arXiv subjects

Mengwei Xu

Publications and source records attributed to Mengwei Xu.

At least 73 records · Page 4Linked to original sources

arXiv2023

Second-order optimality conditions for bilevel programs

Second-order optimality conditions of the bilevel programming problems are dependent on the second-order directional derivatives of the value functions or the solution mappings of the lower level problems under some regular conditions, which can not be calculated or evaluated. To overcome this difficulty, we propose the notion of the bi-local solution. Under the Jacobian uniqueness conditions for the lower level problem, we prove that the bi-local solution is a local minimizer of some one-level minimization problem. Basing on this property, the first-order necessary optimality conditions and second-order necessary and sufficient optimality conditions for the bi-local optimal solution of a given bilevel program are established. The second-order optimality conditions proposed here only involve second-order derivatives of the defining functions of the bilevel problem. The second-order sufficient optimality conditions are used to derive the Q-linear convergence rate of the classical augmented Lagrangian method.

math.OC↗

arXiv2023

Uncertain Machine Ethical Decisions Using Hypothetical Retrospection

We propose the use of the hypothetical retrospection argumentation procedure, developed by Sven Ove Hansson to improve existing approaches to machine ethical reasoning by accounting for probability and uncertainty from a position of Philosophy that resonates with humans. Actions are represented with a branching set of potential outcomes, each with a state, utility, and either a numeric or poetic probability estimate. Actions are chosen based on comparisons between sets of arguments favouring actions from the perspective of their branches, even those branches that led to an undesirable outcome. This use of arguments allows a variety of philosophical theories for ethical reasoning to be used, potentially in flexible combination with each other. We implement the procedure, applying consequentialist and deontological ethical theories, independently and concurrently, to an autonomous library system use case. We introduce a preliminary framework that seems to meet the varied requirements of a machine ethics system: versatility under multiple theories and a resonance with humans that enables transparency and explainability.

cs.AI↗

arXiv2023

A Comprehensive Survey on Orbital Edge Computing: Systems, Applications, and Algorithms

The number of satellites, especially those operating in low-earth orbit (LEO), is exploding in recent years. Additionally, the use of COTS hardware into those satellites enables a new paradigm of computing: orbital edge computing (OEC). OEC entails more technically advanced steps compared to single-satellite computing. This feature allows for vast design spaces with multiple parameters, rendering several novel approaches feasible. The mobility of LEO satellites in the network and limited resources of communication, computation, and storage make it challenging to design an appropriate scheduling algorithm for specific tasks in comparison to traditional ground-based edge computing. This article comprehensively surveys the significant areas of focus in orbital edge computing, which include protocol optimization, mobility management, and resource allocation. This article provides the first comprehensive survey of OEC. Previous survey papers have only concentrated on ground-based edge computing or the integration of space and ground technologies. This article presents a review of recent research from 2000 to 2023 on orbital edge computing that covers network design, computation offloading, resource allocation, performance analysis, and optimization. Moreover, having discussed several related works, both technological challenges and future directions are highlighted in the field.

cs.DC↗

arXiv2023

FedAdapter: Efficient Federated Learning for Modern NLP

Transformer-based pre-trained models have revolutionized NLP for superior performance and generality. Fine-tuning pre-trained models for downstream tasks often requires private data, for which federated learning is the de-facto approach (i.e., FedNLP). However, our measurements show that FedNLP is prohibitively slow due to the large model sizes and the resultant high network/computation cost. Towards practical FedNLP, we identify as the key building blocks adapters, small bottleneck modules inserted at a variety of model layers. A key challenge is to properly configure the depth and width of adapters, to which the training speed and efficiency is highly sensitive. No silver-bullet configuration exists: the optimal choice varies across downstream NLP tasks, desired model accuracy, and mobile resources. To automate adapter configuration, we propose FedAdapter, a framework that enhances the existing FedNLP with two novel designs. First, FedAdapter progressively upgrades the adapter configuration throughout a training session; the principle is to quickly learn shallow knowledge by only training fewer and smaller adapters at the model's top layers, and incrementally learn deep knowledge by incorporating deeper and larger adapters. Second, FedAdapter continuously profiles future adapter configurations by allocating participant devices to trial groups. Extensive experiments show that FedAdapter can reduce FedNLP's model convergence delay to no more than several hours, which is up to 155.5$\times$ faster compared to vanilla FedNLP and 48$\times$ faster compared to strong baselines.

cs.LG↗

arXiv2023

Relaxed constant positive linear dependence constraint qualification for disjunctive programs

