Large Language Models (LLMs) have shown remarkable potential in code generation, making them increasingly important in the field. However, the security issues of generated code have not been fully addressed, and the usability of LLMs in code generation still requires further exploration. This work introduces SecCode, a framework that leverages an innovative interactive encouragement prompting (EP) technique for secure code generation with only NL prompts. This approach ensures that the prompts can be easily shared and understood by general users. SecCode functions through three stages: 1) Code Generation using NL Prompts; 2) Code Vulnerability Detection and Fixing, utilising our proposed encouragement prompting; 3) Vulnerability Cross-Checking and Code Security Refinement. These stages are executed in multiple interactive iterations to progressively enhance security. By using both proprietary LLMs (i.e., GPT-3.5 Turbo, GPT-4 and GPT-4o) and open-source LLMs (i.e., Llama 3.1 8B Instruct, DeepSeek Coder V2 Lite Instruct) evaluated on three benchmark datasets, extensive experimental results show that our proposed SecCode greatly outperforms compared baselines, generating secure code with a high vulnerability correction rate. For example, SecCode exhibits a high fix success rate of over 76% after running 5 automated EP interactive iterations and over 89% after running 10 automated EP interactive iterations. To the best of our knowledge, this work is the first to formulate secure code generation with NL prompts only. We have open-sourced our code and encourage the community to focus on secure code generation.

Artificial Intelligence (AI) has advanced significantly in various domains like healthcare, finance, and cybersecurity, with successes such as DeepMind's medical imaging and Tesla's autonomous vehicles. As telecommunications transition from 5G to 6G, integrating AI is crucial for complex demands like data processing, network optimization, and security. Despite ongoing research, there's a gap in consolidating AI-enabled Spectrum Management (AISM) advancements. Traditional spectrum management methods are inadequate for 6G due to its dynamic and complex demands, making AI essential for spectrum optimization, security, and network efficiency. This study aims to address this gap by: (i) Conducting a systematic review of AISM methodologies, focusing on learning models, data handling techniques, and performance metrics. (ii) Examining security and privacy concerns related to AI and traditional network threats within AISM contexts. Using the Systematic Literature Review (SLR) methodology, we meticulously analyzed 110 primary studies to: (a) Identify AI's utility in spectrum management. (b) Develop a taxonomy of AI approaches. (c) Classify datasets and performance metrics used. (d) Detail security and privacy threats and countermeasures. Our findings reveal challenges such as under-explored AI usage in critical AISM systems, computational resource demands, transparency issues, the need for real-world datasets, imbalances in security and privacy research, and the absence of testbeds, benchmarks, and security analysis tools. Addressing these challenges is vital for maximizing AI's potential in advancing 6G technology.

Deep learning solutions in critical domains like autonomous vehicles, facial recognition, and sentiment analysis require caution due to the severe consequences of errors. Research shows these models are vulnerable to adversarial attacks, such as data poisoning and neural trojaning, which can covertly manipulate model behavior, compromising reliability and safety. Current defense strategies like watermarking have limitations: they fail to detect all model modifications and primarily focus on attacks on CNNs in the image domain, neglecting other critical architectures like RNNs. To address these gaps, we introduce DeepiSign-G, a versatile watermarking approach designed for comprehensive verification of leading DNN architectures, including CNNs and RNNs. DeepiSign-G enhances model security by embedding an invisible watermark within the Walsh-Hadamard transform coefficients of the model's parameters. This watermark is highly sensitive and fragile, ensuring prompt detection of any modifications. Unlike traditional hashing techniques, DeepiSign-G allows substantial metadata incorporation directly within the model, enabling detailed, self-contained tracking and verification. We demonstrate DeepiSign-G's applicability across various architectures, including CNN models (VGG, ResNets, DenseNet) and RNNs (Text sentiment classifier). We experiment with four popular datasets: VGG Face, CIFAR10, GTSRB Traffic Sign, and Large Movie Review. We also evaluate DeepiSign-G under five potential attacks. Our comprehensive evaluation confirms that DeepiSign-G effectively detects these attacks without compromising CNN and RNN model performance, highlighting its efficacy as a robust security measure for deep learning applications. Detection of integrity breaches is nearly perfect, while hiding only a bit in approximately 1

