arXiv subjects

Peter Garraghan

Publications and source records attributed to Peter Garraghan.

15 recordsLinked to original sources

arXiv2026

Geometry of Divergence: Tracking Hidden-State Trajectories for Adaptive Multi-Turn Reasoning

LLM agents need to sustain goal-consistent reasoning across long multi-turn interactions under strict resource constraints. However, as the multi-turn context accumulates, it can destabilize the underlying LLM's internal representation of task-relevant information from earlier turns, blurring the boundary between constructive reasoning and representation drift. We formulate multi-turn reasoning as a hidden-state trajectory of the underlying LLM that is characterized via two complementary signals: temporal curvature that captures the directional consistency of turn-to-turn updates, and variance slope which measures the expansion or contraction of the exploration space. Across four tasks and three underlying LLMs, we observed that these geometric signals distinguish between correct and incorrect episodes prior to completion. We further decompose each episode into three-action chains formed from four actions (Read, Write, Respond, Transfer) and show that separability is action-dependent, with different signals distinguishing various chain patterns. Our experiments demonstrate that trajectory geometry can identify critical turns in the reasoning process, increasing task success rates on $\tau$-Bench from 24.1% to 39.6% while reducing token cost by 11.2%.

cs.AI↗

arXiv2026

kNNGuard: Turning LLM Hidden Activations into a Training-Free Configurable Guardrail

Large language models (LLMs) are increasingly deployed in domains requiring guardrails to detect unsafe, off-topic, or adversarial prompts. Existing guardrails predominantly rely on fine-tuning to build classifiers, which often suffer from low generalization and high inference latency. We present kNNGuard, a training-free guardrail that utilizes the activation space of an off-the-shelf LLM. Given a small bank of 50 safe and unsafe prompts, kNNGuard extracts hidden activations and performs multi-layer kNN fusing activation-space and embedding-space scores for classification. Across six domains spanning topical and security prompts, kNNGuard achieves competitive or superior F1 compared to fine-tuned state-of-the-art guardrails while running 2.7x faster than the best comparable guardrail, and 10x faster than a fine-tuned safety classifier without gradient updates or fine-tuning. Domain adaptation requires only updating the labeled bank, which can be constructed in under 10 seconds and several orders of magnitude faster than established guardrails. We also analyze the impact of system prompts, layer selection, and integration into production LLM pipelines as a configurable, low-latency guardrail.

cs.LG↗

arXiv2026

Behind the Refusal: Determining Guardrail Activation via Behavioral Monitoring

As Large Language Models (LLMs) and agentic systems become integrated into real-world applications, ensuring their safety and security is critical. Guardrail systems that detect and block malicious instructions sent to and from an LLM are an essential component of AI security. However, researchers conducting black-box adversarial emulation against production AI systems often struggle to determine whether a guardrail block or an LLM rejection has occurred. This distinction is important because the techniques used to bypass guardrails can differ substantially from those used to bypass LLM safety alignment, and has a material impact on attack technique selection and optimization. We propose the first black-box guardrail reconnaissance methodology, which detects the presence of a guardrail within a target AI system through behavioral monitoring of HTTP, lexical, and timing signals, assuming only black-box access and zero prior knowledge of the guardrail or AI system. Experiments demonstrate that our approach detects guardrail presence with 100% accuracy, with statistically significant behavioral separation between benign and malicious interactions (q < 0.001). Our approach further identifies the content categories a guardrail is designed to block, and distinguishes guardrail blocks from LLM rejection on unseen prompts with an average F1 score of 98%.

cs.CR↗

arXiv2025

Cost-Aware Model Orchestration for LLM-based Systems

As modern artificial intelligence (AI) systems become more advanced and capable, they can leverage a wide range of tools and models to perform complex tasks. The task of orchestrating these models is increasingly performed by Large Language Models (LLMs) that rely on qualitative descriptions of models for decision-making. However, the descriptions provided to existing LLM-based orchestrators frequently do not reflect true model capabilities and performance characteristics, leading to suboptimal model selection, reduced task accuracy, and increased cost. In this paper, we conduct an empirical analysis of LLM-based orchestration limitations and propose a cost-aware model selection method that accounts for performance-cost trade-offs by incorporating quantitative model performance characteristics within decision-making. Initial experimental results demonstrate that our proposed method increases accuracy by 0.90%-11.92% across various evaluated tasks, achieves up to a 54% energy efficiency improvement, and reduces orchestrator model selection latency from 4.51 s to 7.2 ms.

