arXiv subjects

Rui Hu

Publications and source records attributed to Rui Hu.

At least 127 records · Page 7Linked to original sources

arXiv2021

Federated Learning with Sparsification-Amplified Privacy and Adaptive Optimization

Federated learning (FL) enables distributed agents to collaboratively learn a centralized model without sharing their raw data with each other. However, data locality does not provide sufficient privacy protection, and it is desirable to facilitate FL with rigorous differential privacy (DP) guarantee. Existing DP mechanisms would introduce random noise with magnitude proportional to the model size, which can be quite large in deep neural networks. In this paper, we propose a new FL framework with sparsification-amplified privacy. Our approach integrates random sparsification with gradient perturbation on each agent to amplify privacy guarantee. Since sparsification would increase the number of communication rounds required to achieve a certain target accuracy, which is unfavorable for DP guarantee, we further introduce acceleration techniques to help reduce the privacy cost. We rigorously analyze the convergence of our approach and utilize Renyi DP to tightly account the end-to-end DP guarantee. Extensive experiments on benchmark datasets validate that our approach outperforms previous differentially-private FL approaches in both privacy guarantee and communication efficiency.

cs.LG↗

arXiv2021

MultiXNet: Multiclass Multistage Multimodal Motion Prediction

One of the critical pieces of the self-driving puzzle is understanding the surroundings of a self-driving vehicle (SDV) and predicting how these surroundings will change in the near future. To address this task we propose MultiXNet, an end-to-end approach for detection and motion prediction based directly on lidar sensor data. This approach builds on prior work by handling multiple classes of traffic actors, adding a jointly trained second-stage trajectory refinement step, and producing a multimodal probability distribution over future actor motion that includes both multiple discrete traffic behaviors and calibrated continuous position uncertainties. The method was evaluated on large-scale, real-world data collected by a fleet of SDVs in several cities, with the results indicating that it outperforms existing state-of-the-art approaches.

cs.CV↗

arXiv2021

PolyTransform: Deep Polygon Transformer for Instance Segmentation

In this paper, we propose PolyTransform, a novel instance segmentation algorithm that produces precise, geometry-preserving masks by combining the strengths of prevailing segmentation approaches and modern polygon-based methods. In particular, we first exploit a segmentation network to generate instance masks. We then convert the masks into a set of polygons that are then fed to a deforming network that transforms the polygons such that they better fit the object boundaries. Our experiments on the challenging Cityscapes dataset show that our PolyTransform significantly improves the performance of the backbone instance segmentation network and ranks 1st on the Cityscapes test-set leaderboard. We also show impressive gains in the interactive annotation setting. We release the code at https://github.com/uber-research/PolyTransform.

cs.CV↗

arXiv2020

Multi-Task Multi-Sensor Fusion for 3D Object Detection

In this paper we propose to exploit multiple related tasks for accurate multi-sensor 3D object detection. Towards this goal we present an end-to-end learnable architecture that reasons about 2D and 3D object detection as well as ground estimation and depth completion. Our experiments show that all these tasks are complementary and help the network learn better representations by fusing information at various levels. Importantly, our approach leads the KITTI benchmark on 2D, 3D and BEV object detection, while being real time.

cs.CV↗

arXiv2020

StrObe: Streaming Object Detection from LiDAR Packets

Many modern robotics systems employ LiDAR as their main sensing modality due to its geometrical richness. Rolling shutter LiDARs are particularly common, in which an array of lasers scans the scene from a rotating base. Points are emitted as a stream of packets, each covering a sector of the 360° coverage. Modern perception algorithms wait for the full sweep to be built before processing the data, which introduces an additional latency. For typical 10Hz LiDARs this will be 100ms. As a consequence, by the time an output is produced, it no longer accurately reflects the state of the world. This poses a challenge, as robotics applications require minimal reaction times, such that maneuvers can be quickly planned in the event of a safety-critical situation. In this paper we propose StrObe, a novel approach that minimizes latency by ingesting LiDAR packets and emitting a stream of detections without waiting for the full sweep to be built. StrObe reuses computations from previous packets and iteratively updates a latent spatial representation of the scene, which acts as a memory, as new evidence comes in, resulting in accurate low-latency perception. We demonstrate the effectiveness of our approach on a large scale real-world dataset, showing that StrObe far outperforms the state-of-the-art when latency is taken into account, and matches the performance in the traditional setting.

