查看更多>>摘要：In the fields of intelligent transportation and multi-task cooperation,many practical problems can be modeled by colored traveling salesman problem(CTSP).When solving large-scale CTSP with a scale of more than 1000 dimensions,their convergence speed and the quality of their solutions are limited.This paper proposes a new hybrid ITÖ(HITÖ)algorithm,which integrates two new strategies,crossover operator and mutation strategy,into the stan-dard ITÖ.In the iteration process of HITÖ,the feasible solution of CTSP is represented by the double chromosome coding,and the random drift and wave operators are used to explore and develop new unknown regions.In this pro-cess,the drift operator is executed by the improved crossover operator,and the wave operator is performed by the optimized mutation strategy.Experiments show that HITÖ is superior to the known comparison algorithms in terms of the quality solution.
查看更多>>摘要：Graph convolutional networks that leverage spatial-temporal information from skeletal data have emerged as a popular approach for 3D human pose estimation.However,comprehensively modeling consistent spatial-temporal dependencies among the body joints remains a challenging task.Current approaches are limited by perform-ing graph convolutions solely on immediate neighbors,deploying separate spatial or temporal modules,and utilizing single-pass feedforward architectures.To solve these limitations,we propose a forward multi-scale residual graph con-volutional network(FMR-GNet)for 3D pose estimation from monocular video.First,we introduce a mix-hop spatial-temporal attention graph convolution layer that effectively aggregates neighboring features with learnable weights over large receptive fields.The attention mechanism enables dynamically computing edge weights at each layer.Second,we devise a cross-domain spatial-temporal residual module to fuse multi-scale spatial-temporal convolutional features through residual connections,explicitly modeling interdependencies across spatial and temporal domains.Third,we integrate a forward dense connection block to propagate spatial-temporal representations across network layers,en-abling high-level semantic skeleton information to enrich lower-level features.Comprehensive experiments conducted on two challenging 3D human pose estimation benchmarks,namely Human3.6M and MPI-INF-3DHP,demonstrate that the proposed FMR-GNet achieves superior performance,surpassing the most state-of-the-art methods.
查看更多>>摘要：Offline reinforcement learning(RL)has gathered increasing attention in recent years,which seeks to learn policies from static datasets without active online exploration.However,the existing offline RL approaches often require a large amount of pre-collected data and hence are hardly implemented by a single agent in practice.Inspired by the advancement of federated learning(FL),this paper studies federated offline reinforcement learning(FORL),whereby multiple agents collaboratively carry out offline policy learning with no need to share their raw trajectories.Clearly,a straightforward solution is to simply retrofit the off-the-shelf offline RL methods for FL,whereas such an approach easily overfits individual datasets during local updating,leading to instability and subpar performance.To overcome this challenge,we propose a new FORL algorithm,named model-free(MF)-FORL,that exploits novel"proximal local policy evaluation"to judiciously push up action values beyond local data support,enabling agents to capture the individual information without forgetting the aggregated knowledge.Further,we introduce a model-based variant,MB-FORL,capable of improving the generalization ability and computational efficiency via utilizing a learned dynamics model.We evaluate the proposed algorithms on a suite of complex and high-dimensional offline RL benchmarks,and the results demonstrate significant performance gains over the baselines.
查看更多>>摘要：The subspace clustering has been addressed by learning the block-diagonal self-expressive matrix.This block-diagonal structure heavily affects the accuracy of clustering but is rather challenging to obtain.A novel and effective subspace clustering model,i.e.,subspace clustering via block-diagonal decomposition(SCBD),is proposed,which can simultaneously capture the block-diagonal structure and gain the clustering result.In our model,a strict block-diagonal decomposition is introduced to directly pursue the k block-diagonal structure corresponding to k clus-ters.In this novel decomposition,the self-expressive matrix is decomposed into the block indicator matrix to demon-strate the cluster each sample belongs to.Based on the strict block-diagonal decomposition,the block-diagonal shift is proposed to capture the local intra-cluster structure,which shifts the samples in the same cluster to get smaller distances and results in more discriminative features for clustering.Extensive experimental results on synthetic and real databases demonstrate the superiority of SCBD over other state-of-the-art methods.
查看更多>>摘要：The tile-based multiplayer game Mahjong is widely played in Asia and has also become increasingly popular worldwide.Face-to-face or online,each player begins with a hand of 13 tiles and players draw and discard tiles in turn until they complete a winning hand.An important notion in Mahjong is the deficiency number(a.k.a.shanten number in Japanese Mahjong)of a hand,which estimates how many tile changes are necessary to complete the hand into a winning hand.The deficiency number plays an essential role in major decision-making tasks such as selecting a tile to discard.This paper proposes a fast algorithm for computing the deficiency number of a Mahjong hand.Compared with the baseline algorithm,the new algorithm is usually 100 times faster and,more importantly,respects the agent's knowledge about available tiles.The algorithm can be used as a basic procedure in all Mahjong variants by both rule-based and machine learning-based Mahjong AI.
