CAAI Transactions on Intelligence Technology
Volume 4, Issue 1, March 2019
Volumes & issues:
Volume 4, Issue 1
March 2019
-
- Author(s): Rui Jiang ; Hui Zhou ; Han Wang ; Shuzhi Sam Ge
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 1 –8
- DOI: 10.1049/trit.2018.1058
- Type: Article
- + Show details - Hide details
-
p.
1
–8
(8)
This study presents a new perspective for autonomous mobile robots path searching by proposing a biasing direction towards causal entropy maximisation during random tree generation. Maximum entropy-biased rapidly-exploring random tree (ME-RRT) is proposed where the searching direction is computed from random path sampling and path integral approximation, and the direction is incorporated into the existing rapidly-exploring random tree (RRT) planner. Properties of ME-RRT including degenerating conditions and additional time complexity are also discussed. The performance of the proposed approach is studied, and the results are compared with conventional RRT/RRT* and goal-biased approach in 2D/3D scenarios. Simulations show that trees are generated efficiently with fewer iteration numbers, and the success rate within limited iterations has been greatly improved in complex environments.
- Author(s): Zheng Yu and Wenmin Wang
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 9 –16
- DOI: 10.1049/trit.2018.1051
- Type: Article
- + Show details - Hide details
-
p.
9
–16
(8)
Cross-modal retrieval has been recently proposed to find an appropriate subspace, where the similarity across different modalities such as image and text can be directly measured. In this study, different from most existing works, the authors propose a novel model for cross-modal retrieval based on a domain-adaptive limited text space (DALTS) rather than a common space or an image space. Experimental results on three widely used datasets, Flickr8K, Flickr30K and Microsoft Common Objects in Context (MSCOCO), show that the proposed method, dubbed DALTS, is able to learn superior text space features which can effectively capture the necessary information for cross-modal retrieval. Meanwhile, DALTS achieves promising improvements in accuracy for cross-modal retrieval compared with the current state-of-the-art methods.
- Author(s): Chunwei Tian ; Yong Xu ; Lunke Fei ; Junqian Wang ; Jie Wen ; Nan Luo
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 17 –23
- DOI: 10.1049/trit.2018.1054
- Type: Article
- + Show details - Hide details
-
p.
17
–23
(7)
Owing to the flexible architectures of deep convolutional neural networks (CNNs) are successfully used for image denoising. However, they suffer from the following drawbacks: (i) deep network architecture is very difficult to train. (ii) Deeper networks face the challenge of performance saturation. In this study, the authors propose a novel method called enhanced convolutional neural denoising network (ECNDNet). Specifically, they use residual learning and batch normalisation techniques to address the problem of training difficulties and accelerate the convergence of the network. In addition, dilated convolutions are used in the proposed network to enlarge the context information and reduce the computational cost. Extensive experiments demonstrate that the ECNDNet outperforms the state-of-the-art methods for image denoising.
- Author(s): Jing Liu ; Yaxiong Chi ; Zongdong Liu ; Shan He
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 24 –36
- DOI: 10.1049/trit.2018.1059
- Type: Article
- + Show details - Hide details
-
p.
24
–36
(13)
Many methods aim to use data, especially data about gene expression based on high throughput genomic methods, to identify complicated regulatory relationships between genes. The authors employ a simple but powerful tool, called fuzzy cognitive maps (FCMs), to accurately reconstruct gene regulatory networks (GRNs). Many automated methods have been carried out for training FCMs from data. These methods focus on simulating the observed time sequence data, but neglect the optimisation of network structure. In fact, the FCM learning problem is multi-objective which contains network structure information, thus, the authors propose a new algorithm combining ensemble strategy and multi-objective evolutionary algorithm (MOEA), called EMOEAFCM-GRN, to reconstruct GRNs based on FCMs. In EMOEAFCM-GRN, the MOEA first learns a series of networks with different structures by analysing historical data simultaneously, which is helpful in finding the target network with distinct optimal local information. Then, the networks which receive small simulation error on the training set are selected from the Pareto front and an efficient ensemble strategy is provided to combine these selected networks to the final network. The experiments on the DREAM4 challenge and synthetic FCMs illustrate that EMOEAFCM-GRN is efficient and able to reconstruct GRNs accurately.
- Author(s): Zedong Tang and Maoguo Gong
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 37 –46
- DOI: 10.1049/trit.2018.1090
- Type: Article
- + Show details - Hide details
-
p.
37
–46
(10)
Existing multifactorial particle swarm optimisation (MFPSO) algorithms only explore a relatively narrow area between the inter-task particles. Meanwhile, these algorithms use a fixed inter-task learning probability throughout the evolution process. However, the parameter is problem dependent and can be various at different stages of the evolution. In this work, the authors devise an inter-task learning-based information transferring mechanism to replace the corresponding part in MFPSO. This inter-task learning mechanism transfers the searching step by using a differential term and updates the personal best position by employing an inter-task crossover. By this mean, the particles can explore a broad search space when utilising the additional searching experiences of other tasks. In addition, to enhance the performance on problems with different complementarity, they design a self-adaption strategy to adjust the inter-task learning probability according to the performance feedback. They compared the proposed algorithm with the state-of-the-art algorithms on various benchmark problems. Experimental results demonstrate that the proposed algorithm can transfer inter-task knowledge efficiently and perform well on the problems with different complementarity.
