Showing papers on "Task (computing) published in 2020"

Language Models are Few-Shot Learners

Abstract: Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from only a few examples or from simple instructions - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting. For all tasks, GPT-3 is applied without any gradient updates or fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic. At the same time, we also identify some datasets where GPT-3's few-shot learning still struggles, as well as some datasets where GPT-3 faces methodological issues related to training on large web corpora. Finally, we find that GPT-3 can generate samples of news articles which human evaluators have difficulty distinguishing from articles written by humans. We discuss broader societal impacts of this finding and of GPT-3 in general.

Language Models are Few-Shot Learners

Abstract: Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from only a few examples or from simple instructions - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting. For all tasks, GPT-3 is applied without any gradient updates or fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic. At the same time, we also identify some datasets where GPT-3's few-shot learning still struggles, as well as some datasets where GPT-3 faces methodological issues related to training on large web corpora. Finally, we find that GPT-3 can generate samples of news articles which human evaluators have difficulty distinguishing from articles written by humans. We discuss broader societal impacts of this finding and of GPT-3 in general.

12-in-1: Multi-Task Vision and Language Representation Learning

Abstract: Much of vision-and-language research focuses on a small but diverse set of independent tasks and supporting datasets often studied in isolation; however, the visually-grounded language understanding skills required for success at these tasks overlap significantly. In this work, we investigate these relationships between vision-and-language tasks by developing a large-scale, multi-task model. Our approach culminates in a single model on 12 datasets from four broad categories of task including visual question answering, caption-based image retrieval, grounding referring expressions, and multimodal verification. Compared to independently trained single-task models, this represents a reduction from approximately 3 billion parameters to 270 million while simultaneously improving performance by 2.05 points on average across tasks. We use our multi-task framework to perform in-depth analysis of the effect of joint training diverse tasks. Further, we show that finetuning task-specific models from our single multi-task model can lead to further improvements, achieving performance at or above the state-of-the-art.

AdapterFusion: Non-Destructive Task Composition for Transfer Learning

Abstract: Sequential fine-tuning and multi-task learning are methods aiming to incorporate knowledge from multiple tasks; however, they suffer from catastrophic forgetting and difficulties in dataset balancing. To address these shortcomings, we propose AdapterFusion, a new two stage learning algorithm that leverages knowledge from multiple tasks. First, in the knowledge extraction stage we learn task specific parameters called adapters, that encapsulate the task-specific information. We then combine the adapters in a separate knowledge composition step. We show that by separating the two stages, i.e., knowledge extraction and knowledge composition, the classifier can effectively exploit the representations learned from multiple tasks in a non-destructive manner. We empirically evaluate AdapterFusion on 16 diverse NLU tasks, and find that it effectively combines various types of knowledge at different layers of the model. We show that our approach outperforms traditional strategies such as full fine-tuning as well as multi-task learning. Our code and adapters are available at this http URL.

Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere

Abstract: Contrastive representation learning has been outstandingly successful in practice. In this work, we identify two key properties related to the contrastive loss: (1) alignment (closeness) of features from positive pairs, and (2) uniformity of the induced distribution of the (normalized) features on the hypersphere. We prove that, asymptotically, the contrastive loss optimizes these properties, and analyze their positive effects on downstream tasks. Empirically, we introduce an optimizable metric to quantify each property. Extensive experiments on standard vision and language datasets confirm the strong agreement between both metrics and downstream task performance. Remarkably, directly optimizing for these two metrics leads to representations with comparable or better performance at downstream tasks than contrastive learning. Project Page: https://ssnl.github.io/hypersphere Code: https://github.com/SsnL/align_uniform

SpineNet: Learning Scale-Permuted Backbone for Recognition and Localization

Abstract: Convolutional neural networks typically encode an input image into a series of intermediate features with decreasing resolutions. While this structure is suited to classification tasks, it does not perform well for tasks requiring simultaneous recognition and localization (e.g., object detection). The encoder-decoder architectures are proposed to resolve this by applying a decoder network onto a backbone model designed for classification tasks. In this paper, we argue encoder-decoder architecture is ineffective in generating strong multi-scale features because of the scale-decreased backbone. We propose SpineNet, a backbone with scale-permuted intermediate features and cross-scale connections that is learned on an object detection task by Neural Architecture Search. Using similar building blocks, SpineNet models outperform ResNet-FPN models by 3%+ AP at various scales while using 10-20% fewer FLOPs. In particular, SpineNet-190 achieves 52.1% AP on COCO, attaining the new state-of-the-art performance for single model object detection without test-time augmentation. SpineNet can transfer to classification tasks, achieving 5% top-1 accuracy improvement on a challenging iNaturalist fine-grained dataset. Code is at: https://github.com/tensorflow/tpu/tree/master/models/official/detection.

