Prompt Variability Effects On LLM Code Generation

Code generation is one of the most active areas of application of Large Language Models (LLMs). While LLMs lower barriers to writing code and accelerate development process, the overall quality of generated programs depends on the quality of given prompts. Specifically, functionality and quality of generated code can be sensitive to user's background and familiarity with software development. It is therefore important to quantify LLM's sensitivity to variations in the input. To this end we propose a synthetic evaluation pipeline for code generation with LLMs, as well as a systematic persona-based evaluation approach to expose qualitative differences of LLM responses dependent on prospective user background. Both proposed methods are completely independent from specific programming tasks and LLMs, and thus are widely applicable. We provide experimental evidence illustrating utility of our methods and share our code for the benefit of the community.

LLM Performance for Code Generation on Noisy Tasks

This paper investigates the ability of large language models (LLMs) to recognise and solve tasks which have been obfuscated beyond recognition. Focusing on competitive programming and benchmark tasks (LeetCode and MATH), we compare performance across multiple models and obfuscation methods, such as noise and redaction. We demonstrate that all evaluated LLMs can solve tasks obfuscated to a level where the text would be unintelligible to human readers, and does not contain key pieces of instruction or context. We introduce the concept of eager pattern matching to describe this behaviour, which is not observed in tasks published after the models' knowledge cutoff date, indicating strong memorisation or overfitting to training data, rather than legitimate reasoning about the presented problem. We report empirical evidence of distinct performance decay patterns between contaminated and unseen datasets. We discuss the implications for benchmarking and evaluations of model behaviour, arguing for caution when designing experiments using standard datasets. We also propose measuring the decay of performance under obfuscation as a possible strategy for detecting dataset contamination and highlighting potential safety risks and interpretability issues for automated software systems.

Self-sustaining Software Systems (S4): Towards Improved Interpretability and Adaptation

Software systems impact society at different levels as they pervasively solve real-world problems. Modern software systems are often so sophisticated that their complexity exceeds the limits of human comprehension. These systems must respond to changing goals, dynamic data, unexpected failures, and security threats, among other variable factors in real-world environments. Systems' complexity challenges their interpretability and requires autonomous responses to dynamic changes. Two main research areas explore autonomous systems' responses: evolutionary computing and autonomic computing. Evolutionary computing focuses on software improvement based on iterative modifications to the source code. Autonomic computing focuses on optimising systems' performance by changing their structure, behaviour, or environment variables. Approaches from both areas rely on feedback loops that accumulate knowledge from the system interactions to inform autonomous decision-making. However, this knowledge is often limited, constraining the systems' interpretability and adaptability. This paper proposes a new concept for interpretable and adaptable software systems: self-sustaining software systems (S4). S4 builds knowledge loops between all available knowledge sources that define modern software systems to improve their interpretability and adaptability. This paper introduces and discusses the S4 concept.

Automated discovery of trade-off between utility, privacy and fairness in machine learning models

Machine learning models are deployed as a central component in decision making and policy operations with direct impact on individuals' lives. In order to act ethically and comply with government regulations, these models need to make fair decisions and protect the users' privacy. However, such requirements can come with decrease in models' performance compared to their potentially biased, privacy-leaking counterparts. Thus the trade-off between fairness, privacy and performance of ML models emerges, and practitioners need a way of quantifying this trade-off to enable deployment decisions. In this work we interpret this trade-off as a multi-objective optimization problem, and propose PFairDP, a pipeline that uses Bayesian optimization for discovery of Pareto-optimal points between fairness, privacy and utility of ML models. We show how PFairDP can be used to replicate known results that were achieved through manual constraint setting process. We further demonstrate effectiveness of PFairDP with experiments on multiple models and datasets.

