arXiv subjects

Cristina David

Publications and source records attributed to Cristina David.

At least 19 recordsLinked to original sources

arXiv2026

Using Semantic Distance to Estimate Uncertainty in LLM-Based Code Generation

LLMs show strong performance in code generation, but their outputs lack correctness guarantees. Sample-based uncertainty estimators address this by generating multiple candidate programs and measuring their disagreement. However, existing estimators make different design choices about how behaviours are identified, aggregated, referenced and compared, making them difficult to assess. We therefore first introduce a taxonomy that disentangles these choices and reveals a missing design point: semantic distance-aware uncertainty estimation, which measures not only whether sampled programs disagree, but how severely their execution behaviours differ. Across LiveCodeBench, MBPP, HumanEval-X and BigCodeBench, spanning Python, Java and C++, our metrics provide strong proxies for correctness, and consistently outperform state-of-the-art sample-based baselines across both closed-source models (GPT-3.5-Turbo, GPT-4o-mini, Gemini-2.5-Flash-Lite, Claude Opus 4.5) and an open-source model (DeepSeek-Coder-V2). The method is practical: it requires neither model internals nor LLM-as-judge calls, remains robust across models, languages, sampling temperatures and fuzzing settings, and reduces runtime by approximately 48-79% relative to existing baselines.

cs.SE↗

arXiv2026

Literate Execution

\emph{Literate programming}, introduced by Knurth, interleaves code and prose so that a program can be read as both executable and explanatory text. We propose \emph{literate execution}, which inverts this relationship: rather than embedding code within a static narrative, we treat documentation -- and other expository elements such as visualisations -- as first-class artefacts that can be computed alongside a running program and then integrated into a view of its execution. We explore this idea through Fluid, a programming language with a provenance-tracking runtime that records fine-grained dependencies between inputs and outputs. These provenance relationships can be surfaced as interactions that allow readers to explore how intermediate values contribute to a result. By integrating visualisation, provenance, and exposition, literate execution aims to make programs more explorable and self-explanatory, and explorable explanations easier to program.

cs.PL↗

arXiv2026

Transformers for Program Termination

Determining whether a program terminates is a core challenge in program analysis with direct implications for correctness, verification, and security. We investigate whether transformer architectures can recognise termination patterns directly from source code and how their strengths can be amplified through ensembles. To overcome the extreme scarcity of non-terminating examples, we design an ensemble framework of compact transformer encoders, systematically trained with a suite of imbalance-aware loss functions and class-aware sampling techniques. By combining models trained with distinct loss functions, our ensembles achieve substantially stronger performance than any single transformer, outperforming both powerful off-the-shelf LLMs and graph-based methods. Finally, we introduce an attribution pipeline that produces syntax-aware explanations for the termination estimation.

cs.PL↗

arXiv2026

Validated Code Translation for Projects with External Libraries

Large Language Models (LLMs) have shown promise for program translation, particularly for migrating systems code to memory-safe languages such as Rust. However, existing approaches struggle when source programs depend on external libraries: LLMs frequently hallucinate non-existent target APIs and fail to generate call-enabling imports; moreover, validating semantic equivalence is challenging when the code manipulates opaque, library-defined types. We present a translation and validation framework for translating Go projects with external dependencies to Rust. Our approach combines (i) a retrieval mechanism that maps Go library APIs to Rust APIs, and (ii) a cross-language validation pipeline that establishes language interoperability in the presence of opaque library types by synthesising adapters exclusively from public library APIs, prior to validating I/O equivalence. We evaluate our system on six real-world Go repositories with non-trivial external dependencies. Our approach significantly increases both the compilation and equivalence success rate (up to 100% in the most dependency-heavy case; approx. 2x on average) by enabling validated translation that manipulate opaque, library-defined types.

cs.SE↗

arXiv2025

TreeCoder: Systematic Exploration and Optimisation of Decoding and Constraints for LLM Code Generation

