Author name cluster

Kai Feng

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

8 papers

1 author row

EAAI Journal 2024 Journal Article

Prompt-based learning framework for zero-shot cross-lingual text classification

Kai Feng
Lan Huang
Kangping Wang
Wei Wei
Rui Zhang

Cross-lingual text classification is a challenging task that aims to train classifiers with data in one language, known as the source language, and apply the acquired knowledge to data in another language, referred to as the target language. Recent advancements in multilingual pre-trained language models (PLMs) have made significant progress in addressing cross-lingual issues, and the application of prompt-based learning has further improved task performance. However, these models still face challenges such as the gap between cross-lingual classification tasks and pre-training tasks of PLMs, as well as issues related to scarce resources and data noise, which hinder the full exploitation of the implicit knowledge in PLMs. In this paper, we propose a Prompt-based Cross-lingual Learning (PCL) framework that combines language-agnostic continuous prompt learning with self-learning process. Specifically, PCL framework leverages language-agnostic prompts and PLMs to achieve semantic transfer between source and target languages. To enhance the semantic relationship between prompts and category labels, a label attention module is introduced. Additionally, a set of self-training rules is proposed, which includes a scoring function. In a few-shot setting, noisy data is dynamically filtered through scoring and ranking of the data. During each training iteration, both the model and scoring function weights are updated, further improving the discrimination capability of the model. In summary, the proposed PCL framework builds upon cross-lingual prompt learning, effectively removing noisy data and applying it to zero-shot cross-lingual text classification, which is beneficial for engineering applications. The findings of this study have implications for prompt learning method. The PCL framework achieves state-of-the-art performance in cross-lingual text classification task, with a 14% performance improvement compared to basic soft prompt learning. This demonstrates its potential in addressing classification problems in resource-limited scenarios.

Details DOI

TCS Journal 2024 Journal Article

Subnetwork reliability of the arrangement graphs under probabilistic fault condition

Kai Feng
Zhangjian Ji
Xuebin Lv
Guozhen Zhang
Wei Wei

The subnetwork reliability of the interconnection network of a multiprocessor system is an important factor for processing the computing tasks. In order to measure reliability of the arbitrary size subnetwork of the ( n, k ) -arrangement graph A n, k, the probability R n, k m ( p ) of existing fault-free A n − m, k − m subnetworks in A n, k under the probabilistic fault condition is studied for m ≥ 1. An upper bound and a lower bound of R n, k m ( p ) are derived, and a heuristic algorithm based on the Monte Carlo simulation to calculate R n, k m ( p ) is proposed. Experimental results show that the upper and lower bounds of R n, k m ( p ) are in accordance with the results of the heuristic algorithm when the upper and lower bounds are close; otherwise, the results of the heuristic algorithm are relatively accurate. Under the consideration that high accuracy requirement and large size of A n, k may greatly reduce the evaluation efficiency when the heuristic algorithm is used, the back propagation neural network model for evaluating R n, k m ( p ) is constructed, and the experiments on the generated data indicate that the constructed model has certain advantages in balancing accuracy and efficiency.

Details DOI

TCS Journal 2021 Journal Article

Subnetwork reliability analysis of bubble-sort graph networks

Kai Feng
Xinyu Ma
Wei Wei

The bubble-sort graph network B n is recognized as an attractive interconnection network topology for building multiprocessor computer systems. In this paper, the subnetwork reliability of B n is analyzed in the presence of node failures. An upper bound and a lower bound on the B n − 1 subnetwork reliability of B n are established under the probability fault model, and the theoretical results are proved to be in accordance with and near to the simulation results. Moreover, under the node fault model, the mean time to failure to maintain the fault-free status of different number of disjoint B n − 1 subnetworks in B n is evaluated by using the fixed partitioning and the flexible partitioning, respectively, and it is shown that B n has better robustness under the flexible partitioning pattern when disjoint B n − 1 subnetworks need to be used.