Transformer-based text classifiers like BERT, Roberta, T5, and GPT-3 have shown impressive performance in NLP. However, their vulnerability to adversarial examples poses a security risk. Existing defense methods lack interpretability, making it hard to understand adversarial classifications and identify model vulnerabilities. To address this, we propose the Interpretability and Transparency-Driven Detection and Transformation (IT-DT) framework. It focuses on interpretability and transparency in detecting and transforming textual adversarial examples. IT-DT utilizes techniques like attention maps, integrated gradients, and model feedback for interpretability during detection. This helps identify salient features and perturbed words contributing to adversarial classifications. In the transformation phase, IT-DT uses pre-trained embeddings and model feedback to generate optimal replacements for perturbed words. By finding suitable substitutions, we aim to convert adversarial examples into non-adversarial counterparts that align with the model's intended behavior while preserving the text's meaning. Transparency is emphasized through human expert involvement. Experts review and provide feedback on detection and transformation results, enhancing decision-making, especially in complex scenarios. The framework generates insights and threat intelligence empowering analysts to identify vulnerabilities and improve model robustness. Comprehensive experiments demonstrate the effectiveness of IT-DT in detecting and transforming adversarial examples. The approach enhances interpretability, provides transparency, and enables accurate identification and successful transformation of adversarial inputs. By combining technical analysis and human expertise, IT-DT significantly improves the resilience and trustworthiness of transformer-based text classifiers against adversarial attacks.

Large-scale Machine Learning (ML) based Software Systems are increasingly developed by distributed teams situated in different trust domains. Insider threats can launch attacks from any domain to compromise ML assets (models and datasets). Therefore, practitioners require information about how and by whom ML assets were developed to assess their quality attributes such as security, safety, and fairness. Unfortunately, it is challenging for ML teams to access and reconstruct such historical information of ML assets (ML provenance) because it is generally fragmented across distributed ML teams and threatened by the same adversaries that attack ML assets. This paper proposes ProML, a decentralised platform that leverages blockchain and smart contracts to empower distributed ML teams to jointly manage a single source of truth about circulated ML assets' provenance without relying on a third party, which is vulnerable to insider threats and presents a single point of failure. We propose a novel architectural approach called Artefact-as-a-State-Machine to leverage blockchain transactions and smart contracts for managing ML provenance information and introduce a user-driven provenance capturing mechanism to integrate existing scripts and tools to ProML without compromising participants' control over their assets and toolchains. We evaluate the performance and overheads of ProML by benchmarking a proof-of-concept system on a global blockchain. Furthermore, we assessed ProML's security against a threat model of a distributed ML workflow.

ML-based Phishing URL (MLPU) detectors serve as the first level of defence to protect users and organisations from being victims of phishing attacks. Lately, few studies have launched successful adversarial attacks against specific MLPU detectors raising questions on their practical reliability and usage. Nevertheless, the robustness of these systems has not been extensively investigated. Therefore, the security vulnerabilities of these systems, in general, remain primarily unknown that calls for testing the robustness of these systems. In this article, we have proposed a methodology to investigate the reliability and robustness of 50 representative state-of-the-art MLPU models. First, we have proposed a cost-effective Adversarial URL generator URLBUG that created an Adversarial URL dataset ( $Adv_\text{data}$ ) . Subsequently, we reproduced 50 MLPU (traditional ML and Deep learning) systems and recorded their baseline performance. Lastly, we tested the considered MLPU systems on $Adv_\text{data}$ and analyzed their robustness and reliability using box plots and heat maps. Our results showed that the generated adversarial URLs have valid syntax and can be registered at a median annual price of ${\$}$ 11.99, and out of 13% of the already registered adversarial URLs, 63.94% were used for malicious purposes. Moreover, the considered MLPU models Matthew Correlation Coefficient (MCC) dropped from median 0.92 to 0.02 when tested against $Adv_\text{data}$ , indicating that the baseline MLPU models are unreliable in their current form. Further, our findings identified several security vulnerabilities of these systems and provided future directions for researchers to design dependable and secure MLPU systems.

