In the context of software development, code completion has become an essential functionality that helps speed up coding and reduce syntax errors. Most previously proposed code completion modules are based on rule-based techniques that utilize language grammar to suggest code. Recently, with the rise of LLMs, several code completion assistants have been made publicly available. While demonstrating strong performance, these assistants raise several concerns related to latency, privacy and data security due to their cloud-based nature. In this work, we present CodeWisp, a local code assistant based on the RetrievalAugmented Generation (RAG) paradigm. CodeWisp combines LLM inference with context-aware retrieval using Abstract Syntax Tree (AST)-guided segmentation and semantic indexing. To validate the effectiveness of the proposed AST-guided chunking approach in enhancing semantic retrieval and code generation quality, CodeWisp was evaluated using two embedding models on a dataset of source files spanning four programming languages. In Python CodeWisp's retreiver achieved a recall of 0.96 with nomic-embed-text and a Mean Reciprocal Rank (MRR) of 0.95.

Federated Learning (FL) allows multiple clients to collaboratively train a machine learning model without directly sharing their raw data. While federated learning provides a degree of data privacy, model updates are still susceptible to various inference attacks. This paper presents FLASS, a lightweight and decentralized federated learning framework that leverages additive secret sharing within a multi-server architecture. Each client encodes its local model update into additive shares and shares them with several non-colluding servers, which execute secure aggregation without gaining knowledge of individual contributions. FLASS does not rely on heavy-weight cryptography or a trusted authority. The security analysis and experiments demonstrate the effectiveness and efficiency of the suggested scheme. The analysis indicates that, while achieving the same accuracy as conventional FL schemes, FLASS ensures robust privacy protection with acceptable computational and communication overhead.

With the proliferation of IoT devices and the exponential growth of data generated at the edge, federated learning (FL) emerges as a powerful method for training machine learning models on decentralized data sources. In security-critical applications, such as anomaly and threat detection, ensuring the confidentiality and integrity of sensitive data is paramount. In this paper, we introduce VeSAFL, a novel scheme for verifiable secure aggregation for privacy-preserving FL designed for edge computing environments. VeSAFL decentralizes model updates across edge nodes, reducing dependency on centralized cloud servers and mitigating risks associated with single points of failure. By leveraging multi-server aggregators, our approach fortifies system resilience and reliability against potential cyber threats. To bolster trust in the learning process, we implement a robust verification mechanism that guarantees the integrity and authenticity of local and global updates. Our experimental results highlight the efficacy and efficiency of VeSAFL in safeguarding against active adversaries and accurately identifying anomalous activities. Furthermore, our comprehensive security analysis affirms the scheme's correctness, verifiability, and privacy preservation in adversarial scenarios.

Decentralized machine learning over peer-to-peer networks is very appealing for it enables to learn personalized models without sharing users data, nor relying on any central server. Peers can improve upon their locally trained model across a network graph of other peers with similar objectives. Whilst they offer an inherently scalable scheme with a very simple cost-efficient learning model, peer-to-peer networks are also fragile. In particular, they can be very easily disrupted by unfairness, free-riding, and adversarial behaviors. In this paper, we present CDPL (Contribution Driven P2P Learning), a novel Byzantine-resilient distributed algorithm to train personalized models across similar peers. We convey theoretically and empirically the effectiveness of CDPL in terms of speed of convergence as well as robustness to Byzantine behavior.

K-nearest neighbors is one of the most popular classification algorithms. It assumes that similar things or people are near to each other. One of the primordial steps in this algorithm is K value, which is given by the user. This value influences the algorithm result performance. In this paper, we suggest an enhanced approach that eliminates the use of K value with keeping the same performance and increasing it for a specific datasets type. We propose a combined approach named `Zone classifier' that provides an excellent performance which is estimated, in average, by more than 85%; for Iris, Wine, Digits, The breast cancer and Olivetti faces dataset.

In big data environments with big number of users and high volume of data, we need to manage the corresponding huge number of security policies. Due to the distributed management of these policies, they may contain several anomalies, such as conflicts and redundancies, which may lead to both safety and availability problems. The distributed systems guided by such security policies produce a huge number of access logs. Due to potential security breaches, the access logs may show the presence of non-allowed accesses. This may also be a consequence of conflicting rules in the security policies. In this paper, we present an ongoing work on developing an environment for verifying and correcting security policies. To make the approach efficient, an access log is used as input to determine suspicious parts of the policy that should be considered. The approach is also made efficient by clustering the policy and the access log and considering separately the obtained clusters. The clustering technique and the use of access log significantly reduces the complexity of the suggested approach, making it scalable for large amounts of data.

Today collaborative applications may enable collaboration among users from the same or different tenants of a given cloud provider. During such collaborations, the participants need to access and use resources held by other collaborating users. These resources often contain sensitive data. They are meant to be shared only during specific collaborative sessions. A collaborative session is an abstract entity, comprising a set of users, called members of the session, playing the same or different roles. These users may have concurrent access to the shared objects during a session depending on their roles. In this work, we propose an approach that ensures access control to the shared resources in a collaborative session in multi-tenants environments. We suggest CRBAC, the Collaboration Role-based Access Control. CRBAC consists of an extended version of the RBAC model. CRBAC defines new entities to support access control in collaborative sessions. The suggested model has been implemented within Swift component in the open source cloud-computing platform OpenStack.

In this paper, we present PDP, a distributed polling protocol that enables a set of participants to gather their opinion on a common interest without revealing their point of view. PDP does not rely on any centralized authority or on heavyweight cryptography. PDP is an overlay-based protocol where a subset of participants may use a simple sharing scheme to express their votes. In a system of M participants arranged in groups of size N where at least 2k-1 participants are honest, PDP bounds the probability for a given participant to have its vote recovered with certainty by a coalition of B dishonest participants by π(B/N)(k+1), where π is the proportion of participants splitting their votes, and k a privacy parameter. PDP bounds the impact of dishonest participants on the global outcome by 2(kα + BN), where represents the number of dishonest nodes using the sharing scheme.

The demand for the ubiquitous service is increasing due to the rapidly growing demand for increased data rates, mobile Internet and the diversity of wireless communication technologies. Also due to the challenges to interconnect heterogeneous network technologies and to offer ubiquitous services, telecommunications operators look after the best way to provide continuity of service during handover and how to give the mobile client the possibility to get the best connection anywhere and anytime. In this paper we propose an architecture and its implementation which guarantees the continuity of service during a communication in the context of heterogeneous access network technologies. The suggested solution named HaVe-2W3G (Handover Vertical WLAN WiMAX 3G) ensures a Vertical handover between heterogeneous access networks technologies: WLAN, WiMAX and 3G. A performance evaluation of such implementation is shown using a streaming application.*