphologically shaped by the interplay between volcanic, tectonic, fluid seepage, and oceanographic processes. Since the study area is considered a hot spot of biodiversity affected by maritime traffic (especially in Ferdinandea Channel) and hosting communication pipelines, this map provides insights both for habitat mapping purposes and preliminary marine geohazard assessment due to the occurrence of historically active submarine volcanoes, pockmarks, and mass transport deposits.

Clustered Scheduling and Communication Pipelining for Efficient Resource Management of Wireless Federated Learning

This paper proposes using communication pipelining to enhance the convergence speed of federated learning in mobile edge computing applications. Due to limited wireless sub-channels, a subset of the total clients is scheduled in each iteration of federated learning algorithms. On the other hand, the scheduled clients wait for the slowest client to finish its computation. We propose to first cluster the clients based on the time they need per iteration to compute the local gradients of the federated learning model. Then, we schedule a mixture of clients from all clusters to send their local updates in a pipelined manner. In this way, instead of just waiting for the slower clients to finish their computations, more clients can participate in each iteration. While the time duration of a single iteration does not change, the proposed method can significantly reduce the number of required iterations to achieve a target accuracy. We provide a generic formulation for optimal client clustering under different settings, and we analytically derive an efficient algorithm for obtaining the optimal solution. We also provide numerical results to demonstrate the gains of the proposed method for different datasets and deep learning architectures.

Data-Efficient Collaborative Decentralized Thermal-Inertial Odometry

We propose a system solution to achieve data-efficient, decentralized state estimation for a team of flying robots using thermal images and inertial measurements. Each robot can fly independently, and exchange data when possible to refine its state estimate. Our system front-end applies an online photometric calibration to refine the thermal images so as to enhance feature tracking and place recognition. Our system back-end uses a covariance-intersection fusion strategy to neglect the cross-correlation between agents so as to lower memory usage and computational cost. The communication pipeline uses Vector of Locally Aggregated Descriptors (VLAD) to construct a request-response policy that requires low bandwidth usage. We test our collaborative method on both synthetic and real-world data. Our results show that the proposed method improves by up to 46 % trajectory estimation with respect to an individual-agent approach, while reducing up to 89 % the communication exchange. Datasets and code are released to the public, extending the already-public JPL xVIO library.

Defending Trace-Back Attack in 3D Wireless Internet of Things

With the development of 5G, it is unsurprising that most of the smart devices in the Internet of Things (IoT) will be wirelessly connected with each other in the near future. This kind of lightweight, scalable and green network architecture will be well-received. In a wide variety of IoT application scenarios, sensor nodes deployed in a local space, such as a multistory building, automatically form a distributed 3D wireless IoT and it can be employed to collect and analyze environmental information. Source-location privacy protection is of great importance in these networks and however, most existing schemes focus on only planar distributed networks which are not suitable for the 3D networks. In this paper, we consider a novel trace-back attack for 3D wireless IoT and then design a source-location privacy protection scheme, named DMR-3D, to defend this kind of novel attacks. In DMR-3D, the source node first selects a set of virtual locations to indirectly choose a set of agent nodes based on the cold start sphere structure and the ellipsoid communication pipeline. Then, a sophisticated mechanism is designed based on both the connected graph and Multiple Delaunay Triangulation (MDT) structure of the network to deliver packets from the source node to the destination node via these agent nodes in a relay manner. Analysis and simulation results illustrate that the proposed scheme can effectively protect source-location privacy with a moderate increment of path stretch, time delay and data transmission amount.

Mass Production of Compassionate Communication in the Era of COVID-19: A How-To Guide from Parkland Hospital's COVID ICU Team (TH105)

