skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Efficient Edge Inference by Selective Query
Edge devices provide inference on predictive tasks to many end-users. However, deploying deep neural networks that achieve state-of-the-art accuracy on these devices is infeasible due to edge resource constraints. Nevertheless, cloud-only processing, the de-facto standard, is also problematic, since uploading large amounts of data imposes severe communication bottlenecks. We propose a novel end-to-end hybrid learning framework that allows the edge to selectively query only those hard examples that the cloud can classify correctly. Our framework optimizes over neural architectures and trains edge predictors and routing models so that the overall accuracy remains high while minimizing the overall latency. Training a hybrid learner is difficult since we lack annotations of hard edge-examples. We introduce a novel proxy supervision in this context and show that our method adapts seamlessly and near optimally across different latency regimes. On the ImageNet dataset, our proposed method deployed on a micro-controller unit exhibits 25% reduction in latency compared to cloud-only processing while suffering no excess loss.  more » « less
Award ID(s):
1955981
PAR ID:
10447699
Author(s) / Creator(s):
Date Published:
Journal Name:
International Conference on Learning Representations
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; (, 2021 IEEE 46th Conference on Local Computer Networks (LCN))
    null (Ed.)
    In today's era of Internet of Things (IoT), where massive amounts of data are produced by IoT and other devices, edge computing has emerged as a prominent paradigm for low-latency data processing. However, applications may have diverse latency requirements: certain latency-sensitive processing operations may need to be performed at the edge, while delay-tolerant operations can be performed on the cloud, without occupying the potentially limited edge computing resources. To achieve that, we envision an environment where computing resources are distributed across edge and cloud offerings. In this paper, we present the design of CLEDGE (CLoud + EDGE), an information-centric hybrid cloud-edge framework, aiming to maximize the on-time completion of computational tasks offloaded by applications with diverse latency requirements. The design of CLEDGE is motivated by the networking challenges that mixed reality researchers face. Our evaluation demonstrates that CLEDGE can complete on-time more than 90% of offloaded tasks with modest overheads. 
    more » « less
  2. ; ; ; ; ; ; ; ; ; ; et al (, Proceedings of the ASPDAC Asia and South Pacific Design Automation Conference)
    Low-latency and low-power edge AI is crucial for Virtual Reality and Augmented Reality applications. Recent advances demonstrate that hybrid models, combining convolution layers (CNN) and transformers (ViT), often achieve a superior accuracy/performance tradeoff on various computer vision and machine learning (ML) tasks. However, hybrid ML models can present system challenges for latency and energy efficiency due to their diverse nature in dataflow and memory access patterns. In this work, we leverage architecture heterogeneity from Neural Processing Units (NPU) and Compute-In-Memory (CIM) and explore diverse execution schemas to efficiently execute these hybrid models. We introduce H4H-NAS, a two-stage Neural Architecture Search (NAS) framework to automate the design of efficient hybrid CNN/ViT models for heterogeneous edge systems featuring both NPU and CIM. We propose a two-phase incremental supernet training in our NAS framework to resolve gradient conflicts between sampled subnets caused by different types of blocks in a hybrid model search space. Our H4H-NAS approach is also powered by a performance estimator built with NPU performance results measured on real silicon, and CIM performance based on industry IPs. H4H-NAS searches hybrid CNN-ViT models with fine granularity and achieves significant (up to 1.34%) top-1 accuracy improvement on ImageNet. Moreover, results from our algorithm/hardware co-design reveal up to 56.08% overall latency and 41.72% energy improvements by introducing heterogeneous computing over baseline solutions. Overall, our framework guides the design of hybrid network architectures and system architectures for NPU+CIM heterogeneous systems. 