Autopilot: workload autoscaling at Google

Abstract: In many public and private Cloud systems, users need to specify a limit for the amount of resources (CPU cores and RAM) to provision for their workloads. A job that exceeds its limits might be throttled or killed, resulting in delaying or dropping end-user requests, so human operators naturally err on the side of caution and request a larger limit than the job needs. At scale, this results in massive aggregate resource wastage. To address this, Google uses Autopilot to configure resources automatically, adjusting both the number of concurrent tasks in a job (horizontal scaling) and the CPU/memory limits for individual tasks (vertical scaling). Autopilot walks the same fine line as human operators: its primary goal is to reduce slack - the difference between the limit and the actual resource usage - while minimizing the risk that a task is killed with an out-of-memory (OOM) error or its performance degraded because of CPU throttling. Autopilot uses machine learning algorithms applied to historical data about prior executions of a job, plus a set of finely-tuned heuristics, to walk this line. In practice, Autopiloted jobs have a slack of just 23%, compared with 46% for manually-managed jobs. Additionally, Autopilot reduces the number of jobs severely impacted by OOMs by a factor of 10. Despite its advantages, ensuring that Autopilot was widely adopted took significant effort, including making potential recommendations easily visible to customers who had yet to opt in, automatically migrating certain categories of jobs, and adding support for custom recommenders. At the time of writing, Autopiloted jobs account for over 48% of Google's fleet-wide resource usage.

RLBench: The Robot Learning Benchmark & Learning Environment

Abstract: We present a challenging new benchmark and learning-environment for robot learning: RLBench. The benchmark features 100 completely unique, hand-designed tasks, ranging in difficulty from simple target reaching and door opening to longer multi-stage tasks, such as opening an oven and placing a tray in it. We provide an array of both proprioceptive observations and visual observations, which include rgb, depth, and segmentation masks from an over-the-shoulder stereo camera and an eye-in-hand monocular camera. Uniquely, each task comes with an infinite supply of demos through the use of motion planners operating on a series of waypoints given during task creation time; enabling an exciting flurry of demonstration-based learning possibilities. RLBench has been designed with scalability in mind; new tasks, along with their motion-planned demos, can be easily created and then verified by a series of tools, allowing users to submit their own tasks to the RLBench task repository. This large-scale benchmark aims to accelerate progress in a number of vision-guided manipulation research areas, including: reinforcement learning, imitation learning, multi-task learning, geometric computer vision, and in particular, few-shot learning. With the benchmark's breadth of tasks and demonstrations, we propose the first large-scale few-shot challenge in robotics. We hope that the scale and diversity of RLBench offers unparalleled research opportunities in the robot learning community and beyond. Benchmarking code and videos can be found at https://sites.google.com/view/rlbench .

Towards Learning a Generic Agent for Vision-and-Language Navigation via Pre-training

Abstract: Learning to navigate in a visual environment following natural-language instructions is a challenging task, because the multimodal inputs to the agent are highly variable, and the training data on a new task is often limited. In this paper, we present the first pre-training and fine-tuning paradigm for vision-and-language navigation (VLN) tasks. By training on a large amount of image-text-action triplets in a self-supervised learning manner, the pre-trained model provides generic representations of visual environments and language instructions. It can be easily used as a drop-in for existing VLN frameworks, leading to the proposed agent called Prevalent. It learns more effectively in new tasks and generalizes better in a previously unseen environment. The performance is validated on three VLN tasks. On the Room-to-Room benchmark, our model improves the state-of-the-art from 47% to 51% on success rate weighted by path length. Further, the learned representation is transferable to other VLN tasks. On two recent tasks, vision-and-dialog navigation and "Help, Anna!" the proposed Prevalent leads to significant improvement over existing methods, achieving a new state of the art.