查看更多>>摘要：Model checking computation tree logic based on multi-valued possibility measures has been studied by Li et al.on Information Sciences in 2019.However,the previous work did not consider the nondeterministic choices inherent in systems represented by multi-valued Kripke structures(MvKSs).This nondeterminism is crucial for accur-ate system modeling,decision making,and control capabilities.To address this limitation,we draw inspiration from the generalization of Markov chains to Markov decision processes in probabilistic systems.By integrating nondeter-minism into MvKS,we introduce multi-valued decision processes(MvDPs)as a framework for modeling MvKSs with nondeterministic choices.We investigate the problems of model checking over MvDPs.Verifying properties are ex-pressed by using multi-valued computation tree logic based on schedulers.Our primary objective is to leverage fix-point techniques to determine the maximum and minimum possibilities of the system satisfying temporal properties.This allows us to identify the optimal or worst-case schedulers for decision making or control purposes.We aim to develop reduction techniques that enhance the efficiency of model checking,thereby reducing the associated time complexity.We mathematically demonstrate three reduction techniques that improve model checking performance in most scenarios.
查看更多>>摘要：With the development of knowledge graphs,a series of applications based on knowledge graphs have emerged.The incompleteness of knowledge graphs makes the effect of the downstream applications affected by the quality of the knowledge graphs.To improve the quality of knowledge graphs,translation-based graph embeddings such as TransE,learn structural information by representing triples as low-dimensional dense vectors.However,it is difficult to generalize to the unseen entities that are not observed during training but appear during testing.Other methods use the powerful representational ability of pre-trained language models to learn entity descriptions and con-textual representation of triples.Although they are robust to incompleteness,they need to calculate the score of all candidate entities for each triple during inference.We consider combining two models to enhance the robustness of unseen entities by semantic information,and prevent combined explosion by reducing inference overhead through structured information.We use a pre-training language model to code triples and learn the semantic information within them,and use a hyperbolic space-based distance model to learn structural information,then integrate the two types of information together.We evaluate our model by performing link prediction experiments on standard datasets.The experimental results show that our model achieves better performances than state-of-the-art methods on two standard datasets.
查看更多>>摘要：As a new computing method,edge computing not only improves the computing efficiency and pro-cessing power of data,but also reduces the transmission delay of data.Due to the wide variety of edge devices and the increasing amount of terminal data,third-party data centers are unable to ensure no user privacy data leaked.To solve these problems,this paper proposes an iterative clustering algorithm named local differential privacy iterative aggregation(LDPIA)based on localized differential privacy,which implements local differential privacy.To address the problem of uncertainty in numerical types of mixed data,random perturbation is applied to the user data at the attribute category level.The server then performs clustering on the perturbed data,and density threshold and distur-bance probability are introduced to update the cluster point set iteratively.In addition,a new distance calculation formula is defined in combination with attribute weights to ensure the availability of data.The experimental results show that LDPIA algorithm achieves better privacy protection and availability simultaneously.
查看更多>>摘要：The WiFi fingerprint-based localization method is considered one of the most popular techniques for indoor localization.In INFOCOM'14,Li et al.proposed a wireless fidelity(WiFi)fingerprint localization system based on Paillier encryption,which is claimed to protect both client C's location privacy and service provider S's database privacy.However,Yang et al.presented a practical data privacy attack in INFOCOM'18,which allows a polynomial time attacker to obtain S's database.We propose a novel WiFi fingerprint localization system based on Castagnos-Laguillaumie(CL)encryption,which has a trustless setup and is efficient due to the excellent properties of CL en-cryption.To prevent Yang et al.'s attack,the system requires that S selects only the locations from its database that can receive the nonzero signals from all the available access points in C's nonzero fingerprint in order to determine C's location.Security analysis shows that our scheme is secure under Li et al.'s threat model.Furthermore,to enhance the security level of privacy-preserving WiFi fingerprint localization scheme based on CL encryption,we propose a se-cure and efficient zero-knowledge proof protocol for the discrete logarithm relations in C's encrypted localization queries.
查看更多>>摘要：Nowadays,people are getting used to upload images to a third party for post-processing,such as image denoising and super-resolution.This may easily lead to the disclosure of the privacy in the confidential images.One possible solution is to encrypt the image before sending it to the third party,the encrypted image can be easily de-tected by a malicious attacker in the transmission channel.We propose a confidential image super-resolution method named HSR-Net which firstly hide the secret image and then super-resolve it in the hidden domain.The method is composed of three important modules:image hiding module(IHM),image super-resolution module(ISM),and image revealing module(IRM).The IHM aims to encode secret image and hide it into a cover image to generate the stego image.The stego image looks similar to the cover image but contains the information of the secret image.The third party uses the ISM to perform image super-resolution on the stego image.The user can reveal the super-resolved secret image from the stego image.The proposed HSR-Net method has two advantages.It ensures that the third party cannot directly operate on the secret image,thus protecting the user's privacy.Due to the similarity between the stego image and cover image,we can reduce the attacker's suspicion to further improve the image security.The experimental re-sults were tested on DIV2K dataset and Flickr2K dataset.The peak signal-to-noise ratios(PSNR)of IHM,ISM,and IRM are 38.81 dB,28.91 dB,and 23.51 dB,respectively,which verify that the proposed HSR-Net method is able to achieve image super-resolution and protect user's privacy simultaneouly.
万方数据