Outcomes 1. Describe and analyze a blueprint for rapid upscaling of inpatient palliative care services in the ICU, both in the context of normal operations and in the context of mass casualties, natural disasters, pandemics, and other events 2. Describe the natural history, prognosis, morbidity, and mortality associated with acute respiratory distress syndrome due to COVID-19 3. Review and simulate use of novel communication tools and scripts for communication with families of critically ill patientsIn March 2020, Parkland Memorial Hospital, Dallas County's safety net hospital and one of the busiest hospitals in the nation, opened its Tactical Care Unit, a surgical space converted into a 100-bed unit for patients suffering from the novel and rapidly spreading COVID-19. At the outset of the pandemic, the team committed to expanding access to specialty-level palliative care and maintaining a pipeline of high-quality daily communication for all families of critically ill patients admitted to the COVID ICU. In this session, members of the multispecialty and multidisciplinary Parkland COVID ICU team will present a blueprint for the novel care model that allowed them to meet these goals, even in the midst of massive surges in the summer and winter. The following components of this care model will be reviewed in detail: a clearly defined structure for efficient co-management and co-rounding between palliative care and critical care specialists; the use of volunteer communication extenders; detailed data analysis regarding natural history, prognosis, morbidity, and mortality associated with acute respiratory distress syndrome due to COVID-19; and the generation of standardized, data-driven communication tools and scripts for daily conversations with families of critically ill patients. Attendees will receive copies of said communication tools and scripts, and we will conduct case-based simulations in small groups. Afterwards, we will review lessons learned and outcomes, answer questions, and review the ways in which the strategies and tools described above are being applied in our hospital outside the context of the COVID-19 pandemic.

Research data infrastructure for high-throughput experimental materials science

SummaryThe High-Throughput Experimental Materials Database (HTEM-DB, htem.nrel.gov) is a repository of inorganic thin-film materials data collected during combinatorial experiments at the National Renewable Energy Laboratory (NREL). This data asset is enabled by NREL's Research Data Infrastructure (RDI), a set of custom data tools that collect, process, and store experimental data and metadata. Here, we describe the experimental data flow from the RDI to the HTEM-DB to illustrate the strategies and best practices currently used for materials data at NREL. Integration of the data tools with experimental instruments establishes a data communication pipeline between experimental researchers and data scientists. This work motivates the creation of similar workflows at other institutions to aggregate valuable data and increase their usefulness for future machine learning studies. In turn, such data-driven studies can greatly accelerate the pace of discovery and design in the materials science domain.

Alarm method of communication intelligent manhole cover based on multiple event fusion

As a comprehensive utilization of pipeline resources, communication intelligent manhole cover (CIMC) can effectively real-time monitor communication manhole cover and protect the safety of communication pipeline. Due to the complex working environment of manhole cover and the random error of sensor, the traditional monitoring method usual leads to frequent false alarm in actual applications. In order to ensure the monitoring service quality and improve the service efficiency, a new alarm method of CIMC with multiple event fusion in this paper via jointing analysis of multi-sensor status signals is proposed based on the equipment status signals generated by the CIMC terminal and abnormal alarm events definition. The experimental result shows that the proposed CIMC alarm method by means of multiple sensor signals in this paper can not only make up for the defect of a single sensor, but also reduces the false alarm rate caused by the random error of sensor and CIMC system. It can promote the intelligent monitoring efficiency of the manhole cover and be conducive to the construction of intelligent transportation and smart city.

Leveraging natural history biorepositories as a global, decentralized, pathogen surveillance network.

The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) pandemic reveals a major gap in global biosecurity infrastructure: a lack of publicly available biological samples representative across space, time, and taxonomic diversity. The shortfall, in this case for vertebrates, prevents accurate and rapid identification and monitoring of emerging pathogens and their reservoir host(s) and precludes extended investigation of ecological, evolutionary, and environmental associations that lead to human infection or spillover. Natural history museum biorepositories form the backbone of a critically needed, decentralized, global network for zoonotic pathogen surveillance, yet this infrastructure remains marginally developed, underutilized, underfunded, and disconnected from public health initiatives. Proactive detection and mitigation for emerging infectious diseases (EIDs) requires expanded biodiversity infrastructure and training (particularly in biodiverse and lower income countries) and new communication pipelines that connect biorepositories and biomedical communities. To this end, we highlight a novel adaptation of Project ECHO’s virtual community of practice model: Museums and Emerging Pathogens in the Americas (MEPA). MEPA is a virtual network aimed at fostering communication, coordination, and collaborative problem-solving among pathogen researchers, public health officials, and biorepositories in the Americas. MEPA now acts as a model of effective international, interdisciplinary collaboration that can and should be replicated in other biodiversity hotspots. We encourage deposition of wildlife specimens and associated data with public biorepositories, regardless of original collection purpose, and urge biorepositories to embrace new specimen sources, types, and uses to maximize strategic growth and utility for EID research. Taxonomically, geographically, and temporally deep biorepository archives serve as the foundation of a proactive and increasingly predictive approach to zoonotic spillover, risk assessment, and threat mitigation.