Large language models (LLMs) have shown remarkable ability to generate code, yet their outputs often violate syntactic or semantic constraints when guided only through natural language prompts. We introduce TreeCoder, the most general and flexible framework to date for exploring decoding strategies, constraints, and hyperparameters in LLMs, and use it in code generation to enforce correctness and structure during decoding rather than relying on prompt engineering. TreeCoder represents decoding as a tree search over candidate programs, where both decoding strategies and constraint functions - such as style, syntax, execution - are treated as first-class, optimisable components. This design enables systematic exploration and automatic tuning of decoding configurations using standard optimisation techniques. Experiments on the MBPP (Python) and SQL-Spider benchmarks show that TreeCoder consistently improves accuracy across open-source models such as CodeLlama, Mistral and DeepSeek, often outperforming their unconstrained baselines by considerable margins.

cs.LG↗

arXiv2025

AI-Assisted Authoring for Transparent, Data-Driven Documents

We introduce _transparent documents_, interactive web-based scholarly articles which allow readers to explore the relationship to the underlying data by hovering over fragments of text, and present an LLM-based tool for authoring transparent documents, building on recent developments in data provenance for general-purpose programming languages. As a target platform, our implementation uses Fluid, an open source programming language with a provenance-tracking runtime. Our agent-based tool supports a human author during the creation of transparent documents, identifying fragments of text which can be computed from data, such as numerical values selected from records or computed by aggregations like sum and mean, comparatives and superlatives like _better than_ and _largest_, trend-adjectives like _growing_, and similar quantitative or semi-quantitative phrases, and then attempts to synthesise a suitable Fluid query over the data which generates the target string. The resulting expression is inserted into the article's web page, turning the static text fragment into an interactable data-driven element able to reveal the data that underwrites the natural language claim. We evaluate our approach on a subset of SciGen, an open source dataset consisting of tables from scientific articles and their corresponding descriptions, which we extend with hand-generated counterfactual test cases to evaluate how well machine-generated expressions generalise. Our results show that gpt4o is often able to synthesise compound expressions extensionally compatible with our gold solutions.

cs.HC↗

arXiv2025

Assessing Correctness in LLM-Based Code Generation via Uncertainty Estimation

In this work, we explore uncertainty estimation as a proxy for correctness in LLM-generated code. To this end, we adapt two state-of-the-art techniques from natural language generation -- one based on entropy and another on mutual information -- to the domain of code generation. Given the distinct semantic properties of code, we introduce modifications, including a semantic equivalence check based on symbolic execution. Our findings indicate a strong correlation between the uncertainty computed through these techniques and correctness, highlighting the potential of uncertainty estimation for quality assessment. Additionally, we propose a simplified version of the entropy-based method that assumes a uniform distribution over the LLM's responses, demonstrating comparable effectiveness. Using these techniques, we develop an abstention policy that prevents the model from making predictions when uncertainty is high, reducing incorrect outputs to near zero. Our evaluation on the LiveCodeBench shows that our approach significantly outperforms a baseline relying solely on LLM-reported log-probabilities.

cs.SE↗

arXiv2024

Scalable, Validated Code Translation of Entire Projects using Large Language Models

Large language models (LLMs) show promise in code translation due to their ability to generate idiomatic code. However, a significant limitation when using LLMs for code translation is scalability: existing works have shown a drop in translation success rates for code exceeding around 100 lines. We overcome this limitation by developing a modular approach to translation, where we partition the code into small code fragments which can be translated independently and semantically validated (that is, checking I/O equivalence). When this approach is applied naively, we discover that LLMs are unreliable when translating features of the source language that do not have a direct mapping to the target language, and that the LLM often gets stuck in repair loops when attempting to fix errors. To address these issues, we introduce two key concepts: (1) feature mapping, which integrates predefined translation rules with LLM-based translation to guide the LLM in navigating subtle language differences and producing semantically accurate code; and (2) type-compatibility, which facilitates localized checks at the function signature level to detect errors early, thereby narrowing the scope of potential repairs. We apply our approach to translating real-world Go codebases to Rust, demonstrating that we can consistently generate reliable Rust translations for projects up to 6,600 lines of code and 369 functions, with an average of 73% of functions successfully validated for I/O equivalence, considerably higher than any existing work.

