arXiv subjects

Kanchana Thilakarathna

Publications and source records attributed to Kanchana Thilakarathna.

At least 19 recordsLinked to original sources

arXiv2026

How People Use ChatGPT in Australia: A WildChat Analysis

Generative AI chatbots are increasingly embedded in everyday life, yet most large-scale studies describe global patterns. This paper presents an Australia-focused analysis of WildChat, a public dataset of real-world ChatGPT interaction logs. Using descriptive analysis and a multi-layer classification scheme, we analysed 37,845 conversations identified as Australian, examining language diversity, work relevance, interaction intent, topic distribution, turn-taking, temporal change, work activities, and Australia-related domains. Our findings show that the Australian subset is strongly action-oriented and comparatively work-oriented, with most interactions classified as doing and a majority of conversations classified as work-related. The dataset also shows multilingual use and a growing presence of self-expression over time. Australia-related conversations frequently invoke local institutions, laws, regulators, education systems, companies, cultural references, and public services. Finally, we outline implications for future research, including local AI evaluation, multilingual participation, context-aware design, and safeguards for everyday high-stakes domains.

cs.HC↗

arXiv2026

FedVideoMAE: Efficient Federated Video Moderation with Differential Privacy and Secure Aggregation

Short-form video moderation is increasingly pushed toward edge and privacy-sensitive settings, where users may intend videos for a limited audience, such as friends or private groups, but sending raw clips to a central server can broaden exposure, consume bandwidth, and add moderation latency. Federated learning can keep videos on device, but unprotected model updates may still leak information, and full-video backbones are expensive to communicate. We present FedVideoMAE, a privacy-preserving federated framework for violence detection that adapts a frozen VideoMAE backbone with lightweight LoRA and prompt parameters. Each training round combines self-supervised masked video reconstruction with client-side differential privacy and pairwise masked aggregation (SA) of adapter updates. Violence labels are held out from federation and used only for downstream evaluation, separating private representation learning from supervised assessment. On RWF-2000, exchanging 5,518,848 trainable parameters instead of the 156,371,328-parameter instantiated pretraining state gives a 28.3x model-state payload ratio. FedVideoMAE reaches 77.25% test accuracy without DP or SA, while accuracy under DP+SA remains in the 65.25-66.00% range. Transfer experiments on RLVS and binary UCF-Crime show similar behavior. These results characterize the privacy-utility trade-off for edge video moderation. Code is available at: https://github.com/zyt-599/FedVideoMAE

cs.CV↗

arXiv2026

SUMI: Scalable Unified Model for 3D Point Cloud Inference

Point cloud completion commonly follows a coarse-to-fine paradigm, where a low-density coarse shape is first predicted and then upsampled to the target resolution. Although recent methods have improved global structure recovery, the fine stage often remains limited by simple upsampling and insufficient interaction with coarse structural features, making local detail reconstruction challenging. We propose SUMI, a diffusion-enhanced refinement module for coarse-to-fine point cloud completion. Unlike prior diffusion-based completion methods that use diffusion as a standalone point generator, SUMI injects noisy geometric features into cross-attention with coarse structural features, enabling reverse denoising to refine local geometry while preserving global consistency. SUMI can also be integrated into existing coarse-to-fine models as a flexible refinement module. Experiments on PCN, ShapeNet-55/34, and MVP demonstrate consistent improvements over strong baselines. SUMI achieves the best overall CD and F1-score on PCN, reduces CD by up to 16.1% on ShapeNet-55, and obtains the best CD across all output densities on MVP.

cs.CV↗

arXiv2026

Dependency Triad: A Metric to Quantify the Dependencies Between Attributes for Local Differential Privacy

Collecting multidimensional user data is essential for extracting rich insights across various applications. Local Differential Privacy (LDP) has emerged as a de facto standard for mitigating privacy risks in such scenarios. A key challenge in privacy-preserving multidimensional data collection lies in inter-attribute dependencies, as they can inadvertently reveal correlated information and increase privacy vulnerabilities. Therefore, accurately measuring correlation-induced privacy leakage (CPL) is essential for privacy analysis and privacy-utility trade-off. However, existing CPL analysis solutions either require accurate prior knowledge or face scalability challenges for large numbers of attributes and high-cardinality attributes. These limit their practical applicability in real data. To address this research gap, we propose a novel metric, ``Dependency Triad'' (DT), which summarizes the pairwise dependency information relevant to CPL using three parameters and yields a \emph{constant-time} conservative estimator of pairwise CPL. DT explicitly models uncertainty in prior distributional knowledge through its parameters, delivering robust leakage estimates. Moreover, its robustness to sparse distributions makes it particularly suitable for high-cardinality attributes, while the pairwise formulation serves as a tractable building block for assessing total leakage in multidimensional settings. Extensive experiments on both synthetic and real datasets demonstrate that DT consistently estimates CPL across diverse dependency regimes and prior uncertainties.

