arXiv subjects

Lingfei Wu

Publications and source records attributed to Lingfei Wu.

At least 37 records · Page 2Linked to original sources

arXiv2025

Hidden Division of Labor in Scientific Teams Revealed Through 1.6 Million LaTeX Files

Recognition of individual contributions is fundamental to the scientific reward system, yet coauthored papers obscure who did what. Traditional proxies-author order and career stage-reinforce biases, while contribution statements remain self-reported and limited to select journals. We construct the first large-scale dataset on writing contributions by analyzing author-specific macros in LaTeX files from 1.6 million papers (1991-2023) by 2 million scientists. Validation against self-reported statements (precision = 0.87), author order patterns, field-specific norms, and Overleaf records (Spearman's rho = 0.6, p < 0.05) confirms the reliability of the created data. Using explicit section information, we reveal a hidden division of labor within scientific teams: some authors primarily contribute to conceptual sections (e.g., Introduction and Discussion), while others focus on technical sections (e.g., Methods and Experiments). These findings provide the first large-scale evidence of implicit labor division in scientific teams, challenging conventional authorship practices and informing institutional policies on credit allocation.

cs.SI↗

arXiv2025

Large Teams Overshadow Individual Recognition

In an ideal world, every scientist's contribution would be fully recognized, driving collective scientific progress. In reality, however, only a few scientists are recognized and remembered. Sociologist Robert Merton first described this disparity between contribution and recognition as the Matthew Effect, where citations disproportionately favor established scientists, even when their contributions are no greater than those of junior peers. Merton's work, however, did not account for coauthored papers, where citations acknowledge teams rather than individual authors. How do teams affect reward systems in science? We hypothesize that teams will divide and obscure intellectual credit, making it even harder to recognize individual contributions. To test this, we developed and analyzed the world's first large-scale observational dataset on author contributions, derived from LaTeX source files of 1.6 million papers authored by 2 million scientists. We also quantified individual credits within teams using a validated algorithm and examined their relationship to contributions, accounting for factors such as team size, career stage, and historical time. Our findings confirm that teams amplify the Matthew Effect and overshadow individual contributions. As scientific research shifts from individual efforts to collaborative teamwork, this study highlights the urgent need for effective credit assignment practices in team-based science.

cs.SI↗

arXiv2024

Uncovering LLM-Generated Code: A Zero-Shot Synthetic Code Detector via Code Rewriting

Large Language Models (LLMs) have demonstrated remarkable proficiency in generating code. However, the misuse of LLM-generated (synthetic) code has raised concerns in both educational and industrial contexts, underscoring the urgent need for synthetic code detectors. Existing methods for detecting synthetic content are primarily designed for general text and struggle with code due to the unique grammatical structure of programming languages and the presence of numerous ''low-entropy'' tokens. Building on this, our work proposes a novel zero-shot synthetic code detector based on the similarity between the original code and its LLM-rewritten variants. Our method is based on the observation that differences between LLM-rewritten and original code tend to be smaller when the original code is synthetic. We utilize self-supervised contrastive learning to train a code similarity model and evaluate our approach on two synthetic code detection benchmarks. Our results demonstrate a significant improvement over existing SOTA synthetic content detectors, with AUROC scores increasing by 20.5% on the APPS benchmark and 29.1% on the MBPP benchmark.

cs.SE↗

arXiv2024

Displacing Science

Recent research on the decline in the paper disruption index (D-index) has sparked heated debates among scholars and garnered significant attention from policymakers and research institution leaders globally. To bridge the gap between policymakers' interest and scholars' skepticism about the D-index, we present this article summarizing key insights from our eight-year investigation, including interviews with scientists across nine disciplines and an analysis of 41 million papers over six decades. Our work confirms the decline in disruptive papers, addresses relevant technical concerns, and makes several original contributions: we clarify that the D-index measures how new ideas render old ones obsolete, suggesting 'Displacing' as an alternative interpretation for 'D'; we show that federal funding agencies like the NIH and NSF are less likely to support disruptive research; and we introduce the 'principle of functional equivalence' to explain the origins of recombining and displacing mechanisms in science, stressing that not all innovation problems are combinatorial, and challenging the belief that AI is the mature solution to scientific innovation. This article aims to promote broader and more accurate use of the D-index in research evaluation and to inspire new funding mechanisms for scientific breakthroughs.