    more » « less
  3. ; ; ; ; ; (, IEEE)
    Point cloud is an important type of geometric data structure for many embedded applications such as autonomous driving and augmented reality. Current Point Cloud Networks (PCNs) have proven to achieve great success in using inference to perform point cloud analysis, including object part segmentation, shape classification, and so on. However, point cloud applications on the computing edge require more than just the inference step. They require an end-to-end (E2E) processing of the point cloud workloads: pre-processing of raw data, input preparation, and inference to perform point cloud analysis. Current PCN approaches to support end-to-end processing of point cloud workload cannot meet the real-time latency requirement on the edge, i.e., the ability of the AI service to keep up with the speed of raw data generation by 3D sensors. Latency for end-to-end processing of the point cloud workloads stems from two reasons: memory-intensive down-sampling in the pre-processing phase and the data structuring step for input preparation in the inference phase. In this paper, we present HgPCN, an end-to-end heterogeneous architecture for real-time embedded point cloud applications. In HgPCN, we introduce two novel methodologies based on spatial indexing to address the two identified bottlenecks. In the Pre-processing Engine of HgPCN, an Octree-Indexed-Sampling method is used to optimize the memory-intensive down-sampling bottleneck of the pre-processing phase. In the Inference Engine, HgPCN extends a commercial DLA with a customized Data Structuring Unit which is based on a Voxel-Expanded Gathering method to fundamentally reduce the workload of the data structuring step in the inference phase. The initial prototype of HgPCN has been implemented on an Intel PAC (Xeon+FPGA) platform. Four commonly available point cloud datasets were used for comparison, running on three baseline devices: Intel Xeon W-2255, Nvidia Xavier NX Jetson GPU, and Nvidia 4060ti GPU. These point cloud datasets were also run on two existing PCN accelerators for comparison: PointACC and Mesorasi. Our results show that for the inference phase, depending on the dataset size, HgPCN achieves speedup from 1.3× to 10.2× vs. PointACC, 2.2× to 16.5× vs. Mesorasi, and 6.4× to 21× vs. Jetson NX GPU. Along with optimization of the memory-intensive down-sampling bottleneck in pre-processing phase, the overall latency shows that HgPCN can reach the real-time requirement by providing end-to-end service with keeping up with the raw data generation rate. 
    more » « less
  4. ; ; ; ; ; (, ACM Transactions on Embedded Computing Systems)
    In the past decade, Deep Neural Networks (DNNs), e.g., Convolutional Neural Networks, achieved human-level performance in vision tasks such as object classification and detection. However, DNNs are known to be computationally expensive and thus hard to be deployed in real-time and edge applications. Many previous works have focused on DNN model compression to obtain smaller parameter sizes and consequently, less computational cost. Such methods, however, often introduce noticeable accuracy degradation. In this work, we optimize a state-of-the-art DNN-based video detection framework—Deep Feature Flow (DFF) from the cloud end using three proposed ideas. First, we propose Asynchronous DFF (ADFF) to asynchronously execute the neural networks. Second, we propose a Video-based Dynamic Scheduling (VDS) method that decides the detection frequency based on the magnitude of movement between video frames. Last, we propose Spatial Sparsity Inference, which only performs the inference on part of the video frame and thus reduces the computation cost. According to our experimental results, ADFF can reduce the bottleneck latency from 89 to 19 ms. VDS increases the detection accuracy by 0.6% mAP without increasing computation cost. And SSI further saves 0.2 ms with a 0.6% mAP degradation of detection accuracy. 
    more » « less
  5. ; ; (, Software practice experience)
    Serverless computing is an emerging event-driven programming model that accelerates the development and deployment of scalable web services on cloud computing systems. Though widely integrated with the public cloud, serverless computing use is nascent for edge-based, IoT deployments. In this work, we design and develop STOIC (Serverless TeleOperable HybrId Cloud), an IoT application deployment and offloading system that extends the serverless model in three ways. First, STOIC adopts a dynamic feedback control mechanism to precisely predict latency and dispatch workloads uniformly across edge and cloud systems using a distributed serverless framework. Second, STOIC leverages hardware acceleration (e.g. GPU resources) for serverless function execution when available from the underlying cloud system. Third, STOIC can be configured in multiple ways to overcome deployment variability associated with public cloud use. Finally, we empirically evaluate STOIC using real-world machine learning applications and multi-tier IoT deployments (edge and cloud). We show that STOIC can be used for training image processing workloads (for object recognition) – once thought too resource intensive for edge deployments. We find that STOIC reduces overall execution time (response latency) and achieves placement accuracy that ranges from 92% to 97%. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email