cs.CR↗

arXiv2026

RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment

Human Activity Recognition (HAR) from wearable sensors supports applications in healthcare, rehabilitation, fitness tracking, and smart environments. Yet, existing deep learning approaches require dataset-specific training, large labeled corpora, and repeated adaptation to new sensor settings or activity taxonomies. Retrieval-Augmented Generation for Human Activity Recognition (RAG-HAR) addresses this by framing HAR as a training-free, retrieval-augmented task, in which statistical descriptions of sensor windows are used to retrieve similar labeled examples that guide LLM-based classification. We introduce RAG-HAR+, a retrieval-first and cost-optimized extension that strengthens retrieval while reducing dependence on LLM-based inference. RAG-HAR+ uses an offline Retrieval Designer Agent to design dataset-specific feature groups from a diverse pool of motion descriptors, enabling sensor windows to be compared using features better aligned with dataset-specific activity patterns. During inference, RAG-HAR+ uses majority voting over retrieved neighbors for samples with strong retrieval evidence and defers only uncertain cases to an LLM-based Ambiguity Resolver Agent. Across six HAR benchmarks, RAG-HAR+ maintains competitive or improved performance while reducing LLM usage, token consumption, and inference time. We further extend the RAG-HAR mobile prototype to demonstrate the practical feasibility of retrieval-first, LLM-assisted HAR in mobile sensing scenarios.

cs.LG↗

arXiv2026

STELLA: Efficient Sensor-to-LLM Translation for On-Device Human Activity Recognition

HAR is increasingly expected to run continuously on edge devices, yet recent LLM-based methods remain hard to deploy: raw sensor prompts are long, cloud inference adds latency and privacy risk, and fine-tuned LLM pipelines turn general-purpose models into task-specific classifiers. We present STELLA, an efficient sensor-to-LLM translation framework for on-device HAR that shifts the burden from LLM adaptation to sensor tokenization. A lightweight hierarchical tokenizer compresses an entire multi-channel inertial window into a fixed set of compact latent sensor tokens, which are projected into the embedding space of a frozen pretrained LLM and combined with a natural-language prompt for label scoring. This preserves activity-relevant temporal and cross-channel structure while keeping LLM-side computation predictable across sensor configurations. STELLA also supports on-device personalization, adapting only the lightweight tokenizer on small amounts of user-specific labelled data and augmenting inference with a local retrieval context, keeping the LLM, user data, and retrieval on device. Across seven public HAR datasets and eight benchmark settings, STELLA achieves new state-of-the-art performance, improving over prior methods by up to 11.83% F1; on-device personalization yields up to a further 21.91% F1 as user data accumulates after deployment. STELLA also outperforms representative time-series tokenizers under the same LLM pipeline and achieves real-time inference under practical mobile and edge budgets, showing that efficient sensor tokenization is a practical path toward accurate, private, and personalized LLM-based HAR on edge devices.

cs.LG↗

arXiv2026

RAG-HAR: Retrieval Augmented Generation-based Human Activity Recognition

Human Activity Recognition (HAR) underpins applications in healthcare, rehabilitation, fitness tracking, and smart environments, yet existing deep learning approaches demand dataset-specific training, large labeled corpora, and significant computational resources.We introduce RAG-HAR, a training-free retrieval-augmented framework that leverages large language models (LLMs) for HAR. RAG-HAR computes lightweight statistical descriptors, retrieves semantically similar samples from a vector database, and uses this contextual evidence to make LLM-based activity identification. We further enhance RAG-HAR by first applying prompt optimization and introducing an LLM-based activity descriptor that generates context-enriched vector databases for delivering accurate and highly relevant contextual information. Along with these mechanisms, RAG-HAR achieves state-of-the-art performance across six diverse HAR benchmarks. Most importantly, RAG-HAR attains these improvements without requiring model training or fine-tuning, emphasizing its robustness and practical applicability. RAG-HAR moves beyond known behaviors, enabling the recognition and meaningful labelling of multiple unseen human activities.

cs.CV↗

arXiv2026

A Comprehensive Survey on Network Traffic Synthesis: From Statistical Models to Deep Learning

