In multimodal emotion recognition, disentangled representation learning method effectively address the inherent heterogeneity among modalities. To facilitate the flexible integration of enhanced disentangled features into multimodal emotional features, we propose a task-driven multimodal emotion recognition method TDMER. Its Cross-Modal Learning module promotes adaptive cross-modal learning of features disentangled into modality-invariant and modality-specific subspaces, based on their contributions to emotional classification probabilities. The Task-Contribution Fusion mechanism then assigns controllable weights to the enhanced features according to their task objectives, generating multimodal fusion features that improve the emotion classifier’s discriminative ability. The proposed TDMER approach has been evaluated on two widely-used multimodal emotion recognition benchmarks and demonstrated significant performance improvements compared with other state-of the-art methods.

Influence maximization (IM) is a very important issue in social network diffusion analysis. The topology of real social network is large-scale, dynamic, and heterogeneous. The heterogeneity, and continuous expansion and evolution of social network pose a challenge to find influential users. Existing IM algorithms usually assume that social networks are static or dynamic but homogeneous to simplify the complexity of the IM problem. We propose a community-based influence maximization algorithm using network embedding in dynamic heterogeneous social networks. We use DyHATR algorithm to obtain the propagation feature vectors of network nodes, and execute k -means cluster algorithm to transform the original network into a coarse granularity network (CGN). On CGN, we propose a community-based three-hop independent cascade model and construct the objective function of IM problem. We design a greedy heuristics algorithm to solve the IM problem with \((1-\frac{1}{e})-\) approximation guarantee and use community structure to quickly identify seed users and estimate their influence value. Experimental results on real social networks demonstrated that compared with existing IM algorithms, our proposed algorithm had better comprehensive performance with respect to the influence value, more less execution time and memory consumption, and better scalability.

Imaging-type monitoring techniques are used in monitoring dynamic processes in many domains, including medicine, engineering, and geophysics. This paper aims to propose an efficient workflow for application of such data for the conditioning of simulation models. Such applications are very common in e.g. the geosciences, where large-scale simulation models and measured data are used to monitor the state of e.g. energy and water systems, predict their future behavior and optimize actions to achieve desired behavior of the system. In order to reduce the high computational cost and complexity of data assimilation workflows for high-dimensional parameter estimation, a residual-in-residual dense block extension of the U-Net convolutional network architecture is proposed, to predict time-evolving features in high-dimensional grids. The network is trained using high-fidelity model simulations. We present two examples of application of the trained network as a surrogate within an iterative ensemble-based workflow to estimate the static parameters of geological reservoirs based on binary-type image data, which represent fluid facies as obtained from time-lapse seismic surveys. The differences between binary images are parameterized in terms of distances between the fluid-facies boundaries, or fronts. We discuss the impact of the choice of network architecture, loss function, and number of training samples on the accuracy of results and on overall computational cost. From comparisons with conventional workflows based entirely on high-fidelity simulation models, we conclude that the proposed surrogate-supported hybrid workflow is able to deliver results with an accuracy equal to or better than the conventional workflow, and at significantly lower cost. Cost reductions are shown to increase with the number of samples of the uncertain parameter fields. The hybrid workflow is generic and should be applicable in addressing inverse problems in many geophysical applications as well as other engineering domains.

Visibility forecast is a meteorological problems which has direct impact to daily lives. For instance, timely prediction of low visibility situations is very important for the safe operation in airports and highways. In this paper, we investigate the use of Long Short-Term Memory(LSTM) model to predict visibility. By adjusting the loss function and network structure, we optimize the original LSTM model to make it more suitable for practical applications, which is superior to previous models in short-term low visibility prediction. In addition, there is a ”time delay problem” when the number of hours time ahead we try to forecast becomes larger, this problem is persistent given the limited amount of available training data. We report our attempt of applying re-sampling to deal with the time delay problem, and we find that this method can improve the accuracy of visibility prediction, especially for the low visibility case.