cs.DL↗

arXiv2024

FAC$^2$E: Better Understanding Large Language Model Capabilities by Dissociating Language and Cognition

Large language models (LLMs) are primarily evaluated by overall performance on various text understanding and generation tasks. However, such a paradigm fails to comprehensively differentiate the fine-grained language and cognitive skills, rendering the lack of sufficient interpretation to LLMs' capabilities. In this paper, we present FAC$^2$E, a framework for Fine-grAined and Cognition-grounded LLMs' Capability Evaluation. Specifically, we formulate LLMs' evaluation in a multi-dimensional and explainable manner by dissociating the language-related capabilities and the cognition-related ones. Besides, through extracting the intermediate reasoning from LLMs, we further break down the process of applying a specific capability into three sub-steps: recalling relevant knowledge, utilizing knowledge, and solving problems. Finally, FAC$^2$E evaluates each sub-step of each fine-grained capability, providing a two-faceted diagnosis for LLMs. Utilizing FAC$^2$E, we identify a common shortfall in knowledge utilization among models and propose a straightforward, knowledge-enhanced method to mitigate this issue. Our results not only showcase promising performance enhancements but also highlight a direction for future LLM advancements.

cs.CL↗

arXiv2024

RCAgent: Cloud Root Cause Analysis by Autonomous Agents with Tool-Augmented Large Language Models

Large language model (LLM) applications in cloud root cause analysis (RCA) have been actively explored recently. However, current methods are still reliant on manual workflow settings and do not unleash LLMs' decision-making and environment interaction capabilities. We present RCAgent, a tool-augmented LLM autonomous agent framework for practical and privacy-aware industrial RCA usage. Running on an internally deployed model rather than GPT families, RCAgent is capable of free-form data collection and comprehensive analysis with tools. Our framework combines a variety of enhancements, including a unique Self-Consistency for action trajectories, and a suite of methods for context management, stabilization, and importing domain knowledge. Our experiments show RCAgent's evident and consistent superiority over ReAct across all aspects of RCA -- predicting root causes, solutions, evidence, and responsibilities -- and tasks covered or uncovered by current rules, as validated by both automated metrics and human evaluations. Furthermore, RCAgent has already been integrated into the diagnosis and issue discovery workflow of the Real-time Compute Platform for Apache Flink of Alibaba Cloud.

cs.SE↗

arXiv2024

The Death of Renaissance Scientist

Scholars are often categorized into two types: hedgehogs (specialists), who focus on working within a specific research field, and foxes (generalists), who actively contribute to a variety of fields. Despite the familiar anecdotes and popularity of this distinction, its empirical foundation has remained largely unexamined. We examine whether the research style of being a fox or a hedgehog is a stable personal trait or an evolving strategy over a scientist's career. Analyzing 2.3 million scholars' publication records over a century, we find that research styles exhibit remarkable stability. Notably, the proportion of fox-like scientists has dramatically declined in the past century, a phenomenon we term "the death of Renaissance scientists." This decline is particularly significant as science shifts toward team collaboration. Teams of foxes consistently outperform teams of hedgehogs in generating new ideas and directions, as confirmed by two emerging innovation metrics for papers: atypicality and disruption. Our research is the first to quantify the process and consequences of the decline of Renaissance scientists. By doing so, we establish a universal link between research styles, demographic shifts, and innovative output.

cs.SI↗

arXiv2024

Ask Questions with Double Hints: Visual Question Generation with Answer-awareness and Region-reference

The visual question generation (VQG) task aims to generate human-like questions from an image and potentially other side information (e.g. answer type). Previous works on VQG fall in two aspects: i) They suffer from one image to many questions mapping problem, which leads to the failure of generating referential and meaningful questions from an image. ii) They fail to model complex implicit relations among the visual objects in an image and also overlook potential interactions between the side information and image. To address these limitations, we first propose a novel learning paradigm to generate visual questions with answer-awareness and region-reference. Concretely, we aim to ask the right visual questions with Double Hints - textual answers and visual regions of interests, which could effectively mitigate the existing one-to-many mapping issue. Particularly, we develop a simple methodology to self-learn the visual hints without introducing any additional human annotations. Furthermore, to capture these sophisticated relationships, we propose a new double-hints guided Graph-to-Sequence learning framework, which first models them as a dynamic graph and learns the implicit topology end-to-end, and then utilizes a graph-to-sequence model to generate the questions with double hints. Experimental results demonstrate the priority of our proposed method.

