Classification of Program Texts Represented as Markov Chains with Biology-Inspired Algorithms-Enhanced Extreme Learning Machines

Abstract

The massive nature of modern university programming courses increases the burden on academic workers. The Digital Teaching Assistant (DTA) system addresses this issue by automating unique programming exercise generation and checking, and provides means for analyzing programs received from students by the end of semester. In this paper, we propose a machine learning-based approach to the classification of student programs represented as Markov chains. The proposed approach enables real-time student submissions analysis in the DTA system. We compare the performance of different multi-class classification algorithms, such as support vector machine (SVM), the k nearest neighbors (KNN) algorithm, random forest (RF), and extreme learning machine (ELM). ELM is a single-hidden layer feedforward network (SLFN) learning scheme that drastically speeds up the SLFN training process. This is achieved by randomly initializing weights of connections among input and hidden neurons, and explicitly computing weights of connections among hidden and output neurons. The experimental results show that ELM is the most computationally efficient algorithm among the considered ones. In addition, we apply biology-inspired algorithms to ELM input weights fine-tuning in order to further improve the generalization capabilities of this algorithm. The obtained results show that ELMs fine-tuned with biology-inspired algorithms achieve the best accuracy on test data in most of the considered problems.

Analysis of Program Representations Based on Abstract Syntax Trees and Higher-Order Markov Chains for Source Code Classification Task

“…Modern static analyzers are often based on data mining and machine learning methods such as clustering algorithms [11,19] and classification algorithms [7,9,15,16,18] that automatically discover and learn patterns hidden in data. In [11], a clustering algorithm was used to infer syntactic program transformations from examples, with the aims of providing assistance with refactoring tasks and suggesting repairs to programming assignments in massive open online courses (MOOCs).…”

“…However, token-based feature extraction techniques do not take the hierarchical nature of program syntax into account, which can lead to the loss of useful information while transforming source code into vector-based representations, especially in problems similar to algorithm classification [18]. This is crucial, especially in cases when analyzed code snippets are obfuscated at the token level, in identifier name prediction [9], bug detection [7], code clone detection [15,16], and vulnerability detection [14] problems that can be reduced to the task classification problem.…”

“…In this paper we consider the task detection problem. Known applications of task detection algorithms include malicious software identification [13,14], detection of semantic code clones [15,16], and hardware-software partitioning [17]. The task detection problem can be informally defined similarly to the algorithm detection problem described in [18]: given a finite set of different task types, and a program which solves a unique task of one of the given types, find the type of the task that a program solves.…”

“…In this study we compare different approaches to source code transformation into vector-based representations, such as word2vec [21,25], code2vec [9,26,27], histograms of CPython assembly language instruction opcodes [18,28], histograms of abstract syntax tree (AST) node types, inspired by [28], and AST-based Markov chain models [15,16]. Additionally, we propose a general algorithm for source code transformation into vectors based on higher-order Markov chains, and include AST-based Markov chains of order 1 and 2 into the comparative study.…”

Disk Drives Remaining Useful Life Prediction Using the Extreme Learning Machine

“…In general, the learning process of an ELM network can be described as follows (Demidova & Gorchakov, 2022b): § randomly determine weights 𝛚 ! and biases 𝐛 !…”