Deep reinforcement learning for dynamic computation offloading and resource allocation in cache-assisted mobile edge computing systems

Abstract: Mobile Edge Computing (MEC) is one of the most promising techniques for next-generation wireless communication systems. In this paper, we study the problem of dynamic caching, computation offloading, and resource allocation in cache-assisted multi-user MEC systems with stochastic task arrivals. There are multiple computationally intensive tasks in the system, and each Mobile User (MU) needs to execute a task either locally or remotely in one or more MEC servers by offloading the task data. Popular tasks can be cached in MEC servers to avoid duplicates in offloading. The cached contents can be either obtained through user offloading, fetched from a remote cloud, or fetched from another MEC server. The objective is to minimize the long-term average of a cost function, which is defined as a weighted sum of energy consumption, delay, and cache contents' fetching costs. The weighting coefficients associated with the different metrics in the objective function can be adjusted to balance the tradeoff among them. The optimum design is performed with respect to four decision parameters: whether to cache a given task, whether to offload a given uncached task, how much transmission power should be used during offloading, and how much MEC resources to be allocated for executing a task. We propose to solve the problems by developing a dynamic scheduling policy based on Deep Reinforcement Learning (DRL) with the Deep Deterministic Policy Gradient (DDPG) method. A new decentralized DDPG algorithm is developed to obtain the optimum designs for multi-cell MEC systems by leveraging on the cooperations among neighboring MEC servers. Simulation results demonstrate that the proposed algorithm outperforms other existing strategies, such as Deep Q-Network (DQN).

Near-Optimal and Truthful Online Auction for Computation Offloading in Green Edge-Computing Systems

Abstract: Utilizing the intelligence at the network edge, edge computing paradigm emerges to provide time-sensitive computing services for Internet of Things. In this paper, we investigate sustainable computation offloading in an edge-computing system that consists of energy harvesting-enabled mobile devices (MDs) and a dispatcher. The dispatcher collects computation tasks generated by IoT devices with limited computation power, and offloads them to resourceful MDs in exchange for rewards. We propose an online Rewards-optimal Auction (RoA) to optimize the long-term sum-of-rewards for processing offloaded tasks, meanwhile adapting to the highly dynamic energy harvesting (EH) process and computation task arrivals. RoA is designed based on Lyapunov optimization and Vickrey-Clarke-Groves auction, the operation of which does not require a prior knowledge of the energy harvesting, task arrivals, or wireless channel statistics. Our analytical results confirm the optimality of tasks assignment. Furthermore, simulation results validate the analytical analysis, and verify the efficacy of the proposed RoA.

PolyLaneNet: Lane Estimation via Deep Polynomial Regression

Abstract: One of the main factors that contributed to the large advances in autonomous driving is the advent of deep learning. For safer self-driving vehicles, one of the problems that has yet to be solved completely is lane detection. Since methods for this task have to work in real-time (+30 FPS), they not only have to be effective (i.e., have high accuracy) but they also have to be efficient (i.e., fast). In this work, we present a novel method for lane detection that uses as input an image from a forward-looking camera mounted in the vehicle and outputs polynomials representing each lane marking in the image, via deep polynomial regression. The proposed method is shown to be competitive with existing state-of-the-art methods in the TuSimple dataset while maintaining its efficiency (115 FPS). Additionally, extensive qualitative results on two additional public datasets are presented, alongside with limitations in the evaluation metrics used by recent works for lane detection. Finally, we provide source code and trained models that allow others to replicate all the results shown in this paper, which is surprisingly rare in state-of-the-art lane detection methods. The full source code and pretrained models are available at this https URL.