Details DOI

AAAI Conference 2019 Conference Paper

SVM-Based Deep Stacking Networks

Jingyuan Wang
Kai Feng
Junjie Wu

The deep network model, with the majority built on neural networks, has been proved to be a powerful framework to represent complex data for high performance machine learning. In recent years, more and more studies turn to nonneural network approaches to build diverse deep structures, and the Deep Stacking Network (DSN) model is one of such approaches that uses stacked easy-to-learn blocks to build a parameter-training-parallelizable deep network. In this paper, we propose a novel SVM-based Deep Stacking Network (SVM-DSN), which uses the DSN architecture to organize linear SVM classifiers for deep learning. A BP-like layer tuning scheme is also proposed to ensure holistic and local optimizations of stacked SVMs simultaneously. Some good math properties of SVM, such as the convex optimization, is introduced into the DSN framework by our model. From a global view, SVM-DSN can iteratively extract data representations layer by layer as a deep neural network but with parallelizability, and from a local view, each stacked SVM can converge to its optimal solution and obtain the support vectors, which compared with neural networks could lead to interesting improvements in anti-saturation and interpretability. Experimental results on both image and text data sets demonstrate the excellent performances of SVM-DSN compared with some competitive benchmark models.

PDF Details

TCS Journal 2017 Journal Article

Strong matching preclusion for non-bipartite torus networks

Kai Feng

The strong matching preclusion number of a graph is the minimum number of vertices and edges whose deletion results in the remaining graph that has neither perfect matchings nor almost perfect matchings. The torus network is one of the most popular interconnection network topologies for massively parallel computing systems because of its desirable properties. It is known that bipartite torus networks have low strong matching preclusion numbers. Hu et al. [13] proved that non-bipartite torus networks with an odd number of vertices have good strong matching preclusion properties. To complete the study of strong matching preclusion problem for non-bipartite torus networks, in this paper, we establish the strong matching preclusion number and classify all optimal strong matching preclusion sets for the n-dimensional non-bipartite torus network with an even number of vertices, where n ≥ 3.

Details DOI

TCS Journal 2014 Journal Article

Fault tolerance in the arrangement graphs

Shiying Wang
Kai Feng

Let n and k be positive integers with n − k ⩾ 1. The arrangement graph A n, k is recognized as an attractive interconnection network. Let f m be the minimum number of faulty vertices that make every sub-arrangement graph A n − m, k − m faulty in A n, k under vertex-failure model. In this paper, we prove that f 0 = 1, f 1 = n, f n − 2 = n! / 2, and n! / ( n − m )! ⩽ f m ⩽ ( k − 1 m − 1 ) n! / ( n − m )! − 2 ( k − 2 m − 1 ) n! / ( n − m + 1 )! for 2 ⩽ m ⩽ k − 1.

Details DOI

TCS Journal 2014 Journal Article

Strong matching preclusion for torus networks

Shiying Wang
Kai Feng

The torus network is one of the most popular interconnection network topologies for massively parallel computing systems. Strong matching preclusion that additionally permits more destructive vertex faults in a graph is a more extensive form of the original matching preclusion that assumes only edge faults. In this paper, we establish the strong matching preclusion number and all minimum strong matching preclusion sets for bipartite torus networks and 2-dimensional nonbipartite torus networks.

Details DOI

TCS Journal 2012 Journal Article

Fault tolerance in k -ary n -cube networks

Shiying Wang
Guozhen Zhang
Kai Feng

The k -ary n -cube Q n k is one of the most commonly used interconnection topologies for parallel and distributed computing systems. Let f ( n, m ) be the minimum number of faulty nodes that make every ( n − m ) -dimensional subcube Q n − m k faulty in Q n k under node-failure models. In this paper, we prove that f ( n, 0 ) = 1, f ( n, 1 ) = k for odd k ≥ 3, f ( n, n − 1 ) = k n − 1 for odd k ≥ 3, and k m ≤ f ( n, m ) ≤ ( n − 1 m − 1 ) k m − ( n − 2 m − 1 ) k m − 1 for odd k ≥ 3.

Details DOI