Context : Research at the intersection of cybersecurity, Machine Learning (ML), and Software Engineering (SE) has recently taken significant steps in proposing countermeasures for detecting sophisticated data exfiltration attacks. It is important to systematically review and synthesize the ML-based data exfiltration countermeasures for building a body of knowledge on this important topic. Objective : This article aims at systematically reviewing ML-based data exfiltration countermeasures to identify and classify ML approaches, feature engineering techniques, evaluation datasets, and performance metrics used for these countermeasures. This review also aims at identifying gaps in research on ML-based data exfiltration countermeasures. Method : We used Systematic Literature Review (SLR) method to select and review 92 papers. Results : The review has enabled us to: (a) classify the ML approaches used in the countermeasures into data-driven, and behavior-driven approaches; (b) categorize features into six types: behavioral, content-based, statistical, syntactical, spatial, and temporal; (c) classify the evaluation datasets into simulated, synthesized, and real datasets; and (d) identify 11 performance measures used by these studies. Conclusion : We conclude that: (i) The integration of data-driven and behavior-driven approaches should be explored; (ii) There is a need of developing high quality and large size evaluation datasets; (iii) Incremental ML model training should be incorporated in countermeasures; (iv) Resilience to adversarial learning should be considered and explored during the development of countermeasures to avoid poisoning attacks; and (v) The use of automated feature engineering should be encouraged for efficiently detecting data exfiltration attacks.

Adversarial Examples (AEs) generated by perturbing original training examples are useful in improving the robustness of Deep Learning (DL) based models. Most prior works, generate AEs that are either unconscionable due to lexical errors or semantically or functionally deviant from original examples. In this paper, we present ReinforceBug, a reinforcement learning framework, that learns a policy that is transferable on unseen datasets and generates utility-preserving and transferable (on other models) AEs. Our results show that our method is on average 10% more successful as compared to the state-of-the-art attack TextFooler. Moreover, the target models have on average 73.64% confidence in the wrong prediction, the generated AEs preserve the functional equivalence and semantic similarity (83.38% ) to their original counterparts, and are transferable on other models with an average success rate of 46%.

Software Engineering researchers are increasingly using Natural Language Processing (NLP) techniques to automate Software Vulnerabilities (SVs) assessment using the descriptions in public repositories. However, the existing NLP-based approaches suffer from concept drift. This problem is caused by a lack of proper treatment of new (out-of-vocabulary) terms for the evaluation of unseen SVs over time. To perform automated SVs assessment with concept drift using SVs' descriptions, we propose a systematic approach that combines both character and word features. The proposed approach is used to predict seven Vulnerability Characteristics (VCs). The optimal model of each VC is selected using our customized time-based cross-validation method from a list of eight NLP representations and six well-known Machine Learning models. We have used the proposed approach to conduct large-scale experiments on more than 100,000 SVs in the National Vulnerability Database (NVD). The results show that our approach can effectively tackle the concept drift issue of the SVs' descriptions reported from 2000 to 2018 in NVD even without retraining the model. In addition, our approach performs competitively compared to the existing word-only method. We also investigate how to build compact concept-drift-aware models with much fewer features and give some recommendations on the choice of classifiers and NLP representations for SVs assessment.

Background: Over the year, Machine Learning Phishing URL classification (MLPU) systems have gained tremendous popularity to detect phishing URLs proactively. Despite this vogue, the security vulnerabilities of MLPUs remain mostly unknown. Aim: To address this concern, we conduct a study to understand the test time security vulnerabilities of the state-of-the-art MLPU systems, aiming at providing guidelines for the future development of these systems. Method: In this paper, we propose an evasion attack framework against MLPU systems. To achieve this, we first develop an algorithm to generate adversarial phishing URLs. We then reproduce 41 MLPU systems and record their baseline performance. Finally, we simulate an evasion attack to evaluate these MLPU systems against our generated adversarial URLs. Results: In comparison to previous works, our attack is: (i) effective as it evades all the models with an average success rate of 66% and 85% for famous (such as Netflix, Google) and less popular phishing targets (e.g., Wish, JBHIFI, Officeworks) respectively; (ii) realistic as it requires only 23ms to produce a new adversarial URL variant that is available for registration with a median cost of only $11.99/year. We also found that popular online services such as Google SafeBrowsing and VirusTotal are unable to detect these URLs. (iii) We find that Adversarial training (successful defence against evasion attack) does not significantly improve the robustness of these systems as it decreases the success rate of our attack by only 6% on average for all the models. (iv) Further, we identify the security vulnerabilities of the considered MLPU systems. Our findings lead to promising directions for future research. Conclusion: Our study not only illustrate vulnerabilities in MLPU systems but also highlights implications for future study towards assessing and improving these systems.

