arXiv subjects

Vincent Ng

Publications and source records attributed to Vincent Ng.

At least 19 recordsLinked to original sources

arXiv2026

An Echo Chamber of One: Should AI Psychosis Be a Distinct Clinical Entity?

"AI psychosis" has entered public and clinical discourse as a label for the onset or exacerbation of psychotic symptoms, most commonly delusions, following intensive interaction with large language model (LLM)-based chatbots. Current evidence is limited to media reports, case reports, and early observational data, yet the scale of potential exposure is considerable, and public concern has prompted responses from industry and regulators. We examine whether AI-associated psychosis warrants recognition as a distinct clinical entity, drawing on clinical and technical viewpoints. We outline the proposed mechanism: LLM sycophancy, a tendency to agree with and flatter users that is reinforced through preference-based fine-tuning, combines with increasingly anthropomorphic design to create a bidirectional "echo chamber of one" capable of amplifying and co-constructing unusual beliefs. We then weigh arguments for and against nosological recognition. Potential benefits include improved case identification, tailored interventions, standardised research criteria, post-market surveillance, and pressure on developers and regulators to act. Reasons for caution include the risk of prematurely reifying a syndrome from anecdotal evidence, the possibility that existing diagnostic constructs already accommodate AI use as a contributing factor, the unproven causal claim in the term itself, stigma, and the risk that a psychosis-centric label obscures a broader spectrum of AI-associated mental health harms. We conclude with recommendations for clinicians, developers, researchers, and regulators, including a "technological history" in psychiatric assessment, pre-deployment benchmarking for sycophancy and delusion reinforcement, and post-deployment surveillance. Regardless of whether AI-associated psychosis earns a place in psychiatric nosology, the phenomenon it describes demands coordinated attention now.

cs.CY↗

arXiv2026

ICLE++: Modeling Fine-Grained Traits for Holistic Essay Scoring

The majority of the recently-developed models for automated essay scoring (AES) are evaluated solely on the ASAP corpus. However, ASAP is not without its limitations. For instance, it is not clear whether models trained on ASAP can generalize well when evaluated on other corpora. In light of these limitations, we introduce ICLE++, a corpus of persuasive student essays annotated with both holistic scores and trait-specific scores. Not only can ICLE++ be used to test the generalizability of AES models trained on ASAP, but it can also facilitate the evaluation of models developed for newer AES problems such as multi-trait scoring and cross-prompt scoring. We believe that ICLE++, which represents a culmination of our long-term effort in annotating the essays in the ICLE corpus, contributes to the set of much-needed annotated corpora for AES research.

cs.CL↗

arXiv2026

FailureMem: A Failure-Aware Multimodal Framework for Autonomous Software Repair

Multimodal Automated Program Repair (MAPR) extends traditional program repair by requiring models to jointly reason over source code, textual issue descriptions, and visual artifacts such as GUI screenshots. While recent LLM-based repair systems have shown promising results, existing approaches face several limitations: rigid workflow pipelines restrict exploration during debugging, visual reasoning is often performed over full-page screenshots without localized grounding, and failed repair attempts are rarely transformed into reusable knowledge. To address these challenges, we propose FailureMem, a multimodal repair framework that integrates three key mechanisms: a hybrid workflow-agent architecture that balances structured localization with flexible reasoning, active perception tools that enable region-level visual grounding, and a Failure Memory Bank that converts past repair attempts into reusable guidance. Experiments on SWE-bench Multimodal demonstrate FailureMem improves the resolved rate over GUIRepair by 3.7%.

cs.SE↗

arXiv2026

Echo: Graph-Enhanced Retrieval and Execution Feedback for Issue Reproduction Test Generation

Identifying the root cause of a bug remains difficult for many developers because bug reports often lack a bug reproducing test case that reliably triggers the failure. Manually writing such test cases is time-consuming and requires substantial effort to understand the codebase and isolate the failing behavior. To address this challenge, we propose Echo, an agent for generating issue reproducing test cases, which advances previous work in several ways. During generation, Echo strengthens context retrieval by leveraging a code graph and a novel automatic query-refinement strategy. Echo also improves upon previous tools by automatically executing generated test cases, a first-of-its-kind feature that seamlessly integrates into practical development workflows. In addition, Echo generates potential patches and uses the patched version to validate whether a candidate test meets the fail-to-pass criterion and to provide actionable feedback for refinement. Unlike prior bug-reproduction agents that sample and rank multiple candidate tests, Echo generates a single test per issue, offering a better cost-performance trade-off. Experiments on SWT-Bench Verified show that Echo establishes a new state of the art among open-source approaches, achieving a 66.28% success rate.