Research Data Infrastructure for High-Throughput Experimental Materials Science

The High-Throughput Experimental Materials Database (HTEM-DB, htem.nrel.gov) is the endpoint repository for inorganic thin-film materials data collected during combinatorial experiments at the National Renewable Energy Laboratory (NREL). This data asset is enabled by NREL’s Research Data Infrastructure (RDI) – a set of custom data tools that collect, process, and store experimental data and metadata. Here, we describe the experimental data-tool workflow from the RDI to the HTEM-DB to illustrate the strategies and best practices currently used for materials data at NREL. Integration of these data tools with experimental processes establishes a data communication pipeline between experimental and data science researchers. In doing so, this work motivates the creation of similar data workflows at other institutions to aggregate valuable data and increase their usefulness for future data studies. These types of efforts can greatly accelerate the pace of learning and discovery in the materials science domain, by making data accessible to new and rapidly evolving data analysis methods.

An investigation into the allosteric mechanism of GPCR A2A adenosine receptor with trajectory-based information theory and complex network model

G protein-coupled receptors (GPCRs), a large superfamily of transmembrane (TM) proteins, allosterically transduce the signal of ligand binding in the extracellular (EC) domain to couple to effector proteins in the intracellular (IC) domain, therefore forming the largest class of drug targets. The A2A adenosine receptor (A2AAR), a class-A GPCR, has been extensively studied as it offers numerous possibilities for therapeutic applications. However, the mechanism of allosteric communication between EC and IC domains is not completely clear. In this work, we utilize torsional mutual information to quantify the correlated motions of residue pairs from its molecular dynamics (MD) simulation trajectories, and further use the complex network model to obtain allosteric pipelines and hubs. The identified allosteric communication pipelines mainly transmit the signal from EC domain to the cytoplasmic ends of TM helix 5 (TM5), TM6 and TM7. The allosteric hubs, mostly located at TM5, TM6 and TM7, play an important role in mediating allosteric signal transmission to keep the receptor rigid and prevent G protein from binding to IC domain, which can explain the reason why their mutations distant from ligand-binding site do not affect the ligand binding affinity but affect the ligand efficacy. Additionally, we identify the key residues located in antagonist ZM241385 binding pocket which mediate multiple allosteric pathways and have been experimentally proven to play a critical role in affecting the ligand potency. This study is helpful for understanding the allosteric communication mechanism of A2AAR, and can provide valuable information for the structure-based drug design of GPCRs. Communicated by Ramaswamy H. Sarma

Mathematical Characterization and Simulation of LoRa

LoRa is a recently developed physical layer modulation technique characterized by low power consumption and long range communication capabilities. Due to the organic relevance of these features in the Internet of Things, LoRa has gained significant traction in the contemporary telecom industry. LoRa has been rigorously analyzed with regard to its functionalities and applications. However, due to its proprietary nature, precise descriptions of the underlying modulation waveforms are not publicly available. In this paper, we first derive a mathematical characterization of LoRa with regard to its chirp frequency and phase. Our analysis shows that a single matched filter is sufficient to decode all LoRa data symbols. Further, when the output of the matched filter is sampled at a rate equal to the transmission bandwidth, a constant frequency output is produced, which directly corresponds to the transmitted symbol value. For a visual understanding of LoRa, we present detailed simulations of its waveforms at different stages of the communication pipeline. In particular, the error performance simulation clearly demonstrates that LoRa can receive and decode data packets even when the SNR is below 0 dB.

Visual Distortions in 360° Videos

Omnidirectional (or 360°) images and videos are emergent signals being used in many areas, such as robotics and virtual/augmented reality. In particular, for virtual reality applications, they allow an immersive experience in which the user can interactively navigate through a scene with three degrees of freedom, wearing a head-mounted display. Current approaches for capturing, processing, delivering, and displaying 360° content, however, present many open technical challenges and introduce several types of distortions in the visual signal. Some of the distortions are specific to the nature of 360° images and often differ from those encountered in classical visual communication frameworks. This paper provides a first comprehensive review of the most common visual distortions that alter 360° signals going through the different processing elements of the visual communication pipeline. While their impact on viewers' visual perception and the immersive experience at large is still unknown-thus, it is an open research topic-this review serves the purpose of proposing a taxonomy of the visual distortions that can be encountered in 360° signals. Their underlying causes in the end-to-end 360° content distribution pipeline are identified. This taxonomy is essential as a basis for comparing different processing techniques, such as visual enhancement, encoding, and streaming strategies, and allowing the effective design of new algorithms and applications. It is also a useful resource for the design of psycho-visual studies aiming to characterize human perception of 360° content in interactive and immersive applications.