cs.PL↗

arXiv2024

Quantifying the benefits of code hints for refactoring deprecated Java APIs

When done manually, refactoring legacy code in order to eliminate uses of deprecated APIs is an error-prone and time-consuming process. In this paper, we investigate to which degree refactorings for deprecated Java APIs can be automated, and quantify the benefit of Javadoc code hints for this task. To this end, we build a symbolic and a neural engine for the automatic refactoring of deprecated APIs. The former is based on type-directed and component-based program synthesis, whereas the latter uses LLMs. We applied our engines to refactor the deprecated methods in the Oracle JDK 15. Our experiments show that code hints are enabling for the automation of this task: even the worst engine correctly refactors 71% of the tasks with code hints, which drops to at best 14% on tasks without. Adding more code hints to Javadoc can hence boost the refactoring of code that uses deprecated APIs.

cs.SE↗

arXiv2024

Integrating Large Language Models and Reinforcement Learning for Non-Linear Reasoning

Large Language Models (LLMs) were shown to struggle with long-term planning, which may be caused by the limited way in which they explore the space of possible solutions. We propose an architecture where a Reinforcement Learning (RL) Agent guides an LLM's space exploration: (1) the Agent has access to domain-specific information, and can therefore make decisions about the quality of candidate solutions based on specific and relevant metrics, which were not explicitly considered by the LLM's training objective; (2) the LLM can focus on generating immediate next steps, without the need for long-term planning. We allow non-linear reasoning by exploring alternative paths and backtracking. We evaluate this architecture on the program equivalence task, and compare it against Chain of Thought (CoT) and Tree of Thoughts (ToT). We assess both the downstream task, denoting the binary classification, and the intermediate reasoning steps. Our approach compares positively against CoT and ToT.

cs.LG↗

arXiv2024

Enhancing SQL Query Generation with Neurosymbolic Reasoning

Neurosymbolic approaches blend the effectiveness of symbolic reasoning with the flexibility of neural networks. In this work, we propose a neurosymbolic architecture for generating SQL queries that builds and explores a solution tree using Best-First Search, with the possibility of backtracking. For this purpose, it integrates a Language Model (LM) with symbolic modules that help catch and correct errors made by the LM on SQL queries, as well as guiding the exploration of the solution tree. We focus on improving the performance of smaller open-source LMs, and we find that our tool, Xander, increases accuracy by an average of 10.9% and reduces runtime by an average of 28% compared to the LM without Xander, enabling a smaller LM (with Xander) to outperform its four-times larger counterpart (without Xander).

cs.DB↗

arXiv2024

Towards Translating Real-World Code with LLMs: A Study of Translating to Rust

Large language models (LLMs) show promise in code translation - the task of translating code written in one programming language to another language - due to their ability to write code in most programming languages. However, LLM's effectiveness on translating real-world code remains largely unstudied. In this work, we perform the first substantial study on LLM-based translation to Rust by assessing the ability of five state-of-the-art LLMs, GPT4, Claude 3, Claude 2.1, Gemini Pro, and Mixtral. We conduct our study on code extracted from real-world open source projects. To enable our study, we develop FLOURINE, an end-to-end code translation tool that uses differential fuzzing to check if a Rust translation is I/O equivalent to the original source program, eliminating the need for pre-existing test cases. As part of our investigation, we assess both the LLM's ability to produce an initially successful translation, as well as their capacity to fix a previously generated buggy one. If the original and the translated programs are not I/O equivalent, we apply a set of automated feedback strategies, including feedback to the LLM with counterexamples. Our results show that the most successful LLM can translate 47% of our benchmarks, and also provides insights into next steps for improvements.

cs.SE↗

arXiv2024

Cognacy Queries over Dependence Graphs for Transparent Visualisations

Charts, figures, and text derived from data play an important role in decision making, from data-driven policy development to day-to-day choices informed by online articles. Making sense of, or fact-checking, outputs means understanding how they relate to the underlying data. Even for domain experts with access to the source code and data sets, this poses a significant challenge. In this paper we introduce a new program analysis framework which supports interactive exploration of fine-grained I/O relationships directly through computed outputs, making use of dynamic dependence graphs. Our main contribution is a novel notion in data provenance which we call related inputs, a relation of mutual relevance or "cognacy" which arises between inputs when they contribute to common features of the output. Queries of this form allow readers to ask questions like "What outputs use this data element, and what other data elements are used along with it?". We show how Jonsson and Tarski's concept of conjugate operators on Boolean algebras appropriately characterises the notion of cognacy in a dependence graph, and give a procedure for computing related inputs over such a graph.