cs.AI↗

arXiv2025

Bypassing LLM Guardrails: An Empirical Analysis of Evasion Attacks against Prompt Injection and Jailbreak Detection Systems

Large Language Models (LLMs) guardrail systems are designed to protect against prompt injection and jailbreak attacks. However, they remain vulnerable to evasion techniques. We demonstrate two approaches for bypassing LLM prompt injection and jailbreak detection systems via traditional character injection methods and algorithmic Adversarial Machine Learning (AML) evasion techniques. Through testing against six prominent protection systems, including Microsoft's Azure Prompt Shield and Meta's Prompt Guard, we show that both methods can be used to evade detection while maintaining adversarial utility achieving in some instances up to 100% evasion success. Furthermore, we demonstrate that adversaries can enhance Attack Success Rates (ASR) against black-box targets by leveraging word importance ranking computed by offline white-box models. Our findings reveal vulnerabilities within current LLM protection mechanisms and highlight the need for more robust guardrail systems.

cs.CR↗

arXiv2025

Green Prompting: Characterizing Prompt-driven Energy Costs of LLM Inference

Large Language Models (LLMs) have become widely used across various domains spanning search engines, code generation, and text creation. However, a major concern associated with their adoption is the high cost of inference, impacting both their sustainability and financial feasibility. In this study, we empirically study how different prompt and response characteristics directly impact LLM inference energy cost. We conduct experiments leveraging three open-source transformer-based LLMs across three task types$-$question answering, sentiment analysis, and text generation. For each inference, we analyzed prompt and response characteristics (length, semantic meaning, time taken, energy consumption). Our results demonstrate that even when presented with identical tasks, models generate responses with varying characteristics and subsequently exhibit distinct energy consumption patterns. We found that prompt length is less significant than the semantic meaning of the task itself. In addition, we identified specific keywords associated with higher or lower energy usage that vary between associated tasks. These findings highlight the importance of prompt design in optimizing inference efficiency. We conclude that the semantic meaning of prompts and certain task-related keywords significantly impact inference costs, leading the way for deeper exploration towards creating energy-adaptive LLMs.

cs.CL↗

arXiv2025

EDoRA: Efficient Weight-Decomposed Low-Rank Adaptation via Singular Value Decomposition

Parameter-efficient fine-tuning methods, such as LoRA, reduces the number of trainable parameters. However, they often suffer from scalability issues and differences between their learning pattern and full fine-tuning. To overcome these limitations, we propose Efficient Weight-Decomposed Low-Rank Adaptation (EDoRA): a novel PEFT method that decomposes pre-trained weights into magnitude and directional components. By freezing low-rank matrices, initializing them by singular value decomposition, and introducing a small trainable matrix between them, EDoRA achieves substantial reduction in trainable parameters while maintaining learning capacity. Experimental results on the GLUE benchmark demonstrate that EDoRA achieves competitive or superior performance compared to state-of-the-art methods, such as LoRA and DoRA, with up to 30x fewer trainable parameters. This makes EDoRA a highly efficient solution for adapting LLMs to diverse tasks under memory-constrained settings. Code is available at https://github.com/Hamid-Nasiri/EDoRA .

cs.LG↗

arXiv2024

ICT Sector Greenhouse Gas Emissions -- Issues and Trends

As Information and Communication Technology (ICT) use has become more prevalent, there has been a growing concern in how its associated greenhouse gas emissions will impact the climate. Estimating such ICT emissions is a difficult undertaking due to its complexity, its rapidly changing nature, and the lack of accurate and up-to-date data on individual stakeholder emissions. In this paper we provide a framework for estimating ICT's carbon footprint and identify some of the issues that impede the task. We attempt to gain greater insight into the factors affecting the ICT sector by drawing on a number of interviews with industry experts. We conclude that more accurate emissions estimates will only be possible with a more more detailed, industry informed, understanding of the whole ICT landscape and much more transparent reporting of energy usage and emissions data by ICT stakeholders.