Synthetic network traffic generation has emerged as a promising alternative for various data-driven applications in the networking domain. It enables the creation of synthetic data that preserves real-world characteristics while addressing key challenges such as data scarcity, privacy concerns, and purity constraints associated with real data. In this survey, we provide a comprehensive review of synthetic network traffic generation approaches, covering essential aspects such as data types and generation models. With the rapid advancements in Artificial Intelligence (AI) and Machine Learning (ML), we focus particularly on deep learning (DL)-based techniques while also providing a detailed discussion of statistical methods and their extensions, including commercially available tools. We present a comprehensive comparision of generation approaches and provide an AI tool to apply this comparision for any network traffic generation papers. Furthermore, we highlight open challenges in this domain and discuss potential future directions for further research and development. This survey serves as a foundational resource for researchers and practitioners, offering a structured analysis of existing methods, challenges, and opportunities in synthetic network traffic generation.

cs.NI↗

arXiv2026

Memory Retrieval in Transformers: Insights from The Encoding Specificity Principle

While explainable artificial intelligence (XAI) for large language models (LLMs) remains an evolving field with many unresolved questions, increasing regulatory pressures have spurred interest in its role in ensuring transparency, accountability, and privacy-preserving machine unlearning. Despite recent advances in XAI have provided some insights, the specific role of attention layers in transformer based LLMs remains underexplored. This study investigates the memory mechanisms instantiated by attention layers, drawing on prior research in psychology and computational psycholinguistics that links Transformer attention to cue based retrieval in human memory. In this view, queries encode the retrieval context, keys index candidate memory traces, attention weights quantify cue trace similarity, and values carry the encoded content, jointly enabling the construction of a context representation that precedes and facilitates memory retrieval. Guided by the Encoding Specificity Principle, we hypothesize that the cues used in the initial stage of retrieval are instantiated as keywords. We provide converging evidence for this keywords-as-cues hypothesis. In addition, we isolate neurons within attention layers whose activations selectively encode and facilitate the retrieval of context-defining keywords. Consequently, these keywords can be extracted from identified neurons and further contribute to downstream applications such as unlearning.

cs.LG↗

arXiv2026

SHIELD: An Auto-Healing Agentic Defense Framework for LLM Resource Exhaustion Attacks

Sponge attacks increasingly threaten LLM systems by inducing excessive computation and DoS. Existing defenses either rely on statistical filters that fail on semantically meaningful attacks or use static LLM-based detectors that struggle to adapt as attack strategies evolve. We introduce SHIELD, a multi-agent, auto-healing defense framework centered on a three-stage Defense Agent that integrates semantic similarity retrieval, pattern matching, and LLM-based reasoning. Two auxiliary agents, a Knowledge Updating Agent and a Prompt Optimization Agent, form a closed self-healing loop, when an attack bypasses detection, the system updates an evolving knowledgebase, and refines defense instructions. Extensive experiments show that SHIELD consistently outperforms perplexity-based and standalone LLM defenses, achieving high F1 scores across both non-semantic and semantic sponge attacks, demonstrating the effectiveness of agentic self-healing against evolving resource-exhaustion threats.

cs.CR↗

arXiv2026

Prompt-Induced Over-Generation as Denial-of-Service: A Black-Box Attack-Side Benchmark

Large Language Models (LLMs) can be driven into over-generation, emitting thousands of tokens before producing an end-of-sequence (EOS) token. This degrades answer quality, inflates latency and cost, and can be weaponized as a denial-of-service (DoS) attack. Recent work has begun to study DoS-style prompt attacks, but typically focuses on a single attack algorithm or assumes white-box access, without an attack-side benchmark that compares prompt-based attackers in a black-box, query-only regime with a known tokenizer. We introduce such a benchmark and study two prompt-only attackers. The first is an Evolutionary Over-Generation Prompt Search (EOGen) that searches the token space for prefixes that suppress EOS and induce long continuations. The second is a goal-conditioned reinforcement learning attacker (RL-GOAL) that trains a network to generate prefixes conditioned on a target length. To characterize behavior, we introduce Over-Generation Factor (OGF): the ratio of produced tokens to a model's context window, along with stall and latency summaries. EOGen discovers short-prefix attacks that raise Phi-3 to OGF = 1.39 +/- 1.14 (Success@>=2: 25.2%); RL-GOAL nearly doubles severity to OGF = 2.70 +/- 1.43 (Success@>=2: 64.3%) and drives budget-hit non-termination in 46% of trials.

cs.CR↗

arXiv2025

A Multi-Drone Multi-View Dataset and Deep Learning Framework for Pedestrian Detection and Tracking

