arXiv subjects

Mike Papadakis

Publications and source records attributed to Mike Papadakis.

At least 37 records · Page 2Linked to original sources

arXiv2026

Automatically Evolving Prompt Guidelines for Task-Specific Optimization

For Large Language Models to reliably answer user queries, users must clearly specify requirements, context, and constraints. In practice, however, user queries are often underspecified, forcing models to infer unstated assumptions that may misalign with the actual user intent. Existing prompt engineering guidelines aim to mitigate this issue, they are typically generic and task-agnostic, limiting their practical utility. Additionally, existing guidelines are formed manually and in a non-systematic way. To this end, we study prompt guideline optimization: the problem of automatically generating task-specific guidelines that help write better-specified prompts for a given task and model. Our key observation is that existing (completed) task examples (aka reference answers) often implicitly encode the missing information required to complete underspecified queries, including behavioral constraints, contextual assumptions, and evaluation criteria. We therefore propose AGOPS, an automatic approach that evolves task-specific guidelines via an optimization scheme that involves a prompt LLM writer, a solver LLM and prompt evolution, which maximize downstream effectiveness on a set of examples (user queries with reference answers). At inference time, our guidelines help users write well-specified prompts, boosting the effectiveness of LLMs. We show across mathematical reasoning, medical question answering, and coding tasks, that prompt underspecification leads to major drops (up to 95.3%) in downstream task performance (compared to well-specified prompts) and, perhaps more importantly, that this drop can hardly be recovered by existing prompt optimization techniques. Users following AGOPS guidelines can regain this loss (increasing performance between 15.5 to 81.7% on average) consistently across all benchmarks.

cs.CL↗

arXiv2026

Learning Generalizable Multimodal Representations for Software Vulnerability Detection

Source code and its accompanying comments are complementary yet naturally aligned modalities-code encodes structural logic while comments capture developer intent. However, existing vulnerability detection methods mostly rely on single-modality code representations, overlooking the complementary semantic information embedded in comments and thus limiting their generalization across complex code structures and logical relationships. To address this, we propose MultiVul, a multimodal contrastive framework that aligns code and comment representations through dual similarity learning and consistency regularization, augmented with diverse code-text pairs to improve robustness. Experiments on widely adopted DiverseVul and Devign datasets across four large language models (LLMs) (i.e., DeepSeek-Coder-6.7B, Qwen2.5-Coder-7B, StarCoder2-7B, and CodeLlama-7B) show that MultiVul achieves up to 27.07% F1 improvement over prompting-based methods and 13.37% over code-only Fine-Tuning, while maintaining comparable inference efficiency.

cs.SE↗

arXiv2026

Defective Task Descriptions in LLM-Based Code Generation: Detection and Analysis

Large language models are widely used for code generation, yet they rely on an implicit assumption that the task descriptions are sufficiently detailed and well-formed. However, in practice, users may provide defective descriptions, which can have a strong effect on code correctness. To address this issue, we develop SpecValidator, a lightweight classifier based on a small model that has been parameter-efficiently finetuned, to automatically detect task description defects. We evaluate SpecValidator on three types of defects, Lexical Vagueness, Under-Specification and Syntax-Formatting on 3 benchmarks with task descriptions of varying structure and complexity. Our results show that SpecValidator achieves defect detection of F1 = 0.804 and MCC = 0.745, significantly outperforming GPT-5-mini (F1 = 0.469 and MCC = 0.281) and Claude Sonnet 4 (F1 = 0.518 and MCC = 0.359). Perhaps more importantly, our analysis indicates that SpecValidator can generalize to unseen issues and detect unknown Under-Specification defects in the original (real) descriptions of the benchmarks used. Our results also show that the robustness of LLMs in task description defects depends primarily on the type of defect and the characteristics of the task description, rather than the capacity of the model, with Under-Specification defects being the most severe. We further found that benchmarks with richer contextual grounding, such as LiveCodeBench, exhibit substantially greater resilience, highlighting the importance of structured task descriptions for reliable LLM-based code generation.

