arXiv subjects

Haifeng Hu

Publications and source records attributed to Haifeng Hu.

44 records · Page 3Linked to original sources

arXiv2020

Modality to Modality Translation: An Adversarial Representation Learning and Graph Fusion Network for Multimodal Fusion

Learning joint embedding space for various modalities is of vital importance for multimodal fusion. Mainstream modality fusion approaches fail to achieve this goal, leaving a modality gap which heavily affects cross-modal fusion. In this paper, we propose a novel adversarial encoder-decoder-classifier framework to learn a modality-invariant embedding space. Since the distributions of various modalities vary in nature, to reduce the modality gap, we translate the distributions of source modalities into that of target modality via their respective encoders using adversarial training. Furthermore, we exert additional constraints on embedding space by introducing reconstruction loss and classification loss. Then we fuse the encoded representations using hierarchical graph neural network which explicitly explores unimodal, bimodal and trimodal interactions in multi-stage. Our method achieves state-of-the-art performance on multiple datasets. Visualization of the learned embeddings suggests that the joint embedding space learned by our method is discriminative. code is available at: \url{https://github.com/TmacMai/ARGF_multimodal_fusion}

cs.CV↗

arXiv2020

Universal Multi-Source Domain Adaptation

Unsupervised domain adaptation enables intelligent models to transfer knowledge from a labeled source domain to a similar but unlabeled target domain. Recent study reveals that knowledge can be transferred from one source domain to another unknown target domain, called Universal Domain Adaptation (UDA). However, in the real-world application, there are often more than one source domain to be exploited for domain adaptation. In this paper, we formally propose a more general domain adaptation setting, universal multi-source domain adaptation (UMDA), where the label sets of multiple source domains can be different and the label set of target domain is completely unknown. The main challenges in UMDA are to identify the common label set between each source domain and target domain, and to keep the model scalable as the number of source domains increases. To address these challenges, we propose a universal multi-source adaptation network (UMAN) to solve the domain adaptation problem without increasing the complexity of the model in various UMDA settings. In UMAN, we estimate the reliability of each known class in the common label set via the prediction margin, which helps adversarial training to better align the distributions of multiple source domains and target domain in the common label set. Moreover, the theoretical guarantee for UMAN is also provided. Massive experimental results show that existing UDA and multi-source DA (MDA) methods cannot be directly applied to UMDA and the proposed UMAN achieves the state-of-the-art performance in various UMDA settings.

cs.CV↗

arXiv2020

Unveiling Class-Labeling Structure for Universal Domain Adaptation

As a more practical setting for unsupervised domain adaptation, Universal Domain Adaptation (UDA) is recently introduced, where the target label set is unknown. One of the big challenges in UDA is how to determine the common label set shared by source and target domains, as there is simply no labeling available in the target domain. In this paper, we employ a probabilistic approach for locating the common label set, where each source class may come from the common label set with a probability. In particular, we propose a novel approach for evaluating the probability of each source class from the common label set, where this probability is computed by the prediction margin accumulated over the whole target domain. Then, we propose a simple universal adaptation network (S-UAN) by incorporating the probabilistic structure for the common label set. Finally, we analyse the generalization bound focusing on the common label set and explore the properties on the target risk for UDA. Extensive experiments indicate that S-UAN works well in different UDA settings and outperforms the state-of-the-art methods by large margins.

cs.AI↗

arXiv2020

Learning-Based Safety-Stability-Driven Control for Safety-Critical Systems under Model Uncertainties

Safety and tracking stability are crucial for safety-critical systems such as self-driving cars, autonomous mobile robots, industrial manipulators. To efficiently control safety-critical systems to ensure their safety and achieve tracking stability, accurate system dynamic models are usually required. However, accurate system models are not always available in practice. In this paper, a learning-based safety-stability-driven control (LBSC) algorithm is presented to guarantee the safety and tracking stability for nonlinear safety-critical systems subject to control input constraints under model uncertainties. Gaussian Processes (GPs) are employed to learn the model error between the nominal model and the actual system dynamics, and the estimated mean and variance of the model error are used to quantify a high-confidence uncertainty bound. Using this estimated uncertainty bound, a safety barrier constraint is devised to ensure safety, and a stability constraint is developed to achieve rapid and accurate tracking. Then the proposed LBSC method is formulated as a quadratic program incorporating the safety barrier, the stability constraint, and the control constraints. The effectiveness of the LBSC method is illustrated on the safety-critical connected cruise control (CCC) system simulator under model uncertainties.