Multi-drone surveillance systems offer enhanced coverage and robustness for pedestrian tracking, yet existing approaches struggle with dynamic camera positions and complex occlusions. This paper introduces MATRIX (Multi-Aerial TRacking In compleX environments), a comprehensive dataset featuring synchronized footage from eight drones with continuously changing positions, and a novel deep learning framework for multi-view detection and tracking. Unlike existing datasets that rely on static cameras or limited drone coverage, MATRIX provides a challenging scenario with 40 pedestrians and a significant architectural obstruction in an urban environment. Our framework addresses the unique challenges of dynamic drone-based surveillance through real-time camera calibration, feature-based image registration, and multi-view feature fusion in bird's-eye-view (BEV) representation. Experimental results demonstrate that while static camera methods maintain over 90\% detection and tracking precision and accuracy metrics in a simplified MATRIX environment without an obstruction, 10 pedestrians and a much smaller observational area, their performance significantly degrades in the complex environment. Our proposed approach maintains robust performance with $\sim$90\% detection and tracking accuracy, as well as successfully tracks $\sim$80\% of trajectories under challenging conditions. Transfer learning experiments reveal strong generalization capabilities, with the pretrained model achieving much higher detection and tracking accuracy performance compared to training the model from scratch. Additionally, systematic camera dropout experiments reveal graceful performance degradation, demonstrating practical robustness for real-world deployments where camera failures may occur. The MATRIX dataset and framework provide essential benchmarks for advancing dynamic multi-view surveillance systems.

cs.CV↗

arXiv2025

The Hidden Cost of Correlation: Rethinking Privacy Leakage in Local Differential Privacy

Local differential privacy (LDP) has emerged as a promising paradigm for privacy-preserving data collection in distributed systems, where users contribute multi-dimensional records with potentially correlated attributes. Recent work has highlighted that correlation-induced privacy leakage (CPL) plays a critical role in shaping the privacy-utility trade-off under LDP, especially when correlations exist among attributes. Nevertheless, it remains unclear to what extent the prevailing assumptions and proposed solutions are valid and how significant CPL is in real-world data. To address this gap, we first perform a comprehensive statistical analysis of five widely used LDP mechanisms -- GRR, RAPPOR, OUE, OLH and Exponential mechanism -- to assess CPL across four real-world datasets. We identify that many primary assumptions and metrics in current approaches fall short of accurately characterising these leakages. Moreover, current studies have been limited to a set of pure LDP (i.e., {δ= 0}) mechanisms. In response, we develop the first algorithmic framework to theoretically quantify CPL for any general approximated LDP (({\varepsilon},δ)-LDP) mechanism. We validate our theoretical results against empirical statistical results and provide a theoretical explanation for the observed statistical patterns. Finally, we propose two novel benchmarks to validate correlation analysis algorithms and evaluate the utility vs CPL of LDP mechanisms. Further, we demonstrate how these findings can be applied to achieve an efficient privacy-utility trade-off in real-world data governance.

cs.CR↗

arXiv2025

Personalizing Federated Learning for Hierarchical Edge Networks with Non-IID Data

Accommodating edge networks between IoT devices and the cloud server in Hierarchical Federated Learning (HFL) enhances communication efficiency without compromising data privacy. However, devices connected to the same edge often share geographic or contextual similarities, leading to varying edge-level data heterogeneity with different subsets of labels per edge, on top of device-level heterogeneity. This hierarchical non-Independent and Identically Distributed (non-IID) nature, which implies that each edge has its own optimization goal, has been overlooked in HFL research. Therefore, existing edge-accommodated HFL demonstrates inconsistent performance across edges in various hierarchical non-IID scenarios. To ensure robust performance with diverse edge-level non-IID data, we propose a Personalized Hierarchical Edge-enabled Federated Learning (PHE-FL), which personalizes each edge model to perform well on the unique class distributions specific to each edge. We evaluated PHE-FL across 4 scenarios with varying levels of edge-level non-IIDness, with extreme IoT device level non-IIDness. To accurately assess the effectiveness of our personalization approach, we deployed test sets on each edge server instead of the cloud server, and used both balanced and imbalanced test sets. Extensive experiments show that PHE-FL achieves up to 83 percent higher accuracy compared to existing federated learning approaches that incorporate edge networks, given the same number of training rounds. Moreover, PHE-FL exhibits improved stability, as evidenced by reduced accuracy fluctuations relative to the state-of-the-art FedAvg with two-level (edge and cloud) aggregation.

cs.LG↗

arXiv2024

Resource-Efficient Multiview Perception: Integrating Semantic Masking with Masked Autoencoders