cs.CV↗

arXiv2020

Uncertainty quantification for Multiphase-CFD simulations of bubbly flows: a machine learning-based Bayesian approach supported by high-resolution experiments

In this paper, we develop a machine learning-based Bayesian approach to inversely quantify and reduce the uncertainties of the two-fluid model-based multiphase computational fluid dynamics (MCFD) for bubbly flow simulations. The proposed approach is supported by high-resolution two-phase flow measurement techniques, including double-sensor conductivity probes, high-speed imaging, and particle image velocimetry. Local distribution of key physical quantities of interest (QoIs), including void fraction and phasic velocities, are obtained to support the modular Bayesian inference. In the process, the epistemic uncertainties of the closure relations are inversely quantified, the aleatory uncertainties from stochastic fluctuation of the system are evaluated based on experimental uncertainty analysis. The combined uncertainties are then propagated through the MCFD solver to obtain uncertainties of QoIs, based on which probability-boxes are constructed for validation. The proposed approach relies on three machine learning methods: feedforward neural networks and principal component analysis for surrogate modeling, and Gaussian processes for model form uncertainty modeling. The whole process is implemented within the framework of open-source deep learning library PyTorch with graphics processing unit (GPU) acceleration, thus ensuring the efficiency of the computation. The results demonstrate that with the support of high-resolution data, the uncertainty of MCFD simulations can be significantly reduced.

physics.flu-dyn↗

arXiv2020

Concentrated Differentially Private and Utility Preserving Federated Learning

Federated learning is a machine learning setting where a set of edge devices collaboratively train a model under the orchestration of a central server without sharing their local data. At each communication round of federated learning, edge devices perform multiple steps of stochastic gradient descent with their local data and then upload the computation results to the server for model update. During this process, the challenge of privacy leakage arises due to the information exchange between edge devices and the server when the server is not fully trusted. While some previous privacy-preserving mechanisms could readily be used for federated learning, they usually come at a high cost on convergence of the algorithm and utility of the learned model. In this paper, we develop a federated learning approach that addresses the privacy challenge without much degradation on model utility through a combination of local gradient perturbation, secure aggregation, and zero-concentrated differential privacy (zCDP). We provide a tight end-to-end privacy guarantee of our approach and analyze its theoretical convergence rates. Through extensive numerical experiments on real-world datasets, we demonstrate the effectiveness of our proposed method and show its superior trade-off between privacy and model utility.

cs.LG↗

arXiv2020

Certified Robustness of Graph Classification against Topology Attack with Randomized Smoothing

Graph classification has practical applications in diverse fields. Recent studies show that graph-based machine learning models are especially vulnerable to adversarial perturbations due to the non i.i.d nature of graph data. By adding or deleting a small number of edges in the graph, adversaries could greatly change the graph label predicted by a graph classification model. In this work, we propose to build a smoothed graph classification model with certified robustness guarantee. We have proven that the resulting graph classification model would output the same prediction for a graph under $l_0$ bounded adversarial perturbation. We also evaluate the effectiveness of our approach under graph convolutional network (GCN) based multi-class graph classification model.

cs.LG↗

arXiv2020

Trading Data For Learning: Incentive Mechanism For On-Device Federated Learning

Federated Learning rests on the notion of training a global model distributedly on various devices. Under this setting, users' devices perform computations on their own data and then share the results with the cloud server to update the global model. A fundamental issue in such systems is to effectively incentivize user participation. The users suffer from privacy leakage of their local data during the federated model training process. Without well-designed incentives, self-interested users will be unwilling to participate in federated learning tasks and contribute their private data. To bridge this gap, in this paper, we adopt the game theory to design an effective incentive mechanism, which selects users that are most likely to provide reliable data and compensates for their costs of privacy leakage. We formulate our problem as a two-stage Stackelberg game and solve the game's equilibrium. Effectiveness of the proposed mechanism is demonstrated by extensive simulations.