Numerical models of chemical transport have been used to simulate the complex processes involved in the formation and transport of air pollutants. Although these models can predict the spatiotemporal variability of a variety of chemical species, the accuracy of these models is often limited. Therefore, in the past two decades, data assimilation methods have been applied to use the available measurements for improving the forecast. Nowadays, machine learning techniques provide new opportunities for improving the air quality forecast. A case study on PM10 concentrations during a dust storm is performed. It is known that the PM10 concentrations are caused by multiple emission sources, e.g., dust from the desert and anthropogenic emissions. Accurate modeling of the PM10 concentration levels owing to the local anthropogenic emissions is essential for an adequate evaluation of the dust level. However, real-time measurement of local emissions is not possible, so no direct data is available. Actually, the lack of in-time emission inventories is one of the main reasons that current numerical chemical transport models cannot produce accurate anthropogenic PM10 simulations. Using machine learning techniques to generate local emissions based on past observations is a promising approach. We report how it can be combined with data assimilation to improve the accuracy of air quality forecast considerably.

In machine vision typical heuristic methods to extract parameterized objects out of raw data points are the Hough transform and RANSAC. Bayesian models carry the promise to optimally extract such parameterized objects given a correct definition of the model and the type of noise at hand. A category of solvers for Bayesian models are Markov chain Monte Carlo methods. Naive implementations of MCMC methods suffer from slow convergence in machine vision due to the complexity of the parameter space. Towards this blocked Gibbs and split-merge samplers have been developed that assign multiple data points to clusters at once. In this paper we introduce a new split-merge sampler, the triadic split-merge sampler, that perform steps between two and three randomly chosen clusters. This has two advantages. First, it reduces the asymmetry between the split and merge steps. Second, it is able to propose a new cluster that is composed out of data points from two different clusters. Both advantages speed up convergence which we demonstrate on a line extraction problem. We show that the triadic split-merge sampler outperforms the conventional split-merge sampler. Although this new MCMC sampler is demonstrated in this machine vision context, its application extend to the very general domain of statistical inference.

Speech emotion recognition (SER) poses one of the major challenges in human-machine interaction. We propose a new algorithm, the Voiced Segment Selection (VSS) algorithm, which can produce an accurate segmentation of speech signals. The VSS algorithm deals with the voiced signal segment as the texture image processing feature which is different from the traditional method. It uses the Log-Gabor filters to extract the voiced and unvoiced features from spectrogram to make the classification. The finding shows that the VSS method is a more accurate algorithm for voiced segment detection. Therefore, it has potential to improve performance of emotion recognition from speech.

In computer vision there are many sophisticated methods to perform inference over multiple lines, however they are quite ad-hoc. In this paper a fully Bayesian approach is used to fit multiple lines to a point cloud simultaneously. Our model extends a linear Bayesian regression model to an infinite mixture model and uses a Dirichlet process as a prior for the partition. We perform Gibbs sampling over non-unique parameters as well as over clusters to fit lines of a fixed length, a variety of orientations, and a variable number of data points. The performance is measured using the Rand Index, the Adjusted Rand Index, and two other clustering performance indicators. This paper is mainly meant to demonstrate that general Bayesian methods can be used for line estimation. Bayesian methods, namely, given a model and noise, perform optimal inference over the data. Moreover, rather than only demonstrating the concept as such, the first results are promising with respect to the described clustering performance indicators. Further research is required to extend the method to inference over multiple line segments and multiple volumetric objects that will need to be built on the mathematical foundation that has been laid down in this paper.

The tridiagonal solver is an important kernel in many scientific and engineering applications. Although quite a few parallel algorithms have been exploited recently, challenges still remain when solving tridiagonal systems on many-core architectures. In this paper, quantitative analysis is conducted to guide the selection of algorithms on different architectures, and a fast register-PCR-pThomas algorithm for Intel MIC architecture is presented to achieve the best utilization of in-core vectorization and registers. By choosing the most suitable number of PCR reductions, we further propose an improved algorithm, named register-PCR-half-pThomas algorithm, which minimizes the computational cost and the number of registers for use. The optimizations of manual prefetching and strength reduction are also discussed for the new algorithm. Evaluation on Intel Xeon Phi 7110P shows that our register-PCR-pThomas can outperform the MKL solver by 7.7 times in single precision and 3.4 times in double precision. Moreover, our optimized register-PCR-half-pThomas can further boost the performance by another 43.9% in single precision and 20.4% in double precision. Our best algorithm can outperform the latest official GPU library (cuSPARSE) on Nvidia K40 by 3.7 times and 2.4 times in single and double precision, respectively.