The purpose of this study is to compare the conventional crop monitoring system with the satellite based crop monitoring system in Pakistan. This study is conducted for SUPARCO (Space and Upper Atmosphere Research Commission). The study focused on the wheat crop, as it is the main cash crop of Pakistan and province of Punjab. This study will answer the following: Which system is better in terms of cost, time and man power? The man power calculated for Punjab CRS is: 1,418 personnel and for SUPARCO: 26 personnel. The total cost calculated for SUPARCO is almost 13.35 million and CRS is 47.705 million. The man hours calculated for CRS (Crop Reporting Service) are 1,543,200 hrs (136 days) and man hours for SUPARCO are 8, 320hrs (40 days). It means that SUPARCO workers finish their work 96 days earlier than CRS workers. The results show that the satellite based crop monitoring system is efficient in terms of manpower, cost and time as compared to the conventional system, and also generates early crop forecasts and estimations. The research instruments used included: Interviews, physical visits, group discussions, questionnaires, study of reports and work flows. A total of 93 employees were selected using Yamane’s formula for data collection, which is done with the help questionnaires and interviews. Comparative graphing is used for the analysis of data to formulate the results of the research. The research findings also demonstrate that although conventional methods have a strong impact still in Pakistan (for crop monitoring) but it is the time to bring a change through technology, so that our agriculture will also be developed along modern lines.

Precise and fast shape matching and retrieval from very large datasets is a challenging task because of the existence of many distortions such as noise, occlusion and affine distortions. In this paper, we aim to propose a time-saving and effective shape matching and retrieval framework, that employs pruning which will enable online shape retrieval from extremely large datasets. First, using a hierarchical tree-based structure supporting parallel processing and efficient feature descriptors, irrelevant shapes are pruned and a subset of shapes relevant to the query is selected, then using more sophisticated feature descriptors, accurate retrieval is performed. Contour representation of an object is considered as most significant visual shape similarity measure by the humans. Using boundary information, we generate two simplified and efficient feature descriptors for fast pruning, and a sophisticated feature descriptor for effective and accurate retrieval. Tests performed on standard datasets unveil that the proposed technique is computationally more efficient than the state-of-the-art techniques while maintaining comparable matching and retrieval performance. Its performance is scalable for huge datasets and is robust against affine transformations, articulations and occlusion.

Outdoor pedestrian detection is one of most important, primary and challenging preprocessing step for any automated visual surveillance activity ranging from event to activity detection. Currently used algorithms are resolution, scale and clothing dependent, their performance decreases as resolution of camera decreases and the size of pedestrian decreases to small scale (30-80 pixels). Moreover, regions like Middle East where loose clothes are mostly used the performance of these systems get worse. The reason behind the failure of these systems is that most of them target the contour information of pedestrian and as the scale decreases or cloth varies the contour information becomes ambiguous. Noise accumulation, wide area view, low resolution, outdoor environment artifacts, low frame rate further add to the complexity of pedestrian detection. The paper proposes a pedestrian detection algorithm that detects the pedestrian of small scale from the low frame rate (5 frames per second) video captured by low resolution CCTV camera (352x288) resolution. The algorithm is clothing and illumination invariant and proposes two main contributions: first, motion cues and edge lets based contour detection (ECD) are used to target temporal and low level pixel details, handling clothing variation and providing a heuristic window for pedestrian detection and second, the heuristic window is searched for presence of pedestrian using linear support vector machine (SVM)classifier trained over hybrid of histogram of oriented gradients (HOG) and statistical shape based feature vector.