International Journal on Artificial Intelligence Tools

This work presents a novel approach to multimodal human action recognition by jointly using visual RGB and depth (including skeleton joint positions) data captured from depth camera. For the depth feature extraction, Local Surface Geometric Feature (LSGF) is adopted to capture the geometric appearance and postures cues. Simultaneously, the improved dense trajectory feature (IDT) is extracted from RGB modality to jointly characterize the motion, visual appearance and trajectory shape information. These features from different modalities are complementary to each other. Then a two-stage integration scheme is proposed, which incorporates the probability weights of each classifier for action recognition. The proposed approach is evaluated on four publicly available human action databases: NJUST RGB-D Action, MSR-ActionPairs, MSR-DailyAct3D, and UTD-MHAD. Experimental results demonstrate that the proposed approach outperforms or is comparable to the state-of-the-art methods.

In this paper a novel technique for building change detection from remote sensing imagery is presented. It includes two main stages: (1) Object-specific discriminative features are extracted using Morphological Building Index (MBI) to automatically detect the existence of buildings in remote sensing images. (2) Pixel-based image matching is measured on the basis of Mutual Information (MI) of the images by Normalized Mutual Information (NMI). Here, the MBI features values are computed for each of the pair images taken over the same region at two different times and then changes in these two MBI images are measured to indicate the building change. MI is estimated locally for all the pixels for image matching and then thresholding is applied for eliminating those pixels which are responsible for strong similarity. Finally, after getting the MBI and NMI images, a further fusion of these two images is done for refinement of the change result. For evaluation purpose, the experiments are carried on QuickBird, IKONOS images and images taken from Google Earth. The results show that the proposed technique can attain acceptable correctness rates above 90% with Overall Accuracy (OA) 89.52%.

Multi-view product image queries can improve retrieval performance over single view queries significantly. In this paper, we investigated the performance of deep convolutional neural networks (ConvNets) on multi-view product image search. First, we trained a VGG-like network to learn deep ConvNets representations of product images. Then, we computed the deep ConvNets representations of database and query images and performed single view queries, and multi-view queries using several early and late fusion approaches.

We performed extensive experiments on the publicly available Multi-View Object Image Dataset (MVOD 5K) with both clean background queries from the internet and cluttered background queries from a mobile phone. We compared the performance of ConvNets to the classical bag-of-visual-words (BoWs). We concluded that (1) multi-view queries with deep ConvNets representations perform significantly better than single view queries, (2) ConvNets perform much better than BoWs and have room for further improvement, (3) pre-training of ConvNets on a different image dataset with background clutter is needed to obtain good performance on cluttered product image queries obtained with a mobile phone.

In this paper, we revisit an important issue of CDCL-based SAT solvers, namely the learned clauses database management policies. Our motivation takes its source from a simple observation on the remarkable performances of both random and size-bounded reduction strategies. We first derive a simple reduction strategy, called Size-Bounded Randomized strategy (in short SBR), that combines maintaining short clauses (of size bounded by k), while deleting randomly clauses of size greater than k. The resulting strategy outperform the state-of-the-art on SAT instances taken from the SAT competitions 2013 and 2018, and remains competitive on a broad range of SAT instances of the SAT Competition 2014. Reinforced by the interest of keeping short clauses, we propose several new dynamic variants, and we discuss their performance. We also propose different ways for adjusting dynamically the size-bounded parameter of the strategy.

In problems involving control of financial processes, it is usually complicated to quantify exactly the state variables. It could be expensive to acquire the exact value of a given state, even if it may be physically possible to do so. In such cases it may be interesting to support the decision-making process on inaccurate information pertaining to the system state. In addition, for modeling real-world application, it is necessary to compute the values of the parameters of the environment (transition probabilities and observation probabilities) and the reward functions, which are typically, hand-tuned by experts in the field until it has acquired a satisfactory value. This results in an undesired process.

To address these shortcomings, this paper provides a new Reinforcement Learning (RL) framework for computing the mean-variance customer portfolio with transaction costs in controllable Partially Observable Markov Decision Processes (POMDPs). The solution is restricted to finite state, action, observation sets and average reward problems. For solving this problem, a controller/actor-critic architecture is proposed, which balance the difficult tasks of exploitation and exploration of the environment. The architecture consists of three modules: controller, fast-tracked portfolio learning and an actor-critic module. Each module involves the design of a convergent Temporal Difference (TD) learning algorithm. We employ three different learning rules to estimate the real values of: (a) the transition matrices ${\widehat{\pi }}_{(j|ik)}(t)$, (b) the rewards ${\widehat{\text{J}}}_{(ijk)}(t)$ and (c) the resources destined for carrying out a promotion ${\widehat{\zeta }}_{(\widehat{i}|\widehat{k})}(t)$. We present a proof for the estimated transition matrix rule ${\widehat{\pi }}_{(j|ik)}(t)$ and showing that it converges when t → ∞. For solving the optimization programming problem we extend the c-variable method for partially observable Markov chains. The c-variable is conceptualized as joint strategy given by the product of the control policy, the observation kernel Q(y|s) and the stationary distribution vector. A major advantage of this procedure is that it can be implemented efficiently for real settings in controllable POMDP. A numerical example illustrates the results of the proposed method.

This work proposes a new powerful meta-heuristic optimization algorithm in education process called Competitive Learning (CLA). The algorithm is benchmarked on 8 well-known test functions, and the results are verified by a comparative study with some meta-heuristic optimization methods including: Imperialist Competitive Algorithm (ICA), Teaching-Learning-Based Optimization (TLBO), Grey Wolf Algorithm (GWO), Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). Analyzing the findings, it is shown that the CLA algorithm is able to provide more accurate results than other well-known meta-heuristic ones. Also, those results applied to famous unimodal and multimodal benchmarks show CLA is efficient in improving accuracy as well as computational speed.

This paper presents a prognostic scheme for estimating the remaining useful life of Lithium-ion batteries. The proposed scheme utilizes a prediction module that aims to obtain precise predictions for both short and long prediction horizons. The prediction module makes use of extreme learning machines for one-step and multi-step ahead predictions, using various prediction strategies, including iterative, direct and DirRec, which use the constant-current experimental capacity data for the estimation of the remaining useful life. The data-driven prognostic approach is highly dependent on the availability of high quantity of quality observations. Insufficient amount of available data can result in unsatisfactory prognostics. In this paper, the prognostics scheme is utilized to estimate the remaining useful life of a battery, with insufficient direct data available, but taking advantage of observations available from a fleet of similar batteries with similar working conditions. Experimental results show that the proposed prognostic scheme provides a fast and efficient estimation of the remaining useful life of the batteries and achieves superior results when compared with various state-of-the-art prediction techniques.