Causal fault localisation in dataflow systems

Dataflow computing was shown to bring significant benefits to multiple niches of systems engineering and has the potential to become a general-purpose paradigm of choice for data-driven application development. One of the characteristic features of dataflow computing is the natural access to the dataflow graph of the entire system. Recently it has been observed that these dataflow graphs can be treated as complete graphical causal models, opening opportunities to apply causal inference techniques to dataflow systems. In this demonstration paper we aim to provide the first practical validation of this idea with a particular focus on causal fault localisation. We provide multiple demonstrations of how causal inference can be used to detect software bugs and data shifts in multiple scenarios with three modern dataflow engines.

Dataflow graphs as complete causal graphs

Component-based development is one of the core principles behind modern software engineering practices. Understanding of causal relationships between components of a software system can yield significant benefits to developers. Yet modern software design approaches make it difficult to track and discover such relationships at system scale, which leads to growing intellectual debt. In this paper we consider an alternative approach to software design, flow-based programming (FBP), and draw the attention of the community to the connection between dataflow graphs produced by FBP and structural causal models. With expository examples we show how this connection can be leveraged to improve day-to-day tasks in software projects, including fault localisation, business analysis and experimentation.

Trieste: Efficiently Exploring The Depths of Black-box Functions with TensorFlow

We present Trieste, an open-source Python package for Bayesian optimization and active learning benefiting from the scalability and efficiency of TensorFlow. Our library enables the plug-and-play of popular TensorFlow-based models within sequential decision-making loops, e.g. Gaussian processes from GPflow or GPflux, or neural networks from Keras. This modular mindset is central to the package and extends to our acquisition functions and the internal dynamics of the decision-making loop, both of which can be tailored and extended by researchers or engineers when tackling custom use cases. Trieste is a research-friendly and production-ready toolkit backed by a comprehensive test suite, extensive documentation, and available at https://github.com/secondmind-labs/trieste.

Real-world Machine Learning Systems: A survey from a Data-Oriented Architecture Perspective

With the upsurge of interest in artificial intelligence machine learning (ML) algorithms, originally developed in academic environments, are now being deployed as parts of real-life systems that deal with large amounts of heterogeneous, dynamic, and high-dimensional data. Deployment of ML methods in real life is prone to challenges across the whole system life-cycle from data management to systems deployment, monitoring, and maintenance. Data-Oriented Architecture (DOA) is an emerging software engineering paradigm that has the potential to mitigate these challenges by proposing a set of principles to create data-driven, loosely coupled, decentralised, and open systems. However DOA as a concept is not widespread yet, and there is no common understanding of how it can be realised in practice. This review addresses that problem by contextualising the principles that underpin the DOA paradigm through the ML system challenges. We explore the extent to which current architectures of ML-based real-world systems have implemented the DOA principles. We also formulate open research challenges and directions for further development of the DOA paradigm.

Desiderata for next generation of ML model serving

Inference is a significant part of ML software infrastructure. Despite the variety of inference frameworks available, the field as a whole can be considered in its early days. This paper puts forth a range of important qualities that next generation of inference platforms should be aiming for. We present our rationale for the importance of each quality, and discuss ways to achieve it in practice. An overarching design pattern is data-centricity, which enables smarter monitoring in ML system operation.

A penalisation method for batch multi-objective Bayesian optimisation with application in heat exchanger design

We present HIghly Parallelisable Pareto Optimisation (HIPPO) -- a batch acquisition function that enables multi-objective Bayesian optimisation methods to efficiently exploit parallel processing resources. Multi-Objective Bayesian Optimisation (MOBO) is a very efficient tool for tackling expensive black-box problems. However, most MOBO algorithms are designed as purely sequential strategies, and existing batch approaches are prohibitively expensive for all but the smallest of batch sizes. We show that by encouraging batch diversity through penalising evaluations with similar predicted objective values, HIPPO is able to cheaply build large batches of informative points. Our extensive experimental validation demonstrates that HIPPO is at least as efficient as existing alternatives whilst incurring an order of magnitude lower computational overhead and scaling easily to batch sizes considerably higher than currently supported in the literature. Additionally, we demonstrate the application of HIPPO to a challenging heat exchanger design problem, stressing the real-world utility of our highly parallelisable approach to MOBO.