cs.PL↗

arXiv2023

Ownership guided C to Rust translation

Dubbed a safer C, Rust is a modern programming language that combines memory safety and low-level control. This interesting combination has made Rust very popular among developers and there is a growing trend of migrating legacy codebases (very often in C) to Rust. In this paper, we present a C to Rust translation approach centred around static ownership analysis. We design a suite of analyses that infer ownership models of C pointers and automatically translate the pointers into safe Rust equivalents. The resulting tool, Crown, scales to real-world codebases (half a million lines of code in less than 10 seconds) and achieves a high conversion rate.

cs.PL↗

arXiv2022

Decomposition Without Regret

Programming languages are embracing both functional and object-oriented paradigms. A key difference between the two paradigms is the way of achieving data abstraction. That is, how to organize data with associated operations. There are important tradeoffs between functional and object-oriented decomposition in terms of extensibility and expressiveness. Unfortunately, programmers are usually forced to select a particular decomposition style in the early stage of programming. Once the wrong design decision has been made, the price for switching to the other decomposition style could be rather high since pervasive manual refactoring is often needed. To address this issue, this paper presents a bidirectional transformation system between functional and object-oriented decomposition. We formalize the core of the system in the FOOD calculus, which captures the essence of functional and object-oriented decomposition. We prove that the transformation preserves the type and semantics of the original program. We further implement FOOD in Scala as a translation tool called Cook and conduct several case studies to demonstrate the applicability and effectiveness of Cook.

cs.PL↗

arXiv2022

Fast Parallel Hypertree Decompositions in Logarithmic Recursion Depth

Modern trends in data collection are bringing current mainstream techniques for database query processing to their limits. Consequently, various novel approaches for efficient query processing are being actively studied. One such approach is based on hypertree decompositions (HDs), which have been shown to carry great potential to process complex queries more efficiently and with stronger theoretical guarantees. However, using HDs for query execution relies on the difficult task of computing decompositions of the query structure, which guides the efficient execution of the query. From theoretical results we know that the performance of purely sequential methods is inherently limited, yet the problem is susceptible to parallelisation. In this paper we propose the first algorithm for computing hypertree decompositions that is well-suited for parallelisation. The proposed algorithm log-k-decomp requires only a logarithmic number of recursion levels and additionally allows for highly parallelised pruning of the search space by restriction to balanced separators. We provide detailed experimental evaluation over the HyperBench benchmark and demonstrate that our approach is highly effective especially for complex queries.

cs.DB↗

arXiv2021

Synbit: Synthesizing Bidirectional Programs using Unidirectional Sketches

We propose a technique for synthesizing bidirectional programs from the corresponding unidirectional code plus a few input/output examples. The core ideas are: (1) constructing a sketch using the given unidirectional program as a specification, and (2) filling the sketch in a modular fashion by exploiting the properties of bidirectional programs. These ideas are enabled by our choice of programming language, HOBiT, which is specifically designed to maintain the unidirectional program structure in bidirectional programming, and keep the parts that control bidirectional behavior modular. To evaluate our approach, we implemented it in a tool called Synbit and used it to generate bidirectional programs for intricate microbenchmarks, as well as for a few larger, more realistic problems. We also compared Synbit to a state-of-the-art unidirectional synthesis tool on the task of synthesizing backward computations.

cs.PL↗

arXiv2017

Kayak: Safe Semantic Refactoring to Java Streams

Refactorings are structured changes to existing software that leave its externally observable behaviour unchanged. Their intent is to improve readability, performance or other non-behavioural properties. State-of-the-art automatic refactoring tools are syntax-driven and, therefore, overly conservative. In this paper we explore semantics-driven refactoring, which enables much more sophisticated refactoring schemata. As an exemplar of this broader idea, we present Kayak, an automatic refactoring tool that transforms Java with external iteration over collections into code that uses Streams, a new abstraction introduced by Java 8. Our refactoring procedure performs semantic reasoning and search in the space of possible refactorings using automated program synthesis. Our experimental results support the conjecture that semantics-driven refactorings are more precise and are able to rewrite more complex code scenarios when compared to syntax-driven refactorings.

cs.PL↗