cs.SE↗

arXiv2026

RepoRepair: Leveraging Code Documentation for Repository-Level Automated Program Repair

Automated program repair (APR) struggles to scale from isolated functions to full repositories, as it demands a global, task-aware understanding to locate necessary changes. Current methods, limited by context and reliant on shallow retrieval or costly agent iterations, falter on complex cross-file issues. To this end, we propose RepoRepair, a novel documentation-enhanced approach for repository-level fault localization and program repair. Our core insight is to leverage LLMs to generate hierarchical code documentation (from functions to files) for code repositories, creating structured semantic abstractions that enable LLMs to comprehend repository-level context and dependencies. Specifically, RepoRepair first employs a text-based LLM (e.g., DeepSeek-V3) to generate file/function-level code documentation for repositories, which serves as auxiliary knowledge to guide fault localization. Subsequently, based on the fault localization results and the issue description, a powerful LLM (e.g., Claude-4) attempts to repair the identified suspicious code snippets. Evaluated on SWE-bench Lite, RepoRepair achieves a 45.7% repair rate at a low cost of $0.44 per fix. On SWE-bench Multimodal, it delivers state-of-the-art performance with a 37.1% repair rate despite a higher cost of $0.56 per fix, demonstrating robust and cost-effective performance across diverse problem domains.

cs.SE↗

arXiv2025

Enhancing Technical Documents Retrieval for RAG

In this paper, we introduce Technical-Embeddings, a novel framework designed to optimize semantic retrieval in technical documentation, with applications in both hardware and software development. Our approach addresses the challenges of understanding and retrieving complex technical content by leveraging the capabilities of Large Language Models (LLMs). First, we enhance user queries by generating expanded representations that better capture user intent and improve dataset diversity, thereby enriching the fine-tuning process for embedding models. Second, we apply summary extraction techniques to encode essential contextual information, refining the representation of technical documents. To further enhance retrieval performance, we fine-tune a bi-encoder BERT model using soft prompting, incorporating separate learning parameters for queries and document context to capture fine-grained semantic nuances. We evaluate our approach on two public datasets, RAG-EDA and Rust-Docs-QA, demonstrating that Technical-Embeddings significantly outperforms baseline models in both precision and recall. Our findings highlight the effectiveness of integrating query expansion and contextual summarization to enhance information access and comprehension in technical domains. This work advances the state of Retrieval-Augmented Generation (RAG) systems, offering new avenues for efficient and accurate technical document retrieval in engineering and product development workflows.

cs.IR↗

arXiv2025

A Sentence-Level Approach to Understanding Software Vulnerability Fixes

Understanding software vulnerabilities and their resolutions is crucial for securing modern software systems. This study presents a novel traceability model that links a pair of sentences describing at least one of the three types of semantics (triggers, crash phenomenon and fix action) for a vulnerability in natural language (NL) vulnerability artifacts, to their corresponding pair of code statements. Different from the traditional traceability models, our tracing links between a pair of related NL sentences and a pair of code statements can recover the semantic relationship between code statements so that the specific role played by each code statement in a vulnerability can be automatically identified. Our end-to-end approach is implemented in two key steps: VulnExtract and VulnTrace. VulnExtract automatically extracts sentences describing triggers, crash phenomenon and/or fix action for a vulnerability using 37 discourse patterns derived from NL artifacts (CVE summary, bug reports and commit messages). VulnTrace employs pre-trained code search models to trace these sentences to the corresponding code statements. Our empirical study, based on 341 CVEs and their associated code snippets, demonstrates the effectiveness of our approach, with recall exceeding 90% in most cases for NL sentence extraction. VulnTrace achieves a Top5 accuracy of over 68.2% for mapping a pair of related NL sentences to the corresponding pair of code statements. The end-to-end combined VulnExtract+VulnTrace achieves a Top5 accuracy of 59.6% and 53.1% for mapping two pairs of NL sentences to code statements. These results highlight the potential of our method in automating vulnerability comprehension and reducing manual effort.

cs.SE↗

arXiv2024

Automatic Prompt Generation and Grounding Object Detection for Zero-Shot Image Anomaly Detection

