Thu-Trang Nguyen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Thu-Trang Nguyen has a total of 38 co-authors. Here are the top co-authors by number of shared publications:
- Hieu Dinh Vo
Faculty of Information Technology VNU University of Engineering and Technology Hanoi Vietnam
19 shared publications - Son Nguyen
Faculty of Information Technology VNU University of Engineering and Technology Hanoi Vietnam
16 shared publications - Kien-Tuan Ngo
University of Southern California, Los Angeles, CA, USA
5 shared publications - Thanh Trong Vu4 shared publications
- Tuan-Dung Bui4 shared publications
- Phong Lam2 shared publications
Latest Publications
- 7 September 2026
- Published by Elsevier in Information and Software Technology
- Vol. 200
Rambo: Enhancing RAG-based repository-level method body completion
- 7 September 2026
- Published by Elsevier in Information and Software Technology
- Vol. 200
Rambo: Enhancing RAG-based repository-level method body completion
- 4 April 2026
- Published by Elsevier in Future Generation Computer Systems
- Vol. 182
A program-centered approach to solving AI tasks with ready-to-use models
- 4 April 2026
- Published by Elsevier in Future Generation Computer Systems
- Vol. 182
A program-centered approach to solving AI tasks with ready-to-use models
- 28 March 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 342
Structured exploration and exploitation of label functions for automated data annotation
- 28 March 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 342
Structured exploration and exploitation of label functions for automated data annotation
- 12 March 2026
- Published by Elsevier in Journal of Systems and Software
- Vol. 237
Model-agnostic quality assessment for LLM-generated code via dynamic internal representation selection
- 12 March 2026
- Published by Elsevier in Journal of Systems and Software
- Vol. 237
Model-agnostic quality assessment for LLM-generated code via dynamic internal representation selection
- Open Access
A Multi-Agent Framework for Code-Guided, Modular, and Verifiable Automated Machine Learning
Automated Machine Learning (AutoML) has revolutionized the development of data-driven solutions; however, traditional frameworks often function as "black boxes", lacking the flexibility and transparency required for complex, real-world engineering tasks. Recent Large Language Model (LLM)-based agents have shifted toward code-driven approaches. However, they frequently suffer from hallucinated logic and logic entanglement, where monolithic code generation leads to unrecoverable runtime failures. In this paper, we present iML, a novel multi-agent framework designed to shift AutoML from black-box prompting to a code-guided, modular, and verifiable architectural paradigm. iML introduces three main ideas: (1) Code-Guided Planning, which synthesizes a strategic blueprint grounded in autonomous empirical profiling to eliminate hallucination; (2) Code-Modular Implementation, which decouples preprocessing and modeling into specialized components governed by strict interface contracts; and (3) Code-Verifiable Integration, which enforces physical feasibility through dynamic contract verification and iterative self-correction. We evaluate iML across MLE-BENCH and the newly introduced iML-BENCH, comprising a diverse range of real-world Kaggle competitions. The experimental results show iML's superiority over state-of-the-art agents, achieving a valid submission rate of 85% and a competitive medal rate of 45% on MLE-BENCH, with an average standardized performance score (APS) of 0.77. On iML-BENCH, iML significantly outperforms the other approaches by 38%-163% in APS. Furthermore, iML maintains a robust 70% success rate even under stripped task descriptions, effectively filling information gaps through empirical profiling. These results highlight iML's potential to bridge the gap between stochastic generation and reliable engineering, marking a meaningful step toward truly AutoML.
Open AccessA Multi-Agent Framework for Code-Guided, Modular, and Verifiable Automated Machine Learning
Automated Machine Learning (AutoML) has revolutionized the development of data-driven solutions; however, traditional frameworks often function as "black boxes", lacking the flexibility and transparency required for complex, real-world engineering tasks. Recent Large Language Model (LLM)-based agents have shifted toward code-driven approaches. However, they frequently suffer from hallucinated logic and logic entanglement, where monolithic code generation leads to unrecoverable runtime failures. In this paper, we present iML, a novel multi-agent framework designed to shift AutoML from black-box prompting to a code-guided, modular, and verifiable architectural paradigm. iML introduces three main ideas: (1) Code-Guided Planning, which synthesizes a strategic blueprint grounded in autonomous empirical profiling to eliminate hallucination; (2) Code-Modular Implementation, which decouples preprocessing and modeling into specialized components governed by strict interface contracts; and (3) Code-Verifiable Integration, which enforces physical feasibility through dynamic contract verification and iterative self-correction. We evaluate iML across MLE-BENCH and the newly introduced iML-BENCH, comprising a diverse range of real-world Kaggle competitions. The experimental results show iML's superiority over state-of-the-art agents, achieving a valid submission rate of 85% and a competitive medal rate of 45% on MLE-BENCH, with an average standardized performance score (APS) of 0.77. On iML-BENCH, iML significantly outperforms the other approaches by 38%-163% in APS. Furthermore, iML maintains a robust 70% success rate even under stripped task descriptions, effectively filling information gaps through empirical profiling. These results highlight iML's potential to bridge the gap between stochastic generation and reliable engineering, marking a meaningful step toward truly AutoML.