Learning to Few-Shot Learn Across Diverse Natural Language Classification Tasks

Abstract: Pre-trained transformer models have shown enormous success in improving performance on several downstream tasks. However, fine-tuning on a new task still requires large amounts of task-specific labeled data to achieve good performance. We consider this problem of learning to generalize to new tasks, with a few examples, as a meta-learning problem. While meta-learning has shown tremendous progress in recent years, its application is still limited to simulated problems or problems with limited diversity across tasks. We develop a novel method, LEOPARD, which enables optimization-based meta-learning across tasks with a different number of classes, and evaluate different methods on generalization to diverse NLP classification tasks. LEOPARD is trained with the state-of-the-art transformer architecture and shows better generalization to tasks not seen at all during training, with as few as 4 examples per label. Across 17 NLP tasks, including diverse domains of entity typing, natural language inference, sentiment analysis, and several other text classification tasks, we show that LEOPARD learns better initial parameters for few-shot learning than self-supervised pre-training or multi-task training, outperforming many strong baselines, for example, yielding 14.6% average relative gain in accuracy on unseen tasks with only 4 examples per label.

Deep Reinforcement Learning for Task Offloading in Mobile Edge Computing Systems

Abstract: In mobile edge computing systems, an edge node may have a high load when a large number of mobile devices offload their tasks to it. Those offloaded tasks may experience large processing delay or even be dropped when their deadlines expire. Due to the uncertain load dynamics at the edge nodes, it is challenging for each device to determine its offloading decision (i.e., whether to offload or not, and which edge node it should offload its task to) in a decentralized manner. In this work, we consider non-divisible and delay-sensitive tasks as well as edge load dynamics, and formulate a task offloading problem to minimize the expected long-term cost. We propose a model-free deep reinforcement learning-based distributed algorithm, where each device can determine its offloading decision without knowing the task models and offloading decision of other devices. To improve the estimation of the long-term cost in the algorithm, we incorporate the long short-term memory (LSTM), dueling deep Q-network (DQN), and double-DQN techniques. Simulation results show that our proposed algorithm can better exploit the processing capacities of the edge nodes and significantly reduce the ratio of dropped tasks and average delay when compared with several existing algorithms.

Deep Reinforcement Learning for Task Offloading in Mobile Edge Computing Systems

Abstract: In mobile edge computing systems, an edge node may have a high load when a large number of mobile devices offload their tasks to it. Those offloaded tasks may experience large processing delay or even be dropped when their deadlines expire. Due to the uncertain load dynamics at the edge nodes, it is challenging for each device to determine its offloading decision (i.e., whether to offload or not, and which edge node it should offload its task to) in a decentralized manner. In this work, we consider non-divisible and delay-sensitive tasks as well as edge load dynamics, and formulate a task offloading problem to minimize the expected long-term cost. We propose a model-free deep reinforcement learning-based distributed algorithm, where each device can determine its offloading decision without knowing the task models and offloading decision of other devices. To improve the estimation of the long-term cost in the algorithm, we incorporate the long short-term memory (LSTM), dueling deep Q-network (DQN), and double-DQN techniques. Simulation results with 50 mobile devices and five edge nodes show that the proposed algorithm can reduce the ratio of dropped tasks and average task delay by 86.4%-95.4% and 18.0%-30.1%, respectively, when compared with several existing algorithms.

CC2Vec: Distributed Representations of Code Changes

Abstract: Existing work on software patches often use features specific to a single task. These works often rely on manually identified features, and human effort is required to identify these features for each task. In this work, we propose CC2Vec, a neural network model that learns a representation of code changes guided by their accompanying log messages, which represent the semantic intent of the code changes. CC2Vec models the hierarchical structure of a code change with the help of the attention mechanism and uses multiple comparison functions to identify the differences between the removed and added code. To evaluate if CC2Vec can produce a distributed representation of code changes that is general and useful for multiple tasks on software patches, we use the vectors produced by CC2Vec for three tasks: log message generation, bug fixing patch identification, and just-in-time defect prediction. In all tasks, the models using CC2Vec outperform the state-of-the-art techniques.

DLT-Net: Joint Detection of Drivable Areas, Lane Lines, and Traffic Objects

Abstract: Perception is an essential task for self-driving cars, but most perception tasks are usually handled independently. We propose a unified neural network named DLT-Net to detect drivable areas, lane lines, and traffic objects simultaneously. These three tasks are most important for autonomous driving, especially when a high-definition map and accurate localization are unavailable. Instead of separating tasks in the decoder, we construct context tensors between sub-task decoders to share designate influence among tasks. Therefore, each task can benefit from others during multi-task learning. Experiments show that our model outperforms the conventional multi-task network in terms of the task-wise accuracy and the overall computational efficiency, in the challenging BDD dataset.

