Leslie Kanthan

Paul Brookes, Vardan Voskanyan, Rafail Giavrimis et al.

Agentic AI systems built on large language models (LLMs) offer significant potential for automating complex workflows, from software development to customer support. However, LLM agents often underperform due to suboptimal configurations; poorly tuned prompts, tool descriptions, and parameters that typically require weeks of manual refinement. Existing optimization methods either are too complex for general use or treat components in isolation, missing critical interdependencies. We present ARTEMIS, a no-code evolutionary optimization platform that jointly optimizes agent configurations through semantically-aware genetic operators. Given only a benchmark script and natural language goals, ARTEMIS automatically discovers configurable components, extracts performance signals from execution logs, and evolves configurations without requiring architectural modifications. We evaluate ARTEMIS on four representative agent systems: the \emph{ALE Agent} for competitive programming on AtCoder Heuristic Contest, achieving a \textbf{$13.6\%$ improvement} in acceptance rate; the \emph{Mini-SWE Agent} for code optimization on SWE-Perf, with a statistically significant \textbf{10.1\% performance gain}; and the \emph{CrewAI Agent} for cost and mathematical reasoning on Math Odyssey, achieving a statistically significant \textbf{$36.9\%$ reduction} in the number of tokens required for evaluation. We also evaluate the \emph{MathTales-Teacher Agent} powered by a smaller open-source model (Qwen2.5-7B) on GSM8K primary-level mathematics problems, achieving a \textbf{22\% accuracy improvement} and demonstrating that ARTEMIS can optimize agents based on both commercial and local models.

Lingbo Li, Leslie Kanthan, Michail Basios et al.

Machine learning model development and optimisation can be a rather cumbersome and resource-intensive process. Custom models are often more difficult to build and deploy, and they require infrastructure and expertise which are often costly to acquire and maintain. Machine learning product development lifecycle must take into account the need to navigate the difficulties of developing and deploying machine learning models. evoML is an AI-powered tool that provides automated functionalities in machine learning model development, optimisation, and model code optimisation. Core functionalities of evoML include data cleaning, exploratory analysis, feature analysis and generation, model optimisation, model evaluation, model code optimisation, and model deployment. Additionally, a key feature of evoML is that it embeds code and model optimisation into the model development process, and includes multi-objective optimisation capabilities.

Mari Ashiga, Vardan Voskanyan, Fateme Dinmohammadi et al.

Recent advancements in Large Language Models (LLMs) for code optimization have enabled industrial platforms to automate software performance engineering at unprecedented scale and speed. Yet, organizations in regulated industries face strict constraints on which LLMs they can use - many cannot utilize commercial models due to data privacy regulations and compliance requirements, creating a significant challenge for achieving high-quality code optimization while maintaining cost-effectiveness. We address this by implementing a Mixture-of-Agents (MoA) approach that directly synthesizes code from multiple specialized LLMs, comparing it against TurinTech AI's vanilla Genetic Algorithm (GA)-based ensemble system and individual LLM optimizers using real-world industrial codebases. Our key contributions include: (1) First MoA application to industrial code optimization using real-world codebases; (2) Empirical evidence that MoA excels with open-source models, achieving 14.3% to 22.2% cost savings and 28.6% to 32.2% faster optimization times for regulated environments; (3) Deployment guidelines demonstrating GA's advantage with commercial models while both ensembles outperform individual LLMs; and (4) Real-world validation across 50 code snippets and seven LLM combinations, generating over 8,700 variants, addresses gaps in industrial LLM ensemble evaluation. This provides actionable guidance for organizations balancing regulatory compliance with optimization performance in production environments.

Jingzhi Gong, Rafail Giavrimis, Paul Brookes et al.

There is a growing interest in leveraging multiple large language models (LLMs) for automated code optimization. However, industrial platforms deploying multiple LLMs face a critical challenge: prompts optimized for one LLM often fail with others, requiring expensive model-specific prompt engineering. This cross-model prompt engineering bottleneck severely limits the practical deployment of multi-LLM systems in production environments. We introduce Meta-Prompted Code Optimization (MPCO), a framework that automatically generates high-quality, task-specific prompts across diverse LLMs while maintaining industrial efficiency requirements. MPCO leverages metaprompting to dynamically synthesize context-aware optimization prompts by integrating project metadata, task requirements, and LLM-specific contexts. It is an essential part of the ARTEMIS code optimization platform for automated validation and scaling. Our comprehensive evaluation on five real-world codebases with 366 hours of runtime benchmarking demonstrates MPCO's effectiveness: it achieves overall performance improvements up to 19.06% with the best statistical rank across all systems compared to baseline methods. Analysis shows that 96% of the top-performing optimizations stem from meaningful edits. Through systematic ablation studies and meta-prompter sensitivity analysis, we identify that comprehensive context integration is essential for effective meta-prompting and that major LLMs can serve effectively as meta-prompters, providing actionable insights for industrial practitioners.

Kenneth T. Co, Luis Muñoz-González, Leslie Kanthan et al.