cs.CY↗

arXiv2023

Compilation as a Defense: Enhancing DL Model Attack Robustness via Tensor Optimization

Adversarial Machine Learning (AML) is a rapidly growing field of security research, with an often overlooked area being model attacks through side-channels. Previous works show such attacks to be serious threats, though little progress has been made on efficient remediation strategies that avoid costly model re-engineering. This work demonstrates a new defense against AML side-channel attacks using model compilation techniques, namely tensor optimization. We show relative model attack effectiveness decreases of up to 43% using tensor optimization, discuss the implications, and direction of future work.

cs.LG↗

arXiv2023

Model Leeching: An Extraction Attack Targeting LLMs

Model Leeching is a novel extraction attack targeting Large Language Models (LLMs), capable of distilling task-specific knowledge from a target LLM into a reduced parameter model. We demonstrate the effectiveness of our attack by extracting task capability from ChatGPT-3.5-Turbo, achieving 73% Exact Match (EM) similarity, and SQuAD EM and F1 accuracy scores of 75% and 87%, respectively for only $50 in API cost. We further demonstrate the feasibility of adversarial attack transferability from an extracted model extracted via Model Leeching to perform ML attack staging against a target LLM, resulting in an 11% increase to attack success rate when applied to ChatGPT-3.5-Turbo.

cs.LG↗

arXiv2022

PINCH: An Adversarial Extraction Attack Framework for Deep Learning Models

Adversarial extraction attacks constitute an insidious threat against Deep Learning (DL) models in-which an adversary aims to steal the architecture, parameters, and hyper-parameters of a targeted DL model. Existing extraction attack literature have observed varying levels of attack success for different DL models and datasets, yet the underlying cause(s) behind their susceptibility often remain unclear, and would help facilitate creating secure DL systems. In this paper we present PINCH: an efficient and automated extraction attack framework capable of designing, deploying, and analyzing extraction attack scenarios across heterogeneous hardware platforms. Using PINCH, we perform extensive experimental evaluation of extraction attacks against 21 model architectures to explore new extraction attack scenarios and further attack staging. Our findings show (1) key extraction characteristics whereby particular model configurations exhibit strong resilience against specific attacks, (2) even partial extraction success enables further staging for other adversarial attacks, and (3) equivalent stolen models uncover differences in expressive power, yet exhibit similar captured knowledge.

cs.CR↗

arXiv2021

START: Straggler Prediction and Mitigation for Cloud Computing Environments using Encoder LSTM Networks

Modern large-scale computing systems distribute jobs into multiple smaller tasks which execute in parallel to accelerate job completion rates and reduce energy consumption. However, a common performance problem in such systems is dealing with straggler tasks that are slow running instances that increase the overall response time. Such tasks can significantly impact the system's Quality of Service (QoS) and the Service Level Agreements (SLA). To combat this issue, there is a need for automatic straggler detection and mitigation mechanisms that execute jobs without violating the SLA. Prior work typically builds reactive models that focus first on detection and then mitigation of straggler tasks, which leads to delays. Other works use prediction based proactive mechanisms, but ignore heterogeneous host or volatile task characteristics. In this paper, we propose a Straggler Prediction and Mitigation Technique (START) that is able to predict which tasks might be stragglers and dynamically adapt scheduling to achieve lower response times. Our technique analyzes all tasks and hosts based on compute and network resource consumption using an Encoder Long-Short-Term-Memory (LSTM) network. The output of this network is then used to predict and mitigate expected straggler tasks. This reduces the SLA violation rate and execution time without compromising QoS. Specifically, we use the CloudSim toolkit to simulate START in a cloud environment and compare it with state-of-the-art techniques (IGRU-SD, SGC, Dolly, GRASS, NearestFit and Wrangler) in terms of QoS parameters such as energy consumption, execution time, resource contention, CPU utilization and SLA violation rate. Experiments show that START reduces execution time, resource contention, energy and SLA violations by 13%, 11%, 16% and 19%, respectively, compared to the state-of-the-art approaches.