The disjunctive system is a system involving a disjunctive set which is the union of finitely many polyhedral convex sets. In this paper, we introduce a notion of the relaxed constant positive linear dependence constraint qualification (RCPLD) for the disjunctive system. For a disjunctive system, our notion is weaker than the one we introduced for a more general system recently (J. Glob. Optim. 2020) and is still a constraint qualification. To obtain the local error bound for the disjunctive system, we introduce the piecewise RCPLD under which the error bound property holds if all inequality constraint functions are subdifferentially regular and the rest of the constraint functions are smooth. We then specialize our results to the ortho-disjunctive program, which includes the mathematical program with equilibrium constraints (MPEC), the mathematical program with vanishing constraints (MPVC) and the mathematical program with switching constraints (MPSC) as special cases. For MPEC, we recover MPEC-RCPLD, an MPEC variant of RCPLD and propose the MPEC piecewise RCPLD to obtain the {error bound property}. For MPVC, we introduce new constraint qualifications MPVC-RCPLD and the piecewise RCPLD, which also implies the local error bound. For MPSC, we show that both RCPLD and the piecewise RCPLD coincide and hence it leads to the local error bound.

math.OC↗

arXiv2022

Federated Neural Architecture Search

To preserve user privacy while enabling mobile intelligence, techniques have been proposed to train deep neural networks on decentralized data. However, training over decentralized data makes the design of neural architecture quite difficult as it already was. Such difficulty is further amplified when designing and deploying different neural architectures for heterogeneous mobile platforms. In this work, we propose an automatic neural architecture search into the decentralized training, as a new DNN training paradigm called Federated Neural Architecture Search, namely federated NAS. To deal with the primary challenge of limited on-client computational and communication resources, we present FedNAS, a highly optimized framework for efficient federated NAS. FedNAS fully exploits the key opportunity of insufficient model candidate re-training during the architecture search process, and incorporates three key optimizations: parallel candidates training on partial clients, early dropping candidates with inferior performance, and dynamic round numbers. Tested on large-scale datasets and typical CNN architectures, FedNAS achieves comparable model accuracy as state-of-the-art NAS algorithm that trains models with centralized data, and also reduces the client cost by up to two orders of magnitude compared to a straightforward design of federated NAS.

cs.LG↗

arXiv2022

Benchmarking of DL Libraries and Models on Mobile Devices

Deploying deep learning (DL) on mobile devices has been a notable trend in recent years. To support fast inference of on-device DL, DL libraries play a critical role as algorithms and hardware do. Unfortunately, no prior work ever dives deep into the ecosystem of modern DL libs and provides quantitative results on their performance. In this paper, we first build a comprehensive benchmark that includes 6 representative DL libs and 15 diversified DL models. We then perform extensive experiments on 10 mobile devices, which help reveal a complete landscape of the current mobile DL libs ecosystem. For example, we find that the best-performing DL lib is severely fragmented across different models and hardware, and the gap between those DL libs can be rather huge. In fact, the impacts of DL libs can overwhelm the optimizations from algorithms or hardware, e.g., model quantization and GPU/DSP-based heterogeneous computing. Finally, atop the observations, we summarize practical implications to different roles in the DL lib ecosystem.

cs.LG↗

arXiv2022

Mandheling: Mixed-Precision On-Device DNN Training with DSP Offloading

This paper proposes Mandheling, the first system that enables highly resource-efficient on-device training by orchestrating the mixed-precision training with on-chip Digital Signal Processing (DSP) offloading. Mandheling fully explores the advantages of DSP in integer-based numerical calculation by four novel techniques: (1) a CPU-DSP co-scheduling scheme to mitigate the overhead from DSP-unfriendly operators; (2) a self-adaptive rescaling algorithm to reduce the overhead of dynamic rescaling in backward propagation; (3) a batch-splitting algorithm to improve the DSP cache efficiency; (4) a DSP-compute subgraph reusing mechanism to eliminate the preparation overhead on DSP. We have fully implemented Mandheling and demonstrate its effectiveness through extensive experiments. The results show that, compared to the state-of-the-art DNN engines from TFLite and MNN, Mandheling reduces the per-batch training time by 5.5$\times$ and the energy consumption by 8.9$\times$ on average. In end-to-end training tasks, Mandheling reduces up to 10.7$\times$ convergence time and 13.1$\times$ energy consumption, with only 1.9%-2.7% accuracy loss compared to the FP32 precision setting.

