Enhancing android malware detection explainability through function call graph APIs

Journal of Information Security and Applications, Journal Year: 2024, Volume and Issue: 80, P. 103691 - 103691

Published: Jan. 2, 2024

Nowadays, mobile devices are massively used in everyday activities. Thus, they contain sensitive data targeted by threat actors like bank accounts and personal information. Through the years, Machine Learning approaches have been proposed to identify malicious Android applications, but recent research highlights need for better explanations model decisions, as existing ones may not be related app's functionalities. This paper proposes an explainable approach based on static analysis detect malware. The novelty lies specific conducted select extract features (i.e., APIs taken from DEX Call Graph) that immediately provide meaningful of functionality, thus allowing a significant correlation malware behavior with its family. Moreover, since we number type features, distinct impacts each one appear more evident. attained results show it is possible reach comparable (in terms accuracy) state-of-the-art models while providing easy-to-understand explanations, which yield insights into functionalities samples.

Language: Английский

---

The Explainability of Transformers: Current Status and Directions

Computers, Journal Year: 2024, Volume and Issue: 13(4), P. 92 - 92

Published: April 4, 2024

An increasing demand for model explainability has accompanied the widespread adoption of transformers in various fields applications. In this paper, we conduct a survey existing literature on transformers. We provide taxonomy methods based combination transformer components that are leveraged to arrive at explanation. For each method, describe its mechanism and find out attention-based methods, both alone conjunction with activation-based gradient-based most employed ones. A growing attention is also devoted deployment visualization techniques help explanation process.

Language: Английский

---

Pulmonary Nodule Detection, Segmentation and Classification Using Deep Learning: A Comprehensive Literature Review

BioMedInformatics, Journal Year: 2024, Volume and Issue: 4(3), P. 2043 - 2106

Published: Sept. 13, 2024

Lung cancer is a leading cause of cancer-related deaths worldwide, emphasizing the significance early detection. Computer-aided diagnostic systems have emerged as valuable tools for aiding radiologists in analysis medical images, particularly context lung screening. A typical pipeline diagnosis involves pulmonary nodule detection, segmentation, and classification. Although traditional machine learning methods been deployed previous years with great success, this literature review focuses on state-of-the-art deep methods. The objective to extract key insights methodologies from studies that exhibit high experimental results domain. This paper delves into databases utilized, preprocessing steps applied, data augmentation techniques employed, proposed exceptional outcomes. reviewed predominantly harness cutting-edge methodologies, encompassing convolutional neural networks (CNNs) advanced variants such 3D CNNs, alongside other innovative approaches Capsule transformers. examined these reflect continuous evolution datasets, discussed here collectively contribute development more efficient computer-aided systems, empowering dfhealthcare professionals fight against deadly disease.

Language: Английский

---

PMANet: Malicious URL detection via post-trained language model guided multi-level feature attention network

Information Fusion, Journal Year: 2024, Volume and Issue: 113, P. 102638 - 102638

Published: Aug. 17, 2024

Language: Английский

---

Natural Language Understanding for Navigation of Service Robots in Low-Resource Domains and Languages: Scenarios in Spanish and Nahuatl

Mathematics, Journal Year: 2024, Volume and Issue: 12(8), P. 1136 - 1136

Published: April 10, 2024

Human–robot interaction is becoming increasingly common to perform useful tasks in everyday life. From the human–machine communication perspective, achieving effective natural language one challenge. To address it, processing strategies have recently been used, commonly following a supervised machine learning framework. In this context, most approaches rely on use of linguistic resources (e.g., taggers or embeddings), including training corpora. Unfortunately, such are scarce for some languages specific domains, increasing complexity solution approaches. Motivated by these challenges, paper explores deep methods understanding commands emitted service robots that guide their movements low-resource scenarios, defined Spanish and Nahuatl languages, which scarcely unavailable task. Particularly, we applied (NLU) techniques using neural networks transformers-based models. As part research methodology, introduced labeled dataset movement mentioned languages. The results show models based transformers work well recognize (intent classification task) parameters quantities units) Spanish, performance 98.70% (accuracy) 96.96% (F1) intent slot-filling tasks, respectively). Nahuatl, best obtained was 93.5% 88.57% respectively. general, study shows robot can be guided through cross-lingual transfer strategies, even scenarios.

Language: Английский

---

A Deep Learning Model for Predicting the State of Energy in Lithium-Ion Batteries Based on Magnetic Field Effects

Energy, Journal Year: 2024, Volume and Issue: 304, P. 132161 - 132161

Published: June 25, 2024

Language: Английский

---

Antibody design using deep learning: from sequence and structure design to affinity maturation

Briefings in Bioinformatics, Journal Year: 2024, Volume and Issue: 25(4)

Published: May 23, 2024

Abstract Deep learning has achieved impressive results in various fields such as computer vision and natural language processing, making it a powerful tool biology. Its applications now encompass cellular image classification, genomic studies drug discovery. While development traditionally focused deep on small molecules, recent innovations have incorporated the discovery of biological particularly antibodies. Researchers devised novel techniques to streamline antibody development, combining vitro silico methods. In particular, computational power expedites lead candidate generation, scaling potential against complex antigens. This survey highlights significant advancements protein design optimization, specifically focusing includes aspects design, folding, antibody–antigen interactions docking affinity maturation.

Language: Английский

---

A time-embedded attention-based transformer for crash likelihood prediction at intersections using connected vehicle data

Transportation Research Part C Emerging Technologies, Journal Year: 2024, Volume and Issue: 169, P. 104831 - 104831

Published: Sept. 11, 2024

Language: Английский

---

High-resolution Spatiotemporal Prediction of PM2.5 Concentration based on Mobile Monitoring and Deep Learning

Environmental Pollution, Journal Year: 2024, Volume and Issue: unknown, P. 125342 - 125342

Published: Nov. 1, 2024

Language: Английский

---

DCUFormer: Enhancing pavement crack segmentation in complex environments with dual-cross/upsampling attention

Expert Systems with Applications, Journal Year: 2024, Volume and Issue: unknown, P. 125891 - 125891

Published: Nov. 1, 2024

Language: Английский

---