cs.RO↗

arXiv2020

Metric-Learning-Assisted Domain Adaptation

Domain alignment (DA) has been widely used in unsupervised domain adaptation. Many existing DA methods assume that a low source risk, together with the alignment of distributions of source and target, means a low target risk. In this paper, we show that this does not always hold. We thus propose a novel metric-learning-assisted domain adaptation (MLA-DA) method, which employs a novel triplet loss for helping better feature alignment. We explore the relationship between the second largest probability of a target sample's prediction and its distance to the decision boundary. Based on the relationship, we propose a novel mechanism to adaptively adjust the margin in the triplet loss according to target predictions. Experimental results show that the use of proposed triplet loss can achieve clearly better results. We also demonstrate the performance improvement of MLA-DA on all four standard benchmarks compared with the state-of-the-art unsupervised domain adaptation methods. Furthermore, MLA-DA shows stable performance in robust experiments.

cs.LG↗

arXiv2020

Adaptive Interaction Modeling via Graph Operations Search

Interaction modeling is important for video action analysis. Recently, several works design specific structures to model interactions in videos. However, their structures are manually designed and non-adaptive, which require structures design efforts and more importantly could not model interactions adaptively. In this paper, we automate the process of structures design to learn adaptive structures for interaction modeling. We propose to search the network structures with differentiable architecture search mechanism, which learns to construct adaptive structures for different videos to facilitate adaptive interaction modeling. To this end, we first design the search space with several basic graph operations that explicitly capture different relations in videos. We experimentally demonstrate that our architecture search framework learns to construct adaptive interaction modeling structures, which provides more understanding about the relations between the structures and some interaction characteristics, and also releases the requirement of structures design efforts. Additionally, we show that the designed basic graph operations in the search space are able to model different interactions in videos. The experiments on two interaction datasets show that our method achieves competitive performance with state-of-the-arts.

cs.CV↗

arXiv2019

Dynamically configurable successively switchable multispectral plasmon induced transparency

Plasmon-induced-transparency (PIT) in nanostructures has been intensively investigated, however, no existing metasurface nanostructure exhibits all-optically tunable properties, where the number of transparency windows can be tuned successively, and switched to off-state. Here, we theoretically investigate and demonstrate dynamically tunable, multichannel PIT at optical frequencies. The inplane destructive interference between bright and dark dipolar resonances in coupled plasmonic nanobar topologies is exploited to produce tunable PIT with unique characteristics. In particular, we demonstrate sequential polarization-selective multispectral operation whereby the number of PIT channels can be varied successively from '3' to '0'. The results provide a promising route for active manipulation of PIT and show potential applications for multifunctional dynamic nanophotonics devices.

physics.optics↗

arXiv2017

Stability and semi-classical limit in a semiconductor full quantum hydrodynamic model with non-flat doping profile

We present the new results on stability and semi-classical limit in a semiconductor full quantum hydrodynamic (FQHD) model with non-flat doping profile. The FQHD model can be used to analyze the thermal and quantum influences on the transport of carriers (electrons or holes) in semiconductor device. Inspired by the physical motivation, we consider the initial-boundary value problem of this model over the one-dimensional bounded domain and adopt the ohmic contact boundary condition and the vanishing bohmenian-type boundary condition. Firstly, the existence and asymptotic stability of a stationary solution are proved by Leray- Schauder fixed-point theorem, Schauder fixed-point theorem and the refined energy method. Secondly, we show the semi-classical limit results for both stationary solutions and global solutions by the elaborate energy estimates and the compactness argument. The strong convergence rates of the related asymptotic sequences of solutions are also obtained.

math.AP↗