Identifying defects and anomalies in industrial products is a critical quality control task. Traditional manual inspection methods are slow, subjective, and error-prone. In this work, we propose a novel zero-shot training-free approach for automated industrial image anomaly detection using a multimodal machine learning pipeline, consisting of three foundation models. Our method first uses a large language model, i.e., GPT-3. generate text prompts describing the expected appearances of normal and abnormal products. We then use a grounding object detection model, called Grounding DINO, to locate the product in the image. Finally, we compare the cropped product image patches to the generated prompts using a zero-shot image-text matching model, called CLIP, to identify any anomalies. Our experiments on two datasets of industrial product images, namely MVTec-AD and VisA, demonstrate the effectiveness of this method, achieving high accuracy in detecting various types of defects and anomalies without the need for model training. Our proposed model enables efficient, scalable, and objective quality control in industrial manufacturing settings.

cs.CV↗

arXiv2024

On Evaluating the Efficiency of Source Code Generated by LLMs

Recent years have seen the remarkable capabilities of large language models (LLMs) for code generation. Different from existing work that evaluate the correctness of the code generated by LLMs, we propose to further evaluate its efficiency. More efficient code can lead to higher performance and execution efficiency of programs and software completed by LLM-assisted programming. First, we evaluate the efficiency of the code generated by LLMs on two benchmarks, HumanEval and MBPP. Then, we choose a set of programming problems from the online judge platform LeetCode to conduct a more difficult evaluation. Finally, we explore several prompts that would enable LLMs to generate more efficient code.

cs.SE↗

arXiv2023

Practical Program Repair via Preference-based Ensemble Strategy

To date, over 40 Automated Program Repair (APR) tools have been designed with varying bug-fixing strategies, which have been demonstrated to have complementary performance in terms of being effective for different bug classes. Intuitively, it should be feasible to improve the overall bug-fixing performance of APR via assembling existing tools. Unfortunately, simply invoking all available APR tools for a given bug can result in unacceptable costs on APR execution as well as on patch validation (via expensive testing). Therefore, while assembling existing tools is appealing, it requires an efficient strategy to reconcile the need to fix more bugs and the requirements for practicality. In light of this problem, we propose a Preference-based Ensemble Program Repair framework (P-EPR), which seeks to effectively rank APR tools for repairing different bugs. P-EPR is the first non-learning-based APR ensemble method that is novel in its exploitation of repair patterns as a major source of knowledge for ranking APR tools and its reliance on a dynamic update strategy that enables it to immediately exploit and benefit from newly derived repair results. Experimental results show that P-EPR outperforms existing strategies significantly both in flexibility and effectiveness.

cs.SE↗

arXiv2023

FAIR: Flow Type-Aware Pre-Training of Compiler Intermediate Representations

While the majority of existing pre-trained models from code learn source code features such as code tokens and abstract syntax trees, there are some other works that focus on learning from compiler intermediate representations (IRs). Existing IR-based models typically utilize IR features such as instructions, control and data flow graphs (CDFGs), call graphs, etc. However, these methods confuse variable nodes and instruction nodes in a CDFG and fail to distinguish different types of flows, and the neural networks they use fail to capture long-distance dependencies and have over-smoothing and over-squashing problems. To address these weaknesses, we propose FAIR, a Flow type-Aware pre-trained model for IR that involves employing (1) a novel input representation of IR programs; (2) Graph Transformer to address over-smoothing, over-squashing and long-dependencies problems; and (3) five pre-training tasks that we specifically propose to enable FAIR to learn the semantics of IR tokens, flow type information, and the overall representation of IR. Experimental results show that FAIR can achieve state-of-the-art results on four code-related downstream tasks.

cs.SE↗

arXiv2023

Oversampling Higher-Performing Minorities During Machine Learning Model Training Reduces Adverse Impact Slightly but Also Reduces Model Accuracy

Organizations are increasingly adopting machine learning (ML) for personnel assessment. However, concerns exist about fairness in designing and implementing ML assessments. Supervised ML models are trained to model patterns in data, meaning ML models tend to yield predictions that reflect subgroup differences in applicant attributes in the training data, regardless of the underlying cause of subgroup differences. In this study, we systematically under- and oversampled minority (Black and Hispanic) applicants to manipulate adverse impact ratios in training data and investigated how training data adverse impact ratios affect ML model adverse impact and accuracy. We used self-reports and interview transcripts from job applicants (N = 2,501) to train 9,702 ML models to predict screening decisions. We found that training data adverse impact related linearly to ML model adverse impact. However, removing adverse impact from training data only slightly reduced ML model adverse impact and tended to negatively affect ML model accuracy. We observed consistent effects across self-reports and interview transcripts, whether oversampling real (i.e., bootstrapping) or synthetic observations. As our study relied on limited predictor sets from one organization, the observed effects on adverse impact may be attenuated among more accurate ML models.