CC2Vec: distributed representations of code changes

Abstract: Existing work on software patches often use features specific to a single task. These works often rely on manually identified features, and human effort is required to identify these features for each task. In this work, we propose CC2Vec, a neural network model that learns a representation of code changes guided by their accompanying log messages, which represent the semantic intent of the code changes. CC2Vec models the hierarchical structure of a code change with the help of the attention mechanism and uses multiple comparison functions to identify the differences between the removed and added code. To evaluate if CC2Vec can produce a distributed representation of code changes that is general and useful for multiple tasks on software patches, we use the vectors produced by CC2Vec for three tasks: log message generation, bug fixing patch identification, and just-in-time defect prediction. In all tasks, the models using CC2Vec outperform the state-of-the-art techniques.

Dynamic Feature Integration for Simultaneous Detection of Salient Object, Edge, and Skeleton

Abstract: Salient object segmentation, edge detection, and skeleton extraction are three contrasting low-level pixel-wise vision problems, where existing works mostly focused on designing tailored methods for each individual task. However, it is inconvenient and inefficient to store a pre-trained model for each task and perform multiple different tasks in sequence. There are methods that solve specific related tasks jointly but require datasets with different types of annotations supported at the same time. In this article, we first show some similarities shared by these tasks and then demonstrate how they can be leveraged for developing a unified framework that can be trained end-to-end. In particular, we introduce a selective integration module that allows each task to dynamically choose features at different levels from the shared backbone based on its own characteristics. Furthermore, we design a task-adaptive attention module, aiming at intelligently allocating information for different tasks according to the image content priors. To evaluate the performance of our proposed network on these tasks, we conduct exhaustive experiments on multiple representative datasets. We will show that though these tasks are naturally quite different, our network can work well on all of them and even perform better than current single-purpose state-of-the-art methods. In addition, we also conduct adequate ablation analyses that provide a full understanding of the design principles of the proposed framework.

Multi-Task Reinforcement Learning with Soft Modularization

Abstract: Multi-task learning is a very challenging problem in reinforcement learning. While training multiple tasks jointly allow the policies to share parameters across different tasks, the optimization problem becomes non-trivial: It remains unclear what parameters in the network should be reused across tasks, and how the gradients from different tasks may interfere with each other. Thus, instead of naively sharing parameters across tasks, we introduce an explicit modularization technique on policy representation to alleviate this optimization issue. Given a base policy network, we design a routing network which estimates different routing strategies to reconfigure the base network for each task. Instead of directly selecting routes for each task, our task-specific policy uses a method called soft modularization to softly combine all the possible routes, which makes it suitable for sequential tasks. We experiment with various robotics manipulation tasks in simulation and show our method improves both sample efficiency and performance over strong baselines by a large margin.

Parameter-Efficient Transfer from Sequential Behaviors for User Modeling and Recommendation

Abstract: Inductive transfer learning has had a big impact on computer vision and NLP domains but has not been used in the area of recommender systems. Even though there has been a large body of research on generating recommendations based on modeling user-item interaction sequences, few of them attempt to represent and transfer these models for serving downstream tasks where only limited data exists. In this paper, we delve on the task of effectively learning a single user representation that can be applied to a diversity of tasks, from cross-domain recommendations to user profile predictions. Fine-tuning a large pre-trained network and adapting it to downstream tasks is an effective way to solve such tasks. However, fine-tuning is parameter inefficient considering that an entire model needs to be re-trained for every new task. To overcome this issue, we develop a parameter-efficient transfer learning architecture, termed as PeterRec, which can be configured on-the-fly to various downstream tasks. Specifically, PeterRec allows the pre-trained parameters to remain unaltered during fine-tuning by injecting a series of re-learned neural networks, which are small but as expressive as learning the entire network. We perform extensive experimental ablation to show the effectiveness of the learned user representation in five downstream tasks. Moreover, we show that PeterRec performs efficient transfer learning in multiple domains, where it achieves comparable or sometimes better performance relative to fine-tuning the entire model parameters. Codes and datasets are available at https://github.com/fajieyuan/sigir2020_peterrec.