cs.SE↗

arXiv2026

On the Robustness of Tabular Foundation Models: Test-Time Attacks and In-Context Defenses

Recent tabular Foundational Models (FM) such as TabPFN and TabICL, leverage in-context learning to achieve strong performance without gradient updates or fine-tuning. However, their robustness to adversarial manipulation remains largely unexplored. In this work, we present a comprehensive study of the adversarial vulnerabilities of tabular FM, focusing on both their fragility to targeted test-time attacks and their potential misuse as adversarial tools. We show on three benchmarks in finance, cybersecurity and healthcare, that small, structured perturbations to test inputs can significantly degrade prediction accuracy, even when training context remain fixed. Additionally, we demonstrate that tabular FM can be repurposed to generate transferable evasion to conventional models such as random forests and XGBoost, and on a lesser extent to deep tabular models. To improve tabular FM, we formulate the robustification problem as an optimization of the weights (adversarial fine-tuning), or the context (adversarial in-context learning). We introduce an in-context adversarial training strategy that incrementally replaces the context with adversarial perturbed instances, without updating model weights. Our approach improves robustness across multiple tabular benchmarks. Together, these findings position tabular FM as both a target and a source of adversarial threats, highlighting the urgent need for robust training and evaluation practices in this emerging paradigm.

cs.LG↗

arXiv2026

Boosting LLMs for Mutation Generation

LLM-based mutation testing is a promising testing technology, but existing approaches typically rely on a fixed set of mutations as few-shot examples or none at all. This can result in generic low-quality mutations, missed context-specific mutation patterns, substantial numbers of redundant and uncompilable mutants, and limited semantic similarity to real bugs. To overcome these limitations, we introduce SMART (Semantic Mutation with Adaptive Retrieval and Tuning). SMART integrates retrieval-augmented generation (RAG) on a vectorized dataset of real-world bugs, focused code chunking, and supervised fine-tuning using mutations coupled with real-world bugs. We conducted an extensive empirical study of SMART using 1,991 real-world Java bugs from the Defects4J and ConDefects datasets, comparing SMART to the state-of-the-art LLM-based approaches, LLMut and LLMorpheus. The results reveal that SMART substantially improves mutation validity, effectiveness, and efficiency (even enabling small-scale 7B-scale models to match or even surpass large models like GPT-4o). We also demonstrate that SMART significantly improves downstream software engineering applications, including test case prioritization and fault localization. More specifically, SMART improves validity (weighted average generation rate) from 42.89% to 65.6%. It raises the non-duplicate rate from 87.38% to 95.62%, and the compilable rate from 88.85% to 90.21%. In terms of effectiveness, it achieves a real bug detection rate of 92.61% (vs. 57.86% for LLMut) and improves the average Ochiai coefficient from 25.61% to 38.44%. For fault localization, SMART ranks 64 more bugs as Top-1 under MUSE and 57 more under Metallaxis.

cs.SE↗

arXiv2026

Impact of LLMs news Sentiment Analysis on Stock Price Movement Prediction

This paper addresses stock price movement prediction by leveraging LLM-based news sentiment analysis. Earlier works have largely focused on proposing and assessing sentiment analysis models and stock movement prediction methods, however, separately. Although promising results have been achieved, a clear and in-depth understanding of the benefit of the news sentiment to this task, as well as a comprehensive assessment of different architecture types in this context, is still lacking. Herein, we conduct an evaluation study that compares 3 different LLMs, namely, DeBERTa, RoBERTa and FinBERT, for sentiment-driven stock prediction. Our results suggest that DeBERTa outperforms the other two models with an accuracy of 75% and that an ensemble model that combines the three models can increase the accuracy to about 80%. Also, we see that sentiment news features can benefit (slightly) some stock market prediction models, i.e., LSTM-, PatchTST- and tPatchGNN-based classifiers and PatchTST- and TimesNet-based regression tasks models.