cs.LG↗

arXiv2020

Conditional Entropy Coding for Efficient Video Compression

We propose a very simple and efficient video compression framework that only focuses on modeling the conditional entropy between frames. Unlike prior learning-based approaches, we reduce complexity by not performing any form of explicit transformations between frames and assume each frame is encoded with an independent state-of-the-art deep image compressor. We first show that a simple architecture modeling the entropy between the image latent codes is as competitive as other neural video compression works and video codecs while being much faster and easier to implement. We then propose a novel internal learning extension on top of this architecture that brings an additional 10% bitrate savings without trading off decoding speed. Importantly, we show that our approach outperforms H.265 and other deep learning baselines in MS-SSIM on higher bitrate UVG video, and against all video codecs on lower framerates, while being thousands of times faster in decoding than deep models utilizing an autoregressive entropy model.

eess.IV↗

arXiv2020

Learning Lane Graph Representations for Motion Forecasting

We propose a motion forecasting model that exploits a novel structured map representation as well as actor-map interactions. Instead of encoding vectorized maps as raster images, we construct a lane graph from raw map data to explicitly preserve the map structure. To capture the complex topology and long range dependencies of the lane graph, we propose LaneGCN which extends graph convolutions with multiple adjacency matrices and along-lane dilation. To capture the complex interactions between actors and maps, we exploit a fusion network consisting of four types of interactions, actor-to-lane, lane-to-lane, lane-to-actor and actor-to-actor. Powered by LaneGCN and actor-map interactions, our model is able to predict accurate and realistic multi-modal trajectories. Our approach significantly outperforms the state-of-the-art on the large scale Argoverse motion forecasting benchmark.

cs.CV↗

arXiv2020

PnPNet: End-to-End Perception and Prediction with Tracking in the Loop

We tackle the problem of joint perception and motion forecasting in the context of self-driving vehicles. Towards this goal we propose PnPNet, an end-to-end model that takes as input sequential sensor data, and outputs at each time step object tracks and their future trajectories. The key component is a novel tracking module that generates object tracks online from detections and exploits trajectory level features for motion forecasting. Specifically, the object tracks get updated at each time step by solving both the data association problem and the trajectory estimation problem. Importantly, the whole model is end-to-end trainable and benefits from joint optimization of all tasks. We validate PnPNet on two large-scale driving datasets, and show significant improvements over the state-of-the-art with better occlusion recovery and more accurate future prediction.

cs.CV↗

arXiv2020

Development of a Data-driven Turbulence Model for 3d Thermal Stratification Simulation during Reactor Transients

SAM, a plant-level system analysis tool for advanced reactors (SFR, LFR, MSR/FHR) is under development at Argonne. As a modern system code, SAM aims to improve the predictions of 3D flows relevant to reactor safety during transient conditions. In order to fulfill this goal, one approach is to implement modeling of turbulent flow in SAM through establishing an embedded surrogate model for Reynolds stress/turbulence viscosity based on machine learning techniques. The proposed approach is based on an assumption that there exists a functional dependency relationship between local flow features and local turbulence viscosity or Reynolds stress. There have been very limited studies performed to validate this assumption. This paper documents a case study to examine the assumption in a scenario of potential reactor applications. The work doesn't aim to theoretically validate the assumption, but practically validate the assumption within the limited application domain. From the methodological point of view, the approach used in this paper could be classified into the so-called Type I machine learning (ML) approach, where a scale separation assumption is proposed claiming that conservation equations and closure relations are scale separable, for which the turbulence models are local rather than global. The CFD case studied in this work is a 3D transient thermal stratification tank flow problem performed using a Reynolds-averaged Navier-Stokes turbulence model in STARCCM+ code. Flow information of all geometric points in all timesteps is collected as training data and test data.

physics.flu-dyn↗

arXiv2020

Differentially Private Federated Learning for Resource-Constrained Internet of Things

With the proliferation of smart devices having built-in sensors, Internet connectivity, and programmable computation capability in the era of Internet of things (IoT), tremendous data is being generated at the network edge. Federated learning is capable of analyzing the large amount of data from a distributed set of smart devices without requiring them to upload their data to a central place. However, the commonly-used federated learning algorithm is based on stochastic gradient descent (SGD) and not suitable for resource-constrained IoT environments due to its high communication resource requirement. Moreover, the privacy of sensitive data on smart devices has become a key concern and needs to be protected rigorously. This paper proposes a novel federated learning framework called DP-PASGD for training a machine learning model efficiently from the data stored across resource-constrained smart devices in IoT while guaranteeing differential privacy. The optimal schematic design of DP-PASGD that maximizes the learning performance while satisfying the limits on resource cost and privacy loss is formulated as an optimization problem, and an approximate solution method based on the convergence analysis of DP-PASGD is developed to solve the optimization problem efficiently. Numerical results based on real-world datasets verify the effectiveness of the proposed DP-PASGD scheme.