Multiview systems have become a key technology in modern computer vision, offering advanced capabilities in scene understanding and analysis. However, these systems face critical challenges in bandwidth limitations and computational constraints, particularly for resource-limited camera nodes like drones. This paper presents a novel approach for communication-efficient distributed multiview detection and tracking using masked autoencoders (MAEs). We introduce a semantic-guided masking strategy that leverages pre-trained segmentation models and a tunable power function to prioritize informative image regions. This approach, combined with an MAE, reduces communication overhead while preserving essential visual information. We evaluate our method on both virtual and real-world multiview datasets, demonstrating comparable performance in terms of detection and tracking performance metrics compared to state-of-the-art techniques, even at high masking ratios. Our selective masking algorithm outperforms random masking, maintaining higher accuracy and precision as the masking ratio increases. Furthermore, our approach achieves a significant reduction in transmission data volume compared to baseline methods, thereby balancing multiview tracking performance with communication efficiency.

cs.CV↗

arXiv2024

SHFL: Secure Hierarchical Federated Learning Framework for Edge Networks

Federated Learning (FL) is a distributed machine learning paradigm designed for privacy-sensitive applications that run on resource-constrained devices with non-Identically and Independently Distributed (IID) data. Traditional FL frameworks adopt the client-server model with a single-level aggregation (AGR) process, where the server builds the global model by aggregating all trained local models received from client devices. However, this conventional approach encounters challenges, including susceptibility to model/data poisoning attacks. In recent years, advancements in the Internet of Things (IoT) and edge computing have enabled the development of hierarchical FL systems with a two-level AGR process running at edge and cloud servers. In this paper, we propose a Secure Hierarchical FL (SHFL) framework to address poisoning attacks in hierarchical edge networks. By aggregating trained models at the edge, SHFL employs two novel methods to address model/data poisoning attacks in the presence of client adversaries: 1) a client selection algorithm running at the edge for choosing IoT devices to participate in training, and 2) a model AGR method designed based on convex optimization theory to reduce the impact of edge models from networks with adversaries in the process of computing the global model (at the cloud level). The evaluation results reveal that compared to state-of-the-art methods, SHFL significantly increases the maximum accuracy achieved by the global model in the presence of client adversaries applying model/data poisoning attacks.

cs.LG↗

arXiv2024

ACCESS-FL: Agile Communication and Computation for Efficient Secure Aggregation in Stable Federated Learning Networks

Federated Learning (FL) is a promising distributed learning framework designed for privacy-aware applications. FL trains models on client devices without sharing the client's data and generates a global model on a server by aggregating model updates. Traditional FL approaches risk exposing sensitive client data when plain model updates are transmitted to the server, making them vulnerable to security threats such as model inversion attacks where the server can infer the client's original training data from monitoring the changes of the trained model in different rounds. Google's Secure Aggregation (SecAgg) protocol addresses this threat by employing a double-masking technique, secret sharing, and cryptography computations in honest-but-curious and adversarial scenarios with client dropouts. However, in scenarios without the presence of an active adversary, the computational and communication cost of SecAgg significantly increases by growing the number of clients. To address this issue, in this paper, we propose ACCESS-FL, a communication-and-computation-efficient secure aggregation method designed for honest-but-curious scenarios in stable FL networks with a limited rate of client dropout. ACCESS-FL reduces the computation/communication cost to a constant level (independent of the network size) by generating shared secrets between only two clients and eliminating the need for double masking, secret sharing, and cryptography computations. To evaluate the performance of ACCESS-FL, we conduct experiments using the MNIST, FMNIST, and CIFAR datasets to verify the performance of our proposed method. The evaluation results demonstrate that our proposed method significantly reduces computation and communication overhead compared to state-of-the-art methods, SecAgg and SecAgg+.

cs.CR↗

arXiv2024

DiffPMAE: Diffusion Masked Autoencoders for Point Cloud Reconstruction

Point cloud streaming is increasingly getting popular, evolving into the norm for interactive service delivery and the future Metaverse. However, the substantial volume of data associated with point clouds presents numerous challenges, particularly in terms of high bandwidth consumption and large storage capacity. Despite various solutions proposed thus far, with a focus on point cloud compression, upsampling, and completion, these reconstruction-related methods continue to fall short in delivering high fidelity point cloud output. As a solution, in DiffPMAE, we propose an effective point cloud reconstruction architecture. Inspired by self-supervised learning concepts, we combine Masked Auto-Encoding and Diffusion Model mechanism to remotely reconstruct point cloud data. By the nature of this reconstruction process, DiffPMAE can be extended to many related downstream tasks including point cloud compression, upsampling and completion. Leveraging ShapeNet-55 and ModelNet datasets with over 60000 objects, we validate the performance of DiffPMAE exceeding many state-of-the-art methods in-terms of auto-encoding and downstream tasks considered.

cs.CV↗