cs.NI↗

arXiv2022

Device-centric Federated Analytics At Ease

Nowadays, high-volume and privacy-sensitive data are generated by mobile devices, which are better to be preserved on devices and queried on demand. However, data analysts still lack a uniform way to harness such distributed on-device data. In this paper, we propose a data querying system, Deck, that enables flexible device-centric federated analytics. The key idea of Deck is to bypass the app developers but allow the data analysts to directly submit their analytics code to run on devices, through a centralized query coordinator service. Deck provides a list of standard APIs to data analysts and handles most of the device-specific tasks underneath. Deck further incorporates two key techniques: (i) a hybrid permission checking mechanism and mandatory cross-device aggregation to ensure data privacy; (ii) a zero-knowledge statistical model that judiciously trades off query delay and query resource expenditure on devices. We fully implement Deck and plug it into 20 popular Android apps. An in-the-wild deployment on 1,642 volunteers shows that Deck significantly reduces the query delay by up to 30x compared to baselines. Our microbenchmarks also demonstrate that the standalone overhead of Deck is negligible.

cs.DC↗

arXiv2021

Tiansuan Constellation: An Open Research Platform

Satellite network is the first step of interstellar voyages. It can provide global Internet connectivity everywhere on earth, where most areas cannot access the Internet by the terrestrial infrastructure due to the geographic accessibility and high cost. The space industry experiences a rise in large low-earth-orbit satellite constellations to achieve universal connectivity. The research community is also urgent to do some leading research to bridge the connectivity divide. Researchers now conduct their work by simulation, which is far from enough. However, experiments on real satellites are blocked by the high threshold of space technology, such as deployment cost and unknown risks. To solve the above dilemma, we are eager to contribute to the universal connectivity and build an open research platform, Tiansuan constellation to support experiments on real satellite networks. We discuss the potential research topics that would benefit from Tiansuan constellation. We provide two case studies that have already deployed in two experimental satellites of Tiansuan constellation.

cs.DC↗

arXiv2021

Boosting Mobile CNN Inference through Semantic Memory

Human brains are known to be capable of speeding up visual recognition of repeatedly presented objects through faster memory encoding and accessing procedures on activated neurons. For the first time, we borrow and distill such a capability into a semantic memory design, namely SMTM, to improve on-device CNN inference. SMTM employs a hierarchical memory architecture to leverage the long-tail distribution of objects of interest, and further incorporates several novel techniques to put it into effects: (1) it encodes high-dimensional feature maps into low-dimensional, semantic vectors for low-cost yet accurate cache and lookup; (2) it uses a novel metric in determining the exit timing considering different layers' inherent characteristics; (3) it adaptively adjusts the cache size and semantic vectors to fit the scene dynamics. SMTM is prototyped on commodity CNN engine and runs on both mobile CPU and GPU. Extensive experiments on large-scale datasets and models show that SMTM can significantly speed up the model inference over standard approach (up to 2X) and prior cache designs (up to 1.5X), with acceptable accuracy loss.

cs.CV↗

arXiv2021

From Cloud to Edge: A First Look at Public Edge Platforms

Public edge platforms have drawn increasing attention from both academia and industry. In this study, we perform a first-of-its-kind measurement study on a leading public edge platform that has been densely deployed in China. Based on this measurement, we quantitatively answer two critical yet unexplored questions. First, from end users' perspective, what is the performance of commodity edge platforms compared to cloud, in terms of the end-to-end network delay, throughput, and the application QoE. Second, from the edge service provider's perspective, how are the edge workloads different from cloud, in terms of their VM subscription, monetary cost, and resource usage. Our study quantitatively reveals the status quo of today's public edge platforms, and provides crucial insights towards developing and operating future edge services.

cs.NI↗

arXiv2021

Observable and Attention-Directing BDI Agents for Human-Autonomy Teaming

Human-autonomy teaming (HAT) scenarios feature humans and autonomous agents collaborating to meet a shared goal. For effective collaboration, the agents must be transparent and able to share important information about their operation with human teammates. We address the challenge of transparency for Belief-Desire-Intention agents defined in the Conceptual Agent Notation (CAN) language. We extend the semantics to model agents that are observable (i.e. the internal state of tasks is available), and attention-directing (i.e. specific states can be flagged to users), and provide an executable semantics via an encoding in Milner's bigraphs. Using an example of unmanned aerial vehicles, the BigraphER tool, and PRISM, we show and verify how the extensions work in practice.

cs.MA↗

arXiv2021

Video Analytics with Zero-streaming Cameras

Low-cost cameras enable powerful analytics. An unexploited opportunity is that most captured videos remain "cold" without being queried. For efficiency, we advocate for these cameras to be zero streaming: capturing videos to local storage and communicating with the cloud only when analytics is requested. How to query zero-streaming cameras efficiently? Our response is a camera/cloud runtime system called DIVA. It addresses two key challenges: to best use limited camera resource during video capture; to rapidly explore massive videos during query execution. DIVA contributes two unconventional techniques. (1) When capturing videos, a camera builds sparse yet accurate landmark frames, from which it learns reliable knowledge for accelerating future queries. (2) When executing a query, a camera processes frames in multiple passes with increasingly more expensive operators. As such, DIVA presents and keeps refining inexact query results throughout the query's execution. On diverse queries over 15 videos lasting 720 hours in total, DIVA runs at more than 100x video realtime and outperforms competitive alternative designs. To our knowledge, DIVA is the first system for querying large videos stored on low-cost remote cameras.