- Author(s): Xiaofei Li ; Laurent Girin ; Radu Horaud
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 47 –53
- DOI: 10.1049/trit.2018.1061
- Type: Article
- + Show details - Hide details
-
p.
47
–53
(7)
This study addresses the problem of under-determined speech source separation from multichannel microphone signals, i.e. the convolutive mixtures of multiple sources. The time-domain signals are first transformed to the short-time Fourier transform (STFT) domain. To represent the room filters in the STFT domain, instead of the widely used narrowband assumption, the authors propose to use a more accurate model, i.e. the convolutive transfer function (CTF). At each frequency band, the CTF coefficients of the mixing filters and the STFT coefficients of the sources are jointly estimated by maximising the likelihood of the microphone signals, which is resolved by an expectation-maximisation algorithm. Experiments show that the proposed method provides very satisfactory performance under highly reverberant environments.
- Author(s): Mohamed Raessa ; Daniel Sánchez ; Weiwei Wan ; Damien Petit ; Kensuke Harada
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 54 –63
- DOI: 10.1049/trit.2018.1062
- Type: Article
- + Show details - Hide details
-
p.
54
–63
(10)
This study presents a straightforward method to teach robots to use tools. Teaching robots is crucial in quickly deploying and reconfiguring robots in next-generation factories. Conventional methods require third-party systems like wearable devices or complicated vision system to capture, analyse, and map human grasps, motion, and tool poses to robots. These systems assume lots of experience from their users. Unlike the conventional methods, this study does not involve learning human motion and skills. Instead, it only learns the object goal poses from the human user whilst employs regrasp planning to generate robot motion. The method is most suitable for a robot to learn the usage of electric tools that can be operated by simply switching on and off. The proposed method is validated using a dual-arm robot with hand-mounted cameras and several tools. Experimental results show that the proposed method is robust, feasible, and simple to teach robots. It can find a collision-free and kino-dynamic feasible grasp sequences and motion trajectories when the goal pose is reachable. The method allows the robot to automatically choose placements or handover considering the surrounding environment as intermediate states to change the pose of the tool and use tools following human demonstrations.
- Author(s): Kazuyuki Matsumoto ; Fuji Ren ; Masaya Matsuoka ; Minoru Yoshida ; Kenji Kita
- Source: CAAI Transactions on Intelligence Technology, Volume 4, Issue 1, p. 64 –71
- DOI: 10.1049/trit.2018.1060
- Type: Article
- + Show details - Hide details
-
p.
64
–71
(8)
Recently, the authors often see words such as youth slang, neologism and Internet slang on social networking sites (SNSs) that are not registered on dictionaries. Since the documents posted to SNSs include a lot of fresh information, they are thought to be useful for collecting information. It is important to analyse these words (hereinafter referred to as ‘slang’) and capture their features for the improvement of the accuracy of automatic information collection. This study aims to analyse what features can be observed in slang by focusing on the topic. They construct topic models from document groups including target slang on Twitter by latent Dirichlet allocation. With the models, they chronologically the analyse change of topics during a certain period of time to find out the difference in the features between slang and general words. Then, they propose a slang classification method based on the change of features.
Maximum entropy searching
Learning DALTS for cross-modal retrieval
Enhanced CNN for image denoising
Ensemble multi-objective evolutionary algorithm for gene regulatory network reconstruction based on fuzzy cognitive maps
Adaptive multifactorial particle swarm optimisation
Expectation-maximisation for speech source separation using convolutive transfer function
Teaching a robot to use electric tools with regrasp planning
Slang feature extraction by analysing topic change on social media
Most viewed content
Most cited content for this Journal
-
Self‐training maximum classifier discrepancy for EEG emotion recognition
- Author(s): Xu Zhang ; Dengbing Huang ; Hanyu Li ; Youjia Zhang ; Ying Xia ; Jinzhuo Liu
- Type: Article
-
A robust deformed convolutional neural network (CNN) for image denoising
- Author(s): Qi Zhang ; Jingyu Xiao ; Chunwei Tian ; Jerry Chun‐Wei Lin ; Shichao Zhang
- Type: Article
-
Boosting image watermarking authenticity spreading secrecy from counting‐based secret‐sharing
- Author(s): Adnan Gutub
- Type: Article
-
A survey on adversarial attacks and defences
- Author(s): Anirban Chakraborty ; Manaar Alam ; Vishal Dey ; Anupam Chattopadhyay ; Debdeep Mukhopadhyay
- Type: Article
-
Deep learning: Applications, architectures, models, tools, and frameworks: A comprehensive survey
- Author(s): Mehdi Gheisari ; Fereshteh Ebrahimzadeh ; Mohamadtaghi Rahimi ; Mahdieh Moazzamigodarzi ; Yang Liu ; Pijush Kanti Dutta Pramanik ; Mohammad Ali Heravi ; Abolfazl Mehbodniya ; Mustafa Ghaderzadeh ; Mohammad Reza Feylizadeh ; Saeed Kosari
- Type: Article