A Crowd-Sensing Framework for Allocation of Time-Constrained and Location-Based Tasks

Abstract: Thanks to the capabilities of the built-in sensors of smart devices, mobile crowd-sensing (MCS) has become a promising technique for massive data collection. In this paradigm, the service provider recruits workers (i.e., common people with smart devices) to perform sensing tasks requested by the consumers. To efficiently handle workers’ recruitment and task allocation, several factors have to be considered such as the quality of the sensed data that the workers can deliver and the different tasks locations. This allocation becomes even more challenging when the MCS tries to efficiently allocate multiple tasks under limited budget, time constraints, and the uncertainty that selected workers will not be able to perform the tasks. In this paper, we propose a service computing framework for time constrained-task allocation in location based crowd-sensing systems. This framework relies on (1) a recruitment algorithm that implements a multi-objective task allocation algorithm based on Particle Swarm Optimization, (2) queuing schemes to handle efficiently the incoming sensing tasks in the server side and at the end-user side, (3) a task delegation mechanism to avoid delaying or declining the sensing requests due to unforeseen user context, and (4) a reputation management component to manage the reputation of users based on their sensing activities and task delegation. The platform goal is to efficiently determine the most appropriate set of workers to assign to each incoming task so that high quality results are returned within the requested response time. Simulations are conducted using real datasets from Foursquare1 and Enron email social network.2 Simulation results show that the proposed framework maximizes the aggregated quality of information, reduces the budget and response time to perform a task and increases the average recommenders’ reputation and their payment.

Parameter-Efficient Transfer from Sequential Behaviors for User Modeling and Recommendation

Abstract: Inductive transfer learning has had a big impact on computer vision and NLP domains but has not been used in the area of recommender systems. Even though there has been a large body of research on generating recommendations based on modeling user-item interaction sequences, few of them attempt to represent and transfer these models for serving downstream tasks where only limited data exists. In this paper, we delve on the task of effectively learning a single user representation that can be applied to a diversity of tasks, from cross-domain recommendations to user profile predictions. Fine-tuning a large pre-trained network and adapting it to downstream tasks is an effective way to solve such tasks. However, fine-tuning is parameter inefficient considering that an entire model needs to be re-trained for every new task. To overcome this issue, we develop a parameter efficient transfer learning architecture, termed as PeterRec, which can be configured on-the-fly to various downstream tasks. Specifically, PeterRec allows the pre-trained parameters to remain unaltered during fine-tuning by injecting a series of re-learned neural networks, which are small but as expressive as learning the entire network. We perform extensive experimental ablation to show the effectiveness of the learned user representation in five downstream tasks. Moreover, we show that PeterRec performs efficient transfer learning in multiple domains, where it achieves comparable or sometimes better performance relative to fine-tuning the entire model parameters. Codes and datasets are available at this https URL.

Offloading Dependent Tasks in Mobile Edge Computing with Service Caching

Abstract: In Mobile Edge Computing (MEC), many tasks require specific service support for execution and in addition, have a dependent order of execution among the tasks. However, previous works often ignore the impact of having limited services cached at the edge nodes on (dependent) task offloading, thus may lead to an infeasible offloading decision or a longer completion time. To bridge the gap, this paper studies how to efficiently offload dependent tasks to edge nodes with limited (and predetermined) service caching. We formally define the problem of offloading dependent tasks with service caching (ODT-SC), and prove that there exists no algorithm with constant approximation for this hard problem. Then, we design an efficient convex programming based algorithm (CP) to solve this problem. Moreover, we study a special case with a homogeneous MEC and propose a favorite successor based algorithm (FS) to solve this special case with a competitive ratio of O(1). Extensive simulation results using Google data traces show that our proposed algorithms can significantly reduce applications’ completion time by about 27-51% compared with other alternatives.

Dynamic Feature Integration for Simultaneous Detection of Salient Object, Edge and Skeleton.