q-fin.ST↗

arXiv2026

Towards a more efficient bias detection in financial language models

Bias in financial language models constitutes a major obstacle to their adoption in real-world applications. Detecting such bias is challenging, as it requires identifying inputs whose predictions change when varying properties unrelated to the decision, such as demographic attributes. Existing approaches typically rely on exhaustive mutation and pairwise prediction analysis over large corpora, which is effective but computationally expensive-particularly for large language models and can become impractical in continuous retraining and releasing processes. Aiming at reducing this cost, we conduct a large-scale study of bias in five financial language models, examining similarities in their bias tendencies across protected attributes and exploring cross-model-guided bias detection to identify bias-revealing inputs earlier. Our study uses approximately 17k real financial news sentences, mutated to construct over 125k original-mutant pairs. Results show that all models exhibit bias under both atomic (0.58\%-6.05\%) and intersectional (0.75\%-5.97\%) settings. Moreover, we observe consistent patterns in bias-revealing inputs across models, enabling substantial reuse and cost reduction in bias detection. For example, up to 73\% of FinMA's biased behaviours can be uncovered using only 20\% of the input pairs when guided by properties derived from DistilRoBERTa outputs.

cs.AI↗

arXiv2026

One Model, Many Skills: Parameter-Efficient Fine-Tuning for Multitask Code Analysis

Large language models have recently surpassed specialized systems on code generation, yet their effectiveness on other code-analysis tasks remains less clear. At the same time, multi-task learning offers a way to unify diverse objectives within a single model, but fully fine-tuning LLMs across tasks is computationally prohibitive. Parameter-efficient fine-tuning mitigates this cost by updating only a small fraction of weights. Although PEFT has proven effective in single-task settings, its potential for multi-task learning has not yet been systematically explored. We present the first comprehensive evaluation of multi-task PEFT for code analysis, comparing several methods across diverse tasks and model architectures. Our experiments show that a single PEFT module shared across tasks can match, and in some cases surpass, full multi-task fine-tuning, confirming that the benefits of PEFT extend beyond isolated tasks. When comparing single-task and multi-task setups, we find that multi-task PEFT achieves a favorable performance-efficiency trade-off: it delivers accuracy close to single-task fine-tuning while reducing storage requirements, cutting the number of trainable parameters by a factor of the task count, and lowering computation costs by as much as 85%. At the same time, multi-task gains remain sensitive to task grouping. Through task-pairing experiments, we identify key factors shaping outcomes: task stability, model architecture, task complementarity, asymmetry, and dataset quality determine the success of co-fine-tuning. Finally, we benchmark efficient multi-task PEFT against direct prompting of open-source general-purpose LLMs, including DeepSeek, Qwen, Mistral, CodeLlama, and StarCoder. Despite their strong performance in code generation, these models underperform on analysis tasks, where even a 1B-parameter model with multi-task PEFT achieves significantly better results.

cs.SE↗

arXiv2026

GenCode: A Generic Data Augmentation Framework for Boosting Deep Learning-Based Code Understanding

Pre-trained code models lead the era of code intelligence, with multiple models designed with impressive performance. However, one important problem, data augmentation for code data that automatically helps developers prepare training data lacks study in this field. In this paper, we introduce a generic data augmentation framework, GenCode, to enhance the training of code understanding models. Simply speaking, GenCode follows a generation-and-selection paradigm to prepare useful training code data. Specifically, it employs code augmentation techniques to generate new code candidates first and then identifies important ones as the training data by influence scores. To evaluate the effectiveness of GenCode, we conduct experiments on four code understanding tasks (e.g., code clone detection) and three pre-trained code models (e.g., CodeT5) and two recent released code-specific Large Language Models (LLMs) (e.g., Qwen2.5-Coder). Compared to the state-of-the-art (SOTA) code augmentation method MixCode, GenCode produces pre-trained code models with 2.92% higher accuracy and 4.90% adversarial robustness on average. For code-specific LLMs, GenCode achieves an average improvement of 0.93% in accuracy and 0.98% in natural robustness.