A Little Knowledge is a Dangerous Thing: Excess Confidence Explains Negative Attitudes Towards Science

Scientific knowledge has been accepted as the main driver of development, allowing for longer, healthier, and more comfortable lives. Still, public support to scientific research is wavering, with large numbers of people being uninterested or even hostile towards science. This is having serious social consequences, from the anti-vaccination community to the recent post-truth movement. Such lack of trust and appreciation for science was first justified as lack of knowledge, leading to the Model. As an increase in scientific information did not necessarily lead to a greater appreciation, this model was largely rejected, giving rise to Public Engagement Models. These try to offer more nuanced, two-way, communication pipelines between experts and the general public, strongly respecting non-expert knowledge, possibly even leading to an undervaluing of science. Therefore, we still lack an encompassing theory that can explain public understanding of science, allowing for more targeted and informed approaches. Here, we use a large dataset from the Science and Technology Eurobarometer surveys, over 25 years in 34 countries, and find evidence that a combination of confidence and knowledge is a good predictor of attitudes towards science. This is contrary to current views, that place knowledge as secondary, and in line with findings in behavioral psychology, particularly the Dunning-Kruger effect, as negative attitudes peak at intermediate levels of knowledge, where confidence is largest. We propose a new model, based on the superposition of the Deficit and Dunning-Kruger models and discuss how this can inform science communication.

Assessment of the effectiveness of forest use in the Republic of Tatarstan

As a result of the research work, the structure of the organization of forestry management in the Russian Federation and its constituent entities, acting following the new forest code of the Russian Federation in 2006, was studied. Assessment of the effectiveness of the use of forests in the republic, a multiple correlation analysis was applied. As a result, it was revealed that in the Republic of Tatarstan, the primary income for the use of forests is accounted for by timber harvesting, the bulk of which is accounted for by the use of coniferous forest areas – 81 %, soft-leaved, and hard-leaved forests accounted for 14 % and 5 %, respectively. Taking into account the share of exploitable forests for the receipt of payments from the lease of forest plots for all types of use (harvesting and collection of non-timber forest resources; hunting; farming; recreational activities; geological exploration of mineral resources; development of mineral deposits; construction, reconstruction and operation of communication lines, roads, pipelines and other linear facilities; payment for the conversion of forest land to non-forest land and the transfer of forest land to lands of other categories) in the Republic of Tatarstan, the structure of forests by farms affects, and on the contribution to payments by type of use, soft-leaved farming brings 78 %, hard-leaved 12 % and coniferous 10 %.

Case Study on Communication Based Cooperative Longitudinal Vehicle Guidance

Abstract An approach of a proven concept for cooperative longitudinal vehicle guidance is presented. The paper briefly describes the communication pipeline as well as the incorporated communication technology. Moreover, the used longitudinal controller, which explicitly considers the communicated states of the vehicles ahead, is addressed. Here, we are expanding a classical Adaptive Cruise Controller (ACC) to realize the cooperative longitudinal guidance functionality simply and robustly. On the one hand, the current states of the vehicle in front are transmitted and on the other hand the desired ones. Results of a real-world test campaign of platooning vehicles illustrate the effectiveness of the proposed approach.

On the Reusability of Postexperimental Field Data for Underwater Acoustic Communications R&D

Field data are often expensive to collect, time-consuming to prepare to collect, and even more time-consuming to process after the experiment has concluded. However, it is often the practice that such data are used for little after the funded research activity that was concomitant with the experiment is completed. Immutability of the original experimental configuration either results in regathering of expensive field data, or in absence of such data, model-dependent analysis that partially captures the real-world dynamics. For underwater acoustic research and development, the standard communication pipeline might be modified to enable greater reusability of experimental field data. This paper first characterizes the necessary modifications to the standard communication pipeline to prepare signals for transmission and subsequent recording such that research trades for different modulation and coding schemes may be undertaken postexperiment, without the need for retransmission of additional waveforms. Then, using the modified mathematical framework, sufficient conditions for reliable postexperimental replay of the environment are recognized. Finally, techniques are discussed to collect sufficient environmental statistics such that subsequent research can be accomplished long after the experiment has been completed, and that results from a given experiment may be reasonably compared with those of another. Examples are provided using both synthetic and experimental data collected from at-sea field tests.