Abstract: In this paper, we solve three low-level pixel-wise vision problems, including salient object segmentation, edge detection, and skeleton extraction, within a unified framework. We first show some similarities shared by these tasks and then demonstrate how they can be leveraged for developing a unified framework that can be trained end-to-end. In particular, we introduce a selective integration module that allows each task to dynamically choose features at different levels from the shared backbone based on its own characteristics. Furthermore, we design a task-adaptive attention module, aiming at intelligently allocating information for different tasks according to the image content priors. To evaluate the performance of our proposed network on these tasks, we conduct exhaustive experiments on multiple representative datasets. We will show that though these tasks are naturally quite different, our network can work well on all of them and even perform better than current single-purpose state-of-the-art methods. In addition, we also conduct adequate ablation analyses that provide a full understanding of the design principles of the proposed framework. To facilitate future research, source code will be released.

Task Scheduling Algorithm Based on Improved Firework Algorithm in Fog Computing

Abstract: As an emerging computing model close to the end-user, fog computing can move tasks from the cloud to the fog device to process, and make up for the lack of cloud computing in terms of mobile support, delay and location-aware. Due to the less computing resources and weak processing ability of fog devices, the lightweight computing has been the primary choice, higher requirements are imposed for time and load. Therefore, in the Cloud-fog computing architecture, a new task scheduling method(I-FASC) is proposed for the characteristics of tasks and resources, including an improved firework algorithm (I-FA) is proposed by introducing the explosion radius detection mechanism of fireworks. By two sets of experiments show this method can better reduce the processing time of the task and ensure better overall load balancing of fog devices in a cloud-fog computing system.

Adaptive Offloading for Time-Critical Tasks in Heterogeneous Internet of Vehicles

Abstract: With the recent development of wireless communication, sensing, and computing technologies, Internet of Vehicles (IoV) has attracted great attention in both academia and industry. Nevertheless, it is challenging to process time-critical tasks due to unique characteristics of IoV, including heterogeneous computation and communication capacities of network nodes, intermittent wireless connections, unevenly distributed workload, massive data transmission, intensive computation demands, and high mobility of vehicles. In this article, we propose a two-layer vehicular fog computing (VFC) architecture to explore the synergistic effect of the cloud, the static fog, and the mobile fog on processing time-critical tasks in IoV. Then, we give a motivational case study by implementing a prototype of a traffic abnormity detection and warning system, which demonstrates the necessity and urgency of developing adaptive task offloading mechanisms in such a scenario and gives insight into the problem formulation. Furthermore, we formulate the offloading model, aiming at maximizing the completion ratio of time-critical tasks. On this basis, we propose an adaptive task offloading algorithm (ATOA). Specifically, it adaptively categorizes all tasks into four types of pending lists by considering the dynamic requirements and resource constraints, and then tasks in each list will be cooperatively offloaded to different nodes based on their features. Finally, we build the simulation model and give a comprehensive performance evaluation. The results demonstrate the superiority of ATOA.

POST: Parallel Offloading of Splittable Tasks in Heterogeneous Fog Networks

Abstract: Fog computing has been promoted to support delay-sensitive applications in future Internet of Things (IoT). For a general heterogeneous fog network consisting of many dispersive fog nodes (FNs), it may well happen that some of them have delay-sensitive tasks to process, i.e., task nodes (TNs), and some have spare resources to help the TNs to process tasks, i.e., helper nodes (HNs). It remains a fundamental challenge to effectively map multiple tasks or TNs into multiple HNs to minimize every task’s service delay in a distributed manner, i.e., the multitask multihelper (MTMH) problem. The problem becomes more challenging as tasks are splittable, i.e., tasks can be divided into multiple subtasks and offloaded to multiple HNs to further reduce the service delay via the scheme similar to distributed computing, because it introduces the more complicated task division problem which results in a much larger and more complex solution space. To tackle this challenge, in this article, a generalized Nash equilibrium problem (GNEP), called parallel offloading of splittable tasks (POST), is formulated and studied thoroughly. The structural properties of the problem are characterized and thus the existence of generalized Nash equilibrium (GNE) is proven via the fixed-point theorem. Furthermore, the corresponding distributed task offloading algorithm is developed via the Gauss–Seidel-type method. The simulation results show that the proposed POST algorithm can offer much better performance in terms of the system average delay, individual delay, delay reduction ratio (DRR), and number of beneficial TNs, compared with the existing solution to the counterpart problem for nonsplittable tasks.