cs.SE↗

arXiv2026

A Comprehensive Study on Large Language Models for Mutation Testing

Large Language Models (LLMs) have recently been used to generate mutants in both research work and in industrial practice. However, there has been no comprehensive empirical study of their performance for this increasingly important LLM-based Software Engineering application. To address this, we conduct a comprehensive empirical study evaluating BugFarm and LLMorpheus (the two state-of-the-art LLM-based approaches), alongside seven LLMs using our newly designed prompt, including both leading open- and closed-source models, on 851 real bugs from two Java real-world bug benchmarks. Our results reveal that, compared to existing rule-based approaches, LLMs generate more diverse mutants, that are behaviorally closer to real bugs and, most importantly, with 111.29% higher fault detection. That is, 87.98% (for LLMs) vs. 41.64% (for rule-based); an increase of 46.34 percentage points. Nevertheless, our results also reveal that these impressive results for improved effectiveness come at a cost: the LLM-generated mutants have worse non-compilability, duplication, and equivalent mutant rates by 26.60, 10.14, and 3.51 percentage points, respectively. These findings are immediately actionable for both research and practice. They allow practitioners to have greater confidence in deploying LLM-based mutation, while researchers now have a baseline for the state-of-the-art, with which they can research techniques to further improve effectiveness and reduce cost.

cs.SE↗

arXiv2025

When Prompts Go Wrong: Evaluating Code Model Robustness to Ambiguous, Contradictory, and Incomplete Task Descriptions

Large Language Models (LLMs) have demonstrated impressive performance in code generation tasks under idealized conditions, where task descriptions are clear and precise. However, in practice, task descriptions frequently exhibit ambiguity, incompleteness, or internal contradictions. In this paper, we present the first empirical study examining the robustness of state-of-the-art code generation models when faced with such unclear task descriptions. We extend the HumanEval and MBPP benchmarks by systematically introducing realistic task descriptions flaws through guided mutation strategies, producing a dataset that mirrors the messiness of informal developer instructions. We evaluate multiple LLMs of varying sizes and architectures, analyzing their functional correctness and failure modes across task descriptions categories. Our findings reveal that even minor imperfections in task description phrasing can cause significant performance degradation, with contradictory task descriptions resulting in numerous logical errors. Moreover, while larger models tend to be more resilient than smaller variants, they are not immune to the challenges posed by unclear requirements. We further analyze semantic error patterns and identify correlations between description clarity, model behavior, and error types. Our results underscore the critical need for developing LLMs that are not only powerful but also robust to the imperfections inherent in natural user tasks, highlighting important considerations for improving model training strategies, designing more realistic evaluation benchmarks, and ensuring reliable deployment in practical software development environments.

cs.SE↗

arXiv2025

YATE: The Role of Test Repair in LLM-Based Unit Test Generation

Recent advances in automated test generation utilises language models to produce unit tests. While effective, language models tend to generate many incorrect tests with respect to both syntax and semantics. Although such incorrect tests can be easily detected and discarded, they constitute a "missed opportunity" -- if fixed, they are often valuable as they directly add testing value (they effectively target the underlying program logic to be tested) and indirectly form good seeds for generating additional tests. To this end, we propose a simple technique for repairing some of these incorrect tests through a combination of rule-based static analysis and re-prompting. We evaluate this simple approach, named YATE, on a set of 6 open-source projects and show that it can effectively produce tests that cover on average 32.06% more lines and kill 21.77% more mutants than a plain LLM-based method. We also compare YATE with four other LLM-based methods, namely HITS, SYMPROMPT, TESTSPARK and COVERUP and show that it produces tests that cover substantially more code. YATE achieves 22% higher line coverage, 20% higher branch coverage and kill 20% more mutants at a comparable cost (number of calls to LLMs).

