Search

Question answering (QA) plays a vital role in biomedical natural language processing. Among question answering tasks, the retrieval question answering (ReQA) aims to directly retrieve the correct answer from candidates and has attracted much attention in the community for its efficiency. Recently, researchers have introduced ReQA into the biomedical domain as BioReQA. Typically BioReQA models rely on the dual-encoder to gain semantic representation and are trained following the settings of dense retrieval. However, they normally utilize easy in-batch negative samples in training process to avoid the extra forwarding cost and GPU memory required by encoding additional negative samples. However, hard negative samples have been proved more important with regard to the overall performance of BioReQA tasks. Therefore in this research, we focus on effectively constructing hard in-batch negative samples. Inspired by the classic linear assignment problem, we propose an Iterative Linear Assignment Grouping (ILAG) algorithm to construct hard in-batch negative samples. To further enhance performance for given hard batches in a low-resource scenario, we also employ adversarial training to augment the difficulty of batches. Extensive experiments have shown our proposed method's promising potential in the area of biomedical retrieval question answering.

The problem of data heterogeneity is one of the main challenges facing federated learning (FL). Non-IID data usually introduce bias in the training process of FL models, which can impact the accuracy and convergence speed of the models. To this end, we propose a personalized federated learning (PFL) algorithm with adaptive dynamic adjustment of the gradient clipping boundaries and the idea of freezing to reduce the influence of non-IID data on the model, called PerFreezeClip. PerFreezeClip is a design decision regarding parameter architecture, comparing the private and federated models. PerFreezeClip facilitates the training of each device based on an adaptive clipping gradient during training, with more rational updates and more stable gradients. The results based on the CIFAR-10 and CIFAR-100 datasets show that the proposed PerFreezeClip algorithm provides higher test accuracy after controlling the gradient: a maximum of a 50% enhancement compared to typical federated learning (non-personalized) algorithms. [ABSTRACT FROM AUTHOR]

Purpose: To present the detailed implementation processes of the IDEAL algorithm for two-dimensional compressible flows based on Delaunay triangular mesh, and compare the performance of the SIMPLE and IDEAL algorithms for solving compressible problems. What's more, the implementation processes of Delaunay mesh generation and derivation of the pressure correction equation are also introduced. Design/methodology/approach: Programming completely in C++. Findings: Five compressible examples are used to test the SIMPLE and IDEAL algorithms, and the comparison with measurement data shows good agreement. The IDEAL algorithm has much better performance in both convergence rate and stability over the SIMPLE algorithm. Originality/value: The detail solution procedure of implementing the IDEAL algorithm for compressible flows based on Delaunay triangular mesh is presented in this work, seemingly first in the literature. [ABSTRACT FROM AUTHOR]

Markov models are extensively used for categorical sequence clustering and classification due to their inherent ability to capture complex chronological dependencies hidden in sequential data. Existing Markov models are based on an implicit assumption that the probability of the next state depends on the preceding context/pattern which is consist of consecutive states. This restriction hampers the models since some patterns, disrupted by noise, may be not frequent enough in a consecutive form, but frequent in a sparse form, which can not make use of the information hidden in the sequential data. A sparse pattern corresponds to a pattern in which one or some of the state(s) between the first and last one in the pattern is/are replaced by wildcard(s) that can be matched by a subset of values in the state set. In this paper, we propose a new model that generalizes the conventional Markov approach making it capable of dealing with the sparse pattern and handling the length of the sparse patterns adaptively, i.e. allowing variable length pattern with variable wildcards. The model, named Dynamic order Markov model (DOMM), allows deriving a new similarity measure between a sequence and a set of sequences/cluster. DOMM builds a sparse pattern from sub-frequent patterns that contain significant statistical information veiled by the noise. To implement DOMM, we propose a sparse pattern detector (SPD) based on the probability suffix tree (PST) capable of discovering both sparse and consecutive patterns, and then we develop a divisive clustering algorithm, named DMSC, for Dynamic order Markov model for categorical sequence clustering. Experimental results on real-world datasets demonstrate the promising performance of the proposed model. [ABSTRACT FROM AUTHOR]

Summary: Robust centralized and weighted observation fusion (CAWOF) prediction algorithm is addressed in this article for an uncertain multi‐sensor generalized system with linear correlation between observation noises and an input white noise. This uncertainty in the generalized system primarily means that the variances of the aforementioned types of noise, as well as the multiplicative noise variances, are uncertain. Through singular value decomposition and virtual noise compensation, the original generalized system is changed to non‐generalized reduced‐order subsystems in which only noise variances are uncertain. Utilizing the minimax robustness estimation criterion, robust CAWOF Kalman predictors are put forward on account of the first subsystem with conservative upper bounds of noise variances. Eventually, robust observation fusion Kalman predictors of the original generalized system are proposed. The Lyapunov equation method is applied to verify two fusion predictors' robustness. With regard to all permissible uncertain practical noise variances, CAWOF predictors are robust, namely, the practical prediction error variances of two robust predictors will have minimum upper bounds. This equivalence between CAWOF Kalman predictors is proved by an information filter. In this article, the precision relationship of fusion predictors is given. Meanwhile, robust Kalman predictors for steady‐state case are proposed, and the astringency of robust time‐variant Kalman predictors is analyzed through the analysis of dynamic error system. The validity and correctness of proposed algorithm are proved by the simulation example of random dynamic input and output system in an economic system. [ABSTRACT FROM AUTHOR]

Data centers are fundamental facilities that support high-performance computing and large-scale data processing. To guarantee that a data center can provide excellent properties of expanding and routing, the interconnection network of a data center should be designed elaborately. Herein, we propose a novel structure for the interconnection network of data centers that can be expanded with a variable coefficient, also known as a variable expanding structure (VES). A VES is designed in a hierarchical manner and built iteratively. A VES can include hundreds of thousands and millions of servers with only a few layers. Meanwhile, a VES has an extremely short diameter, which implies better performance on routing between every pair of servers. Furthermore, we design an address space for the servers and switches in a VES. In addition, we propose a construction algorithm and routing algorithm associated with the address space. The results and analysis of simulations verify that the expanding rate of a VES depends on three factors: n, m, and k where the n is the number of ports on a switch, the m is the expanding speed and the k is the number of layers. However, the factor m yields the optimal effect. Hence, a VES can be designed with factor m to achieve the expected expanding rate and server scale based on the initial planning objectives. [ABSTRACT FROM AUTHOR]