Universal Adversarial Perturbations (UAPs) are a prominent class of adversarial examples that exploit the systemic vulnerabilities and enable physically realizable and robust attacks against Deep Neural Networks (DNNs). UAPs generalize across many different inputs; this leads to realistic and effective attacks that can be applied at scale. In this paper we propose HyperNeuron, an efficient and scalable algorithm that allows for the real-time detection of UAPs by identifying suspicious neuron hyper-activations. Our results show the effectiveness of HyperNeuron on multiple tasks (image classification, object detection), against a wide variety of universal attacks, and in realistic scenarios, like perceptual ad-blocking and adversarial patches. HyperNeuron is able to simultaneously detect both adversarial mask and patch UAPs with comparable or better performance than existing UAP defenses whilst introducing a significantly reduced latency of only 0.86 milliseconds per image. This suggests that many realistic and practical universal attacks can be reliably mitigated in real-time, which shows promise for the robust deployment of machine learning systems.

Fan Fang, Carmine Ventre, Lingbo Li et al.

Feature importance aims at measuring how crucial each input feature is for model prediction. It is widely used in feature engineering, model selection and explainable artificial intelligence (XAI). In this paper, we propose a new tree-model explanation approach for model selection. Our novel concept leverages the Coefficient of Variation of a feature weight (measured in terms of the contribution of the feature to the prediction) to capture the dispersion of importance over samples. Extensive experimental results show that our novel feature explanation performs better than general cross validation method in model selection both in terms of time efficiency and accuracy performance.

Yuxi Huan, Fan Wu, Michail Basios et al.

Hyperparameter optimisation is a crucial process in searching the optimal machine learning model. The efficiency of finding the optimal hyperparameter settings has been a big concern in recent researches since the optimisation process could be time-consuming, especially when the objective functions are highly expensive to evaluate. In this paper, we introduce an intelligent evolutionary optimisation algorithm which applies machine learning technique to the traditional evolutionary algorithm to accelerate the overall optimisation process of tuning machine learning models in classification problems. We demonstrate our Intelligent Evolutionary Optimisation (IEO)in a series of controlled experiments, comparing with traditional evolutionary optimisation in hyperparameter tuning. The empirical study shows that our approach accelerates the optimisation speed by 30.40% on average and up to 77.06% in the best scenarios.

Fan Fang, Waichung Chung, Carmine Ventre et al.

The cryptocurrency market is amongst the fastest-growing of all the financial markets in the world. Unlike traditional markets, such as equities, foreign exchange and commodities, cryptocurrency market is considered to have larger volatility and illiquidity. This paper is inspired by the recent success of using deep learning for stock market prediction. In this work, we analyze and present the characteristics of the cryptocurrency market in a high-frequency setting. In particular, we applied a deep learning approach to predict the direction of the mid-price changes on the upcoming tick. We monitored live tick-level data from $8$ cryptocurrency pairs and applied both statistical and machine learning techniques to provide a live prediction. We reveal that promising results are possible for cryptocurrencies, and in particular, we achieve a consistent $78\%$ accuracy on the prediction of the mid-price movement on live exchange rate of Bitcoins vs US dollars.

Kenneth T. Co, Luis Muñoz-González, Leslie Kanthan et al.

Increasing shape-bias in deep neural networks has been shown to improve robustness to common corruptions and noise. In this paper we analyze the adversarial robustness of texture and shape-biased models to Universal Adversarial Perturbations (UAPs). We use UAPs to evaluate the robustness of DNN models with varying degrees of shape-based training. We find that shape-biased models do not markedly improve adversarial robustness, and we show that ensembles of texture and shape-biased models can improve universal adversarial robustness while maintaining strong performance.

Michail Basios, Lingbo Li, Fan Wu et al.

Data structure selection and tuning is laborious but can vastly improve an application's performance and memory footprint. Some data structures share a common interface and enjoy multiple implementations. We call them Darwinian Data Structures (DDS), since we can subject their implementations to survival of the fittest. We introduce ARTEMIS a multi-objective, cloud-based search-based optimisation framework that automatically finds optimal, tuned DDS modulo a test suite, then changes an application to use that DDS. ARTEMIS achieves substantial performance improvements for \emph{every} project in $5$ Java projects from DaCapo benchmark, $8$ popular projects and $30$ uniformly sampled projects from GitHub. For execution time, CPU usage, and memory consumption, ARTEMIS finds at least one solution that improves \emph{all} measures for $86\%$ ($37/43$) of the projects. The median improvement across the best solutions is $4.8\%$, $10.1\%$, $5.1\%$ for runtime, memory and CPU usage. These aggregate results understate ARTEMIS's potential impact. Some of the benchmarks it improves are libraries or utility functions. Two examples are gson, a ubiquitous Java serialization framework, and xalan, Apache's XML transformation tool. ARTEMIS improves gson by $16.5$\%, $1\%$ and $2.2\%$ for memory, runtime, and CPU; ARTEMIS improves xalan's memory consumption by $23.5$\%. \emph{Every} client of these projects will benefit from these performance improvements.