- Open Access
Model-Agnostic Correctness Assessment for LLM-Generated Code via Dynamic Internal Representation Selection
Large Language Models (LLMs) have demonstrated impressive capabilities in code generation and are increasingly integrated into the software development process. However, ensuring the correctness of LLM-generated code remains a critical concern. Prior work has shown that the internal representations of LLMs encode meaningful signals for assessing code correctness. Nevertheless, the existing methods rely on representations from pre-selected/fixed layers and token positions, which could limit its generalizability across diverse model architectures and tasks. In this work, we introduce AUTOPROBE, a novel model-agnostic approach that dynamically selects the most informative internal representations for code correctness assessment. AUTOPROBE employs an attention-based mechanism to learn importance scores for hidden states, enabling it to focus on the most relevant features. These weighted representations are then aggregated and passed to a probing classifier to predict code correctness across multiple dimensions, including compilability, functionality, and security. To evaluate the performance of AUTOPROBE, we conduct extensive experiments across multiple benchmarks and code LLMs. Our experimental results show that AUTOPROBE consistently outperforms the baselines. For security assessment, AUTOPROBE surpasses the state-of-the-art white-box approach by 18%. For compilability and functionality assessment, AUTOPROBE demonstrates its highest robustness to code complexity, with the performance higher than the other approaches by up to 19% and 111%, respectively. These findings highlight that dynamically selecting important internal signals enables AUTOPROBE to serve as a robust and generalizable solution for assessing the correctness of code generated by various LLMs.
Open AccessModel-Agnostic Correctness Assessment for LLM-Generated Code via Dynamic Internal Representation Selection
Large Language Models (LLMs) have demonstrated impressive capabilities in code generation and are increasingly integrated into the software development process. However, ensuring the correctness of LLM-generated code remains a critical concern. Prior work has shown that the internal representations of LLMs encode meaningful signals for assessing code correctness. Nevertheless, the existing methods rely on representations from pre-selected/fixed layers and token positions, which could limit its generalizability across diverse model architectures and tasks. In this work, we introduce AUTOPROBE, a novel model-agnostic approach that dynamically selects the most informative internal representations for code correctness assessment. AUTOPROBE employs an attention-based mechanism to learn importance scores for hidden states, enabling it to focus on the most relevant features. These weighted representations are then aggregated and passed to a probing classifier to predict code correctness across multiple dimensions, including compilability, functionality, and security. To evaluate the performance of AUTOPROBE, we conduct extensive experiments across multiple benchmarks and code LLMs. Our experimental results show that AUTOPROBE consistently outperforms the baselines. For security assessment, AUTOPROBE surpasses the state-of-the-art white-box approach by 18%. For compilability and functionality assessment, AUTOPROBE demonstrates its highest robustness to code complexity, with the performance higher than the other approaches by up to 19% and 111%, respectively. These findings highlight that dynamically selecting important internal signals enables AUTOPROBE to serve as a robust and generalizable solution for assessing the correctness of code generated by various LLMs.
- Open Access
CABENCH: Benchmarking Composable AI for Solving Complex Tasks through Composing Ready-to-Use Models
Composable AI offers a scalable and effective paradigm for tackling complex AI tasks by decomposing them into sub-tasks and solving each sub-task using ready-to-use well-trained models. However, systematically evaluating methods under this setting remains largely unexplored. In this paper, we introduce CABENCH, the first public benchmark comprising 70 realistic composable AI tasks, along with a curated pool of 700 models across multiple modalities and domains. We also propose an evaluation framework to enable end-to-end assessment of composable AI solutions. To establish initial baselines, we provide human-designed reference solutions and compare their performance with two LLM-based approaches. Our results illustrate the promise of composable AI in addressing complex real-world problems while highlighting the need for methods that can fully unlock its potential by automatically generating effective execution pipelines.
Open AccessCABENCH: Benchmarking Composable AI for Solving Complex Tasks through Composing Ready-to-Use Models
Composable AI offers a scalable and effective paradigm for tackling complex AI tasks by decomposing them into sub-tasks and solving each sub-task using ready-to-use well-trained models. However, systematically evaluating methods under this setting remains largely unexplored. In this paper, we introduce CABENCH, the first public benchmark comprising 70 realistic composable AI tasks, along with a curated pool of 700 models across multiple modalities and domains. We also propose an evaluation framework to enable end-to-end assessment of composable AI solutions. To establish initial baselines, we provide human-designed reference solutions and compare their performance with two LLM-based approaches. Our results illustrate the promise of composable AI in addressing complex real-world problems while highlighting the need for methods that can fully unlock its potential by automatically generating effective execution pipelines.
- 19 July 2025
- Published by Elsevier in Journal of Systems and Software
- Vol. 230
Correctness assessment of code generated by Large Language Models using internal representations
- 19 July 2025
- Published by Elsevier in Journal of Systems and Software
- Vol. 230
Correctness assessment of code generated by Large Language Models using internal representations
- 22 May 2025
- Published by Elsevier in Information and Software Technology
- Vol. 185
An empirical study on capability of Large Language Models in understanding code semantics
- 22 May 2025
- Published by Elsevier in Information and Software Technology
- Vol. 185
An empirical study on capability of Large Language Models in understanding code semantics
- 24 January 2025
- Published by Elsevier in Future Generation Computer Systems
- Vol. 166
Leveraging local and global relationships for corrupted label detection
- 24 January 2025
- Published by Elsevier in Future Generation Computer Systems
- Vol. 166
Leveraging local and global relationships for corrupted label detection