cs.SE↗

arXiv2025

You Can REST Now: Automated REST API Documentation and Testing via LLM-Assisted Request Mutations

REST APIs are prevalent among web service implementations, easing interoperability through the HTTP protocol. API testers and users exploit the widely adopted OpenAPI Specification (OAS), a machine-readable standard to document REST APIs. However, documenting APIs is a time-consuming and error-prone task, and existing documentation is not always complete, publicly accessible, or up-to-date. This situation limits the efficiency of testing tools and hinders human comprehension. Large Language Models (LLMs) offer the potential to automatically infer API documentation, using their colossal training data. In this paper, we present RESTSpecIT, the first automated approach that infers documentation and performs black-box testing of REST APIs by leveraging LLMs. Our approach requires minimal user input compared to state-of-the-art tools; Given an API name and an LLM access key, RESTSpecIT generates API request seeds and mutates them with data returned by the LLM. The tool then analyzes API responses for documentation inference and testing purposes. RESTSpecIT utilizes an in-context prompt masking strategy, requiring no prior model fine-tuning. We evaluate the quality of our tool with three state-of-the-art LLMs: DeepSeek V3, GPT-4.1, and GPT-3.5. Our evaluation demonstrates that RESTSpecIT can (1) infer documentation with 88.62% of routes and 89.25% of query parameters found on average, (2) discover undocumented API data, (3) operate efficiently (in terms of model costs, requests sent, runtime), and (4) assist REST API testing by uncovering server errors and generating valid OpenAPI Specification inputs for testing tools.

cs.SE↗

arXiv2025

HInter: Exposing Hidden Intersectional Bias in Large Language Models

Large Language Models (LLMs) may portray discrimination towards certain individuals, especially those characterized by multiple attributes (aka intersectional bias). Discovering intersectional bias in LLMs is challenging, as it involves complex inputs on multiple attributes (e.g. race and gender). To address this challenge, we propose HInter, a test technique that synergistically combines mutation analysis, dependency parsing and metamorphic oracles to automatically detect intersectional bias in LLMs. HInter generates test inputs by systematically mutating sentences using multiple mutations, validates inputs via a dependency invariant and detects biases by checking the LLM response on the original and mutated sentences. We evaluate HInter using six LLM architectures and 18 LLM models (GPT3.5, Llama2, BERT, etc) and find that 14.61% of the inputs generated by HInter expose intersectional bias. Results also show that our dependency invariant reduces false positives (incorrect test inputs) by an order of magnitude. Finally, we observed that 16.62% of intersectional bias errors are hidden, meaning that their corresponding atomic cases do not trigger biases. Overall, this work emphasize the importance of testing LLMs for intersectional bias.

cs.CL↗

arXiv2025

Boosting Source Code Learning with Text-Oriented Data Augmentation: An Empirical Study

Recent studies have demonstrated remarkable advancements in source code learning, which applies deep neural networks (DNNs) to tackle various software engineering tasks. Similar to other DNN-based domains, source code learning also requires massive high-quality training data to achieve the success of these applications. Data augmentation, a technique used to produce additional training data, is widely adopted in other domains (e.g. computer vision). However, the existing practice of data augmentation in source code learning is limited to simple syntax-preserved methods, such as code refactoring. In this paper, considering that source code can also be represented as text data, we take an early step to investigate the effectiveness of data augmentation methods originally designed for natural language texts in the context of source code learning. To this end, we focus on code classification tasks and conduct a comprehensive empirical study across four critical code problems and four DNN architectures to assess the effectiveness of 25 data augmentation methods. Our results reveal specific data augmentation methods that yield more accurate and robust models for source code learning. Additionally, we discover that the data augmentation methods remain beneficial even when they slightly break source code syntax.

cs.SE↗

arXiv2024

On the Effectiveness of Hybrid Pooling in Mixup-Based Graph Learning for Language Processing