A Scalable and Adaptable ILP-Based Approach for Task Mapping on MPSoC Considering Load Balance and Communication Optimization

Task mapping has been a hot topic in multiprocessor system-on-chip software design for decades. During the mapping process, load balance (LB) and communication optimization have been two important performance optimization factors. This paper studies the relations between LB, interprocessor communications, and communication pipeline technique during the mapping process, and proposes an integer linear programming (ILP)-based static task mapping approach, which considers both LB and communication optimization. The approach consists of an optimized ILP model for task mapping with fewer variables compared to previous ILP mapping works. Moreover, to enhance the scalability of the ILP task mapping, the task-processor-cluster algorithm is proposed to reduce the scale of the task graph and the number of processors and then solve the coarse-grained input by the ILP mapping. To increase the adaptability of the ILP task mapping, the improved augmented $\epsilon $ -constraint method is further integrated with the ILP formulations to select the best mapping for different applications. Experimental results on a 2/4/8/16/24-CPU platform of both synthetic and real-life benchmarks demonstrate the efficiency of the proposed approach.

Predicting the Quality of Images Compressed after Distortion in Two Steps.

In a typical communication pipeline, images undergo a series of processing steps that can cause visual distortions before being viewed. Given a high quality reference image, a reference (R) image quality assessment (IQA) algorithm can be applied after compression or transmission. However, the assumption of a high quality reference image is often not fulfilled in practice, thus contributing to less accurate quality predictions when using stand-alone R IQA models. This is particularly common on social media, where hundreds of billions of usergenerated photos and videos containing diverse, mixed distortions are uploaded, compressed, and shared annually on sites like Facebook, YouTube, and Snapchat. The qualities of the pictures that are uploaded to these sites vary over a very wide range. While this is an extremely common situation, the problem of assessing the qualities of compressed images against their precompressed, but often severely distorted (reference) pictures has been little studied. Towards ameliorating this problem, we propose a novel two-step image quality prediction concept that combines NR with R quality measurements. Applying a first stage of NR IQA to determine the possibly degraded quality of the source image yields information that can be used to quality-modulate the R prediction to improve its accuracy. We devise a simple and efficient weighted product model of R and NR stages, which combines a pre-compression NR measurement with a post-compression R measurement. This first-of-a-kind two-step approach produces more reliable objective prediction scores. We also constructed a new, first-of-a-kind dedicated database specialized for the design and testing of two-step IQA models. Using this new resource, we show that twostep approaches yield outstanding performance when applied to compressed images whose original, pre-compression quality covers a wide range of realistic distortion types and severities. The two-step concept is versatile as it can use any desired R and NR components. We are making the source code of a particularly efficient model that we call 2stepQA publicly available at https://github.com/xiangxuyu/2stepQA. We are also providing the dedicated new two-step database free of charge at http://live.ece.utexas.edu/research/twostep/index.html.

Fine-Grained Communication-Aware Task Scheduling Approach for Acyclic and Cyclic Applications on MPSoCs

Fine-grained task models can exploit parallelism to achieve high performance for multiprocessor system-on-chip (MPSoC). However, fine-grained models face the issues of high-communication overhead and difficult scheduling decisions, and the two challenges are inter-dependent. To address the issues, this paper gives a full analysis of the fine-grained communication optimization technique and communication pipeline, from both time and topology perspectives, and proposes a static fine-grained communication-aware task scheduling (FCATS) approach, which integrates scheduling with communication pipeline for acyclic and cyclic applications based on the fine-grained Simulink model. The approach contains search-based scheduling with high-quality solutions utilizing genetic algorithm-integer linear programming (GA-ILP) and hybrid GA-heuristic scheduling with short solving time to meet different demands for users. The experimental results with both synthetic and real-life benchmarks on the 4/8/16-CPU platform demonstrate the efficiency of the approach on performance improvements compared to previous works.