cs.LG↗

arXiv2023

Multimodal Propaganda Processing

Propaganda campaigns have long been used to influence public opinion via disseminating biased and/or misleading information. Despite the increasing prevalence of propaganda content on the Internet, few attempts have been made by AI researchers to analyze such content. We introduce the task of multimodal propaganda processing, where the goal is to automatically analyze propaganda content. We believe that this task presents a long-term challenge to AI researchers and that successful processing of propaganda could bring machine understanding one important step closer to human understanding. We discuss the technical challenges associated with this task and outline the steps that need to be taken to address it.

cs.CL↗

arXiv2023

CrossCodeBench: Benchmarking Cross-Task Generalization of Source Code Models

Despite the recent advances showing that a model pre-trained on large-scale source code data is able to gain appreciable generalization capability, it still requires a sizeable amount of data on the target task for fine-tuning. And the effectiveness of the model generalization is largely affected by the size and quality of the fine-tuning data, which is detrimental for target tasks with limited or unavailable resources. Therefore, cross-task generalization, with the goal of improving the generalization of the model to unseen tasks that have not been seen before, is of strong research and application value. In this paper, we propose a large-scale benchmark that includes 216 existing code-related tasks. Then, we annotate each task with the corresponding meta information such as task description and instruction, which contains detailed information about the task and a solution guide. This also helps us to easily create a wide variety of ``training/evaluation'' task splits to evaluate the various cross-task generalization capabilities of the model. Then we perform some preliminary experiments to demonstrate that the cross-task generalization of models can be largely improved by in-context learning methods such as few-shot learning and learning from task instructions, which shows the promising prospects of conducting cross-task learning research on our benchmark. We hope that the collection of the datasets and our benchmark will facilitate future work that is not limited to cross-task generalization.

cs.SE↗

arXiv2023

An Empirical Comparison of Pre-Trained Models of Source Code

While a large number of pre-trained models of source code have been successfully developed and applied to a variety of software engineering (SE) tasks in recent years, our understanding of these pre-trained models is arguably fairly limited. With the goal of advancing our understanding of these models, we perform the first systematic empirical comparison of 19 recently-developed pre-trained models of source code on 13 SE tasks. To gain additional insights into these models, we adopt a recently-developed 4-dimensional categorization of pre-trained models, and subsequently investigate whether there are correlations between different categories of pre-trained models and their performances on different SE tasks.

cs.SE↗

arXiv2022

End-to-End Neural Discourse Deixis Resolution in Dialogue

We adapt Lee et al.'s (2018) span-based entity coreference model to the task of end-to-end discourse deixis resolution in dialogue, specifically by proposing extensions to their model that exploit task-specific characteristics. The resulting model, dd-utt, achieves state-of-the-art results on the four datasets in the CODI-CRAC 2021 shared task.

cs.CL↗

arXiv2022

DiscoSense: Commonsense Reasoning with Discourse Connectives

We present DiscoSense, a benchmark for commonsense reasoning via understanding a wide variety of discourse connectives. We generate compelling distractors in DiscoSense using Conditional Adversarial Filtering, an extension of Adversarial Filtering that employs conditional generation. We show that state-of-the-art pre-trained language models struggle to perform well on DiscoSense, which makes this dataset ideal for evaluating next-generation commonsense reasoning systems.

cs.CL↗

arXiv2022

SPT-Code: Sequence-to-Sequence Pre-Training for Learning Source Code Representations

Recent years have seen the successful application of large pre-trained models to code representation learning, resulting in substantial improvements on many code-related downstream tasks. But there are issues surrounding their application to SE tasks. First, the majority of the pre-trained models focus on pre-training only the encoder of the Transformer. For generation tasks that are addressed using models with the encoder-decoder architecture, however, there is no reason why the decoder should be left out during pre-training. Second, many existing pre-trained models, including state-of-the-art models such as T5-learning, simply reuse the pre-training tasks designed for natural languages. Moreover, to learn the natural language description of source code needed eventually for code-related tasks such as code summarization, existing pre-training tasks require a bilingual corpus composed of source code and the associated natural language description, which severely limits the amount of data for pre-training. To this end, we propose SPT-Code, a sequence-to-sequence pre-trained model for source code. In order to pre-train SPT-Code in a sequence-to-sequence manner and address the aforementioned weaknesses associated with existing pre-training tasks, we introduce three pre-training tasks that are specifically designed to enable SPT-Code to learn knowledge of source code, the corresponding code structure, as well as a natural language description of the code without relying on any bilingual corpus, and eventually exploit these three sources of information when it is applied to downstream tasks. Experimental results demonstrate that SPT-Code achieves state-of-the-art performance on five code-related downstream tasks after fine-tuning.

cs.SE↗