cs.LG↗

arXiv2019

DeepPruner: Learning Efficient Stereo Matching via Differentiable PatchMatch

Our goal is to significantly speed up the runtime of current state-of-the-art stereo algorithms to enable real-time inference. Towards this goal, we developed a differentiable PatchMatch module that allows us to discard most disparities without requiring full cost volume evaluation. We then exploit this representation to learn which range to prune for each pixel. By progressively reducing the search space and effectively propagating such information, we are able to efficiently compute the cost volume for high likelihood hypotheses and achieve savings in both memory and computation. Finally, an image guided refinement module is exploited to further improve the performance. Since all our components are differentiable, the full network can be trained end-to-end. Our experiments show that our method achieves competitive results on KITTI and SceneFlow datasets while running in real-time at 62ms.

cs.CV↗

arXiv2019

DP-ADMM: ADMM-based Distributed Learning with Differential Privacy

Alternating Direction Method of Multipliers (ADMM) is a widely used tool for machine learning in distributed settings, where a machine learning model is trained over distributed data sources through an interactive process of local computation and message passing. Such an iterative process could cause privacy concerns of data owners. The goal of this paper is to provide differential privacy for ADMM-based distributed machine learning. Prior approaches on differentially private ADMM exhibit low utility under high privacy guarantee and often assume the objective functions of the learning problems to be smooth and strongly convex. To address these concerns, we propose a novel differentially private ADMM-based distributed learning algorithm called DP-ADMM, which combines an approximate augmented Lagrangian function with time-varying Gaussian noise addition in the iterative process to achieve higher utility for general objective functions under the same differential privacy guarantee. We also apply the moments accountant method to bound the end-to-end privacy loss. The theoretical analysis shows that DP-ADMM can be applied to a wider class of distributed learning problems, is provably convergent, and offers an explicit utility-privacy tradeoff. To our knowledge, this is the first paper to provide explicit convergence and utility properties for differentially private ADMM-based distributed learning algorithms. The evaluation results demonstrate that our approach can achieve good convergence and model accuracy under high end-to-end differential privacy guarantee.

cs.LG↗

arXiv2019

A portable potentiometric electronic tongue leveraging smartphone and cloud platforms

Electronic tongues based on potentiometry offer the prospect of rapid and continuous chemical fingerprinting for portable and remote systems. The present contribution presents a technology platform including a miniaturized electronic tongue based on electropolymerized ion-sensitive films, microcontroller-based data acquisition, a smartphone interface and cloud computing back-end for data storage and deployment of machine learning models. The sensor array records a series of differential voltages without use of a true reference electrode and the resulting time-series potentiometry data is used to train supervised machine learning algorithms. For trained systems, inferencing tasks such as the classification of liquids are realized within less than 1 minute including data acquisition at the edge and inference using the cloud-deployed machine learning model. Preliminary demonstration of the complete electronic tongue technology stack is reported for the classification of beverages and mineral water.

cs.ET↗

arXiv2019

Channel Covariance Matrix Estimation via Dimension Reduction for Hybrid MIMO MmWave Communication Systems

Hybrid massive MIMO structures with lower hardware complexity and power consumption have been considered as a potential candidate for millimeter wave (mmWave) communications. Channel covariance information can be used for designing transmitter precoders, receiver combiners, channel estimators, etc. However, hybrid structures allow only a lower-dimensional signal to be observed, which adds difficulties for channel covariance matrix estimation. In this paper, we formulate the channel covariance estimation as a structured low-rank matrix sensing problem via Kronecker product expansion and use a low-complexity algorithm to solve this problem. Numerical results with uniform linear arrays (ULA) and uniform squared planar arrays (USPA) are provided to demonstrate the effectiveness of our proposed method.

eess.SP↗