Diagnosis by comparison is a realistic approach to detect faults of multiprocessor systems. This paper considers a pessimistic diagnostic strategy for hypercube-like multiprocessor systems under the comparison model. The pessimistic strategy is a diagnostic process whereby all faulty nodes can be correctly identified and at most one fault-free node may be misjudged as a faulty node. We propose a pessimistic diagnosis algorithm based on the largest component in the faulty system. For a system with$N=2^n$nodes and$n\ge 5$, when the number of faulty nodes is bounded by$2n-2$, the algorithm can correctly identify all nodes except at most one node left undiagnosed. The time complexity of the algorithm is$O(Nlog_2N)$.

In this article, we discuss the performance modeling and optimization of Sparse Matrix-Vector Multiplication ( ) on NVIDIA GPUs using CUDA. has a very low computation-data ratio and its performance is mainly bound by the memory bandwidth. We propose optimization of based on ELLPACK from two aspects: (1) enhanced performance for the dense vector by reducing cache misses, and (2) reduce accessed matrix data by index reduction. With matrix bandwidth reduction techniques, both cache usage enhancement and index compression can be enabled. For GPU with better cache support, we propose differentiated memory access scheme to avoid contamination of caches by matrix data. Performance evaluation shows that the combined speedups of proposed optimizations for GT-200 are 16% (single-precision) and 12.6% (double-precision) for GT-200 GPU, and 19% (single-precision) and 15% (double-precision) for GF-100 GPU.

An integrated approach for the parallel solution of large sparse systems arisen in finite element computations is presented. The approach includes a three-phase preprocessor and a macro dataflow execution scheme. The three phases of the preprocessor are: (1) Extracting parallelism by means of an automatic domain decomposer; (2) Building the distributed data structure and (partial) scheduling for parallel computation during symbolic factorization; (3) Assigning processes (tasks) onto processors. The proposed approach has been implemented in the finite element analysis software package DIANA. Experimental results show that this integrated approach is an efficient method for both shared-and distributed-memory parallel systems.

A pipelined vector-reduction method that is based on L.M. Ni and K. Hwang's (1985) symmetric and asymmetric reduction methods is discussed. It is shown that the proposed method is the fastest among known pipelined vector-reduction methods.>

A number of new algorithms to make an online CORDIC implementation are introduced. The online CORDIC algorithm takes n+6 clock cycles to compute a CORDIC function. It is estimated that an implementation of the proposed algorithm is six to seven times as fast as the traditional CORDIC approach for Given's rotation, and seven to eight times as fast for SVD computation.>

The CORDIC algorithms provide a fast way to calculate a number of basic arithmetic functions. A CORDIC calculation takes O(n) steps for a function, where n is the word length of the operands. The speed is limited by the carry propagation in the adders and the I/O throughput. Speed can be improved by introducing redundancy in the calculation circuitry and I/O throughput by doing I/O transfers while calculating. The latter is characteristic for the class of so-called online arithmetic. At the same time, the pin requirements are limited to a single digit per operand. The authors introduce a number of new algorithms that provide an online CORDIC implementation.>

This paper describes a new online division (reciprocal) algorithm for (maximally) redundant floating-point numbers of arbitrary radix. The algorithm works for normalized, quasi-normalized, and pseudo-normailized numbers and can therefore be applied in chained online compuatations. The online delay of die proposed algorithm is the smallest reported so far. The algorithm consista of two steps: the first m digits of the result are generated by a simple table lookup method; the remaining n-m digits are generated by using an adapted Newton-Raphson iteration method. In the second step, the online digits are created by using a fast and simple selection mechanism.