cs.DC↗

arXiv2021

HUNTER: AI based Holistic Resource Management for Sustainable Cloud Computing

The worldwide adoption of cloud data centers (CDCs) has given rise to the ubiquitous demand for hosting application services on the cloud. Further, contemporary data-intensive industries have seen a sharp upsurge in the resource requirements of modern applications. This has led to the provisioning of an increased number of cloud servers, giving rise to higher energy consumption and, consequently, sustainability concerns. Traditional heuristics and reinforcement learning based algorithms for energy-efficient cloud resource management address the scalability and adaptability related challenges to a limited extent. Existing work often fails to capture dependencies across thermal characteristics of hosts, resource consumption of tasks and the corresponding scheduling decisions. This leads to poor scalability and an increase in the compute resource requirements, particularly in environments with non-stationary resource demands. To address these limitations, we propose an artificial intelligence (AI) based holistic resource management technique for sustainable cloud computing called HUNTER. The proposed model formulates the goal of optimizing energy efficiency in data centers as a multi-objective scheduling problem, considering three important models: energy, thermal and cooling. HUNTER utilizes a Gated Graph Convolution Network as a surrogate model for approximating the Quality of Service (QoS) for a system state and generating optimal scheduling decisions. Experiments on simulated and physical cloud environments using the CloudSim toolkit and the COSCO framework show that HUNTER outperforms state-of-the-art baselines in terms of energy consumption, SLA violation, scheduling time, cost and temperature by up to 12, 35, 43, 54 and 3 percent respectively.

cs.DC↗

arXiv2020

ThermoSim: Deep Learning based Framework for Modeling and Simulation of Thermal-aware Resource Management for Cloud Computing Environments

Current cloud computing frameworks host millions of physical servers that utilize cloud computing resources in the form of different virtual machines (VM). Cloud Data Center (CDC) infrastructures require significant amounts of energy to deliver large scale computational services. Computing nodes generate large volumes of heat, requiring cooling units in turn to eliminate the effect of this heat. Thus, the overall energy consumption of the CDC increases tremendously for servers as well as for cooling units. However, current workload allocation policies do not take into account the effect on temperature and it is challenging to simulate the thermal behavior of CDCs. There is a need for a thermal-aware framework to simulate and model the behavior of nodes and measure the important performance parameters which can be affected by its temperature. In this paper, we propose a lightweight framework, ThermoSim, for modeling and simulation of thermal-aware resource management for cloud computing environments. This work presents a Recurrent Neural Network based deep learning temperature predictor for CDCs which is utilized by ThermoSim for lightweight resource management in constrained cloud environments. ThermoSim extends the CloudSim toolkit helping to analyze the performance of various key parameters such as energy consumption, SLA violation rate, number of VM migrations and temperature during the management of cloud resources for execution of workloads. Further, different energy-aware and thermal-aware resource management techniques are tested using the proposed ThermoSim framework in order to validate it against the existing framework. The experimental results demonstrate the proposed framework is capable of modeling and simulating the thermal behavior of a CDC and the ThermoSim framework is better than Thas in terms of energy consumption, cost, time, memory usage & prediction accuracy.

cs.DC↗

arXiv2019

Transformative effects of IoT, Blockchain and Artificial Intelligence on cloud computing: Evolution, vision, trends and open challenges

Cloud computing plays a critical role in modern society and enables a range of applications from infrastructure to social media. Such system must cope with varying load and evolving usage reflecting societies interaction and dependency on automated computing systems whilst satisfying Quality of Service (QoS) guarantees. Enabling these systems are a cohort of conceptual technologies, synthesized to meet demand of evolving computing applications. In order to understand current and future challenges of such system, there is a need to identify key technologies enabling future applications. In this study, we aim to explore how three emerging paradigms (Blockchain, IoT and Artificial Intelligence) will influence future cloud computing systems. Further, we identify several technologies driving these paradigms and invite international experts to discuss the current status and future directions of cloud computing. Finally, we proposed a conceptual model for cloud futurology to explore the influence of emerging paradigms and technologies on evolution of cloud computing.

cs.DC↗