cs.DB↗

arXiv2021

Modelling and Verifying BDI Agents with Bigraphs

The Belief-Desire-Intention (BDI) architecture is a popular framework for rational agents; most verification approaches are based on reasoning about implementations of BDI programming languages. We investigate an alternative approach based on reasoning about BDI agent semantics, through a model of the execution of an agent program. We employ Milner's bigraphs as the modelling framework and present an encoding for the Conceptual Agent Notation (CAN) language - a superset of AgentSpeak featuring declarative goals, concurrency, and failure recovery. We provide an encoding of the syntax and semantics of CAN agents, and give a rigorous proof that the encoding is faithful. Verification is based on the use of mainstream software tools including BigraphER, and a small case study verifying several properties of Unmanned Aerial Vehicles (UAVs) illustrates the framework in action. The executable framework is a foundational step that will enable more advanced reasoning such as plan preference, intention priorities and trade-offs, and interactions with an environment under uncertainty.

cs.PL↗

arXiv2021

Characterizing Impacts of Heterogeneity in Federated Learning upon Large-Scale Smartphone Data

Federated learning (FL) is an emerging, privacy-preserving machine learning paradigm, drawing tremendous attention in both academia and industry. A unique characteristic of FL is heterogeneity, which resides in the various hardware specifications and dynamic states across the participating devices. Theoretically, heterogeneity can exert a huge influence on the FL training process, e.g., causing a device unavailable for training or unable to upload its model updates. Unfortunately, these impacts have never been systematically studied and quantified in existing FL literature. In this paper, we carry out the first empirical study to characterize the impacts of heterogeneity in FL. We collect large-scale data from 136k smartphones that can faithfully reflect heterogeneity in real-world settings. We also build a heterogeneity-aware FL platform that complies with the standard FL protocol but with heterogeneity in consideration. Based on the data and the platform, we conduct extensive experiments to compare the performance of state-of-the-art FL algorithms under heterogeneity-aware and heterogeneity-unaware settings. Results show that heterogeneity causes non-trivial performance degradation in FL, including up to 9.2% accuracy drop, 2.32x lengthened training time, and undermined fairness. Furthermore, we analyze potential impact factors and find that device failure and participant bias are two potential factors for performance degradation. Our study provides insightful implications for FL practitioners. On the one hand, our findings suggest that FL algorithm designers consider necessary heterogeneity during the evaluation. On the other hand, our findings urge system providers to design specific mechanisms to mitigate the impacts of heterogeneity.

cs.LG↗

arXiv2021

DeepWear: Adaptive Local Offloading for On-Wearable Deep Learning

Due to their on-body and ubiquitous nature, wearables can generate a wide range of unique sensor data creating countless opportunities for deep learning tasks. We propose DeepWear, a deep learning (DL) framework for wearable devices to improve the performance and reduce the energy footprint. DeepWear strategically offloads DL tasks from a wearable device to its paired handheld device through local network. Compared to the remote-cloud-based offloading, DeepWear requires no Internet connectivity, consumes less energy, and is robust to privacy breach. DeepWear provides various novel techniques such as context-aware offloading, strategic model partition, and pipelining support to efficiently utilize the processing capacity from nearby paired handhelds. Deployed as a user-space library, DeepWear offers developer-friendly APIs that are as simple as those in traditional DL libraries such as TensorFlow. We have implemented DeepWear on the Android OS and evaluated it on COTS smartphones and smartwatches with real DL models. DeepWear brings up to 5.08X and 23.0X execution speedup, as well as 53.5% and 85.5% energy saving compared to wearable-only and handheld-only strategies, respectively.

cs.CY↗

arXiv2021

A First Look at Deep Learning Apps on Smartphones

We are in the dawn of deep learning explosion for smartphones. To bridge the gap between research and practice, we present the first empirical study on 16,500 the most popular Android apps, demystifying how smartphone apps exploit deep learning in the wild. To this end, we build a new static tool that dissects apps and analyzes their deep learning functions. Our study answers threefold questions: what are the early adopter apps of deep learning, what do they use deep learning for, and how do their deep learning models look like. Our study has strong implications for app developers, smartphone vendors, and deep learning R\&D. On one hand, our findings paint a promising picture of deep learning for smartphones, showing the prosperity of mobile deep learning frameworks as well as the prosperity of apps building their cores atop deep learning. On the other hand, our findings urge optimizations on deep learning models deployed on smartphones, the protection of these models, and validation of research ideas on these models.

cs.LG↗