Graph neural network (GNN)-based graph learning has been popular in natural language and programming language processing, particularly in text and source code classification. Typically, GNNs are constructed by incorporating alternating layers which learn transformations of graph node features, along with graph pooling layers that use graph pooling operators (e.g., Max-pooling) to effectively reduce the number of nodes while preserving the semantic information of the graph. Recently, to enhance GNNs in graph learning tasks, Manifold-Mixup, a data augmentation technique that produces synthetic graph data by linearly mixing a pair of graph data and their labels, has been widely adopted. However, the performance of Manifold-Mixup can be highly affected by graph pooling operators, and there have not been many studies that are dedicated to uncovering such affection. To bridge this gap, we take an early step to explore how graph pooling operators affect the performance of Mixup-based graph learning. To that end, we conduct a comprehensive empirical study by applying Manifold-Mixup to a formal characterization of graph pooling based on 11 graph pooling operations (9 hybrid pooling operators, 2 non-hybrid pooling operators). The experimental results on both natural language datasets (Gossipcop, Politifact) and programming language datasets (JAVA250, Python800) demonstrate that hybrid pooling operators are more effective for Manifold-Mixup than the standard Max-pooling and the state-of-the-art graph multiset transformer (GMT) pooling, in terms of producing more accurate and robust GNN models.

cs.LG↗

arXiv2023

Large Language Models in Fault Localisation

Large Language Models (LLMs) have shown promise in multiple software engineering tasks including code generation, program repair, code summarisation, and test generation. Fault localisation is instrumental in enabling automated debugging and repair of programs and was prominently featured as a highlight during the launch event of ChatGPT-4. Nevertheless, the performance of LLMs compared to state-of-the-art methods, as well as the impact of prompt design and context length on their efficacy, remains unclear. To fill this gap, this paper presents an in-depth investigation into the capability of ChatGPT-3.5 and ChatGPT-4, the two state-of-the-art LLMs, on fault localisation. Using the widely-adopted large-scale Defects4J dataset, we compare the two LLMs with the existing fault localisation techniques. We also investigate the consistency of LLMs in fault localisation, as well as how prompt engineering and the length of code context affect the fault localisation effectiveness. Our findings demonstrate that within function-level context, ChatGPT-4 outperforms all the existing fault localisation methods. Additional error logs can further improve ChatGPT models' localisation accuracy and consistency, with an average 46.9% higher accuracy over the state-of-the-art baseline SmartFL on the Defects4J dataset in terms of TOP-1 metric. However, when the code context of the Defects4J dataset expands to the class-level, ChatGPT-4's performance suffers a significant drop, with 49.9% lower accuracy than SmartFL under TOP-1 metric. These observations indicate that although ChatGPT can effectively localise faults under specific conditions, limitations are evident. Further research is needed to fully harness the potential of LLMs like ChatGPT for practical fault localisation applications.

cs.SE↗

arXiv2023

Hazards in Deep Learning Testing: Prevalence, Impact and Recommendations

Much research on Machine Learning testing relies on empirical studies that evaluate and show their potential. However, in this context empirical results are sensitive to a number of parameters that can adversely impact the results of the experiments and potentially lead to wrong conclusions (Type I errors, i.e., incorrectly rejecting the Null Hypothesis). To this end, we survey the related literature and identify 10 commonly adopted empirical evaluation hazards that may significantly impact experimental results. We then perform a sensitivity analysis on 30 influential studies that were published in top-tier SE venues, against our hazard set and demonstrate their criticality. Our findings indicate that all 10 hazards we identify have the potential to invalidate experimental findings, such as those made by the related literature, and should be handled properly. Going a step further, we propose a point set of 10 good empirical practices that has the potential to mitigate the impact of the hazards. We believe our work forms the first step towards raising awareness of the common pitfalls and good practices within the software engineering community and hopefully contribute towards setting particular expectations for empirical research in the field of deep learning testing.

cs.SE↗