cs.CV↗

arXiv2024

Tram: A Token-level Retrieval-augmented Mechanism for Source Code Summarization

Automatically generating human-readable text describing the functionality of a program is the intent of source code summarization. Although neural language models achieve significant performance in this field, they are limited by their inability to access external knowledge. To address this limitation, an emerging trend is combining neural models with external knowledge through retrieval methods. Previous methods have relied on the sentence-level retrieval paradigm on the encoder side. However, this paradigm is coarse-grained, noise-filled and cannot directly take advantage of the high-quality retrieved summary tokens on the decoder side. In this paper, we propose a fine-grained Token-level retrieval-augmented mechanism (Tram) on the decoder side rather than the encoder side to enhance the performance of neural models and produce more low-frequency tokens in generating summaries. Furthermore, to overcome the challenge of token-level retrieval in capturing contextual code semantics, we also propose integrating code semantics into individual summary tokens. The results of extensive experiments and human evaluation show that our token-level retrieval-augmented approach significantly improves performance and is more interpretable.

cs.AI↗

arXiv2024

AdaCCD: Adaptive Semantic Contrasts Discovery Based Cross Lingual Adaptation for Code Clone Detection

Code Clone Detection, which aims to retrieve functionally similar programs from large code bases, has been attracting increasing attention. Modern software often involves a diverse range of programming languages. However, current code clone detection methods are generally limited to only a few popular programming languages due to insufficient annotated data as well as their own model design constraints. To address these issues, we present AdaCCD, a novel cross-lingual adaptation method that can detect cloned codes in a new language without annotations in that language. AdaCCD leverages language-agnostic code representations from pre-trained programming language models and propose an Adaptively Refined Contrastive Learning framework to transfer knowledge from resource-rich languages to resource-poor languages. We evaluate the cross-lingual adaptation results of AdaCCD by constructing a multilingual code clone detection benchmark consisting of 5 programming languages. AdaCCD achieves significant improvements over other baselines, and achieve comparable performance to supervised fine-tuning.

cs.SE↗

arXiv2024

KG-TREAT: Pre-training for Treatment Effect Estimation by Synergizing Patient Data with Knowledge Graphs

Treatment effect estimation (TEE) is the task of determining the impact of various treatments on patient outcomes. Current TEE methods fall short due to reliance on limited labeled data and challenges posed by sparse and high-dimensional observational patient data. To address the challenges, we introduce a novel pre-training and fine-tuning framework, KG-TREAT, which synergizes large-scale observational patient data with biomedical knowledge graphs (KGs) to enhance TEE. Unlike previous approaches, KG-TREAT constructs dual-focus KGs and integrates a deep bi-level attention synergy method for in-depth information fusion, enabling distinct encoding of treatment-covariate and outcome-covariate relationships. KG-TREAT also incorporates two pre-training tasks to ensure a thorough grounding and contextualization of patient data and KGs. Evaluation on four downstream TEE tasks shows KG-TREAT's superiority over existing methods, with an average improvement of 7% in Area under the ROC Curve (AUC) and 9% in Influence Function-based Precision of Estimating Heterogeneous Effects (IF-PEHE). The effectiveness of our estimated treatment effects is further affirmed by alignment with established randomized clinical trial findings.

cs.LG↗

arXiv2023

Heterogeneous Graph Neural Networks using Self-supervised Reciprocally Contrastive Learning

Heterogeneous graph neural network (HGNN) is a very popular technique for the modeling and analysis of heterogeneous graphs. Most existing HGNN-based approaches are supervised or semi-supervised learning methods requiring graphs to be annotated, which is costly and time-consuming. Self-supervised contrastive learning has been proposed to address the problem of requiring annotated data by mining intrinsic information hidden within the given data. However, the existing contrastive learning methods are inadequate for heterogeneous graphs because they construct contrastive views only based on data perturbation or pre-defined structural properties (e.g., meta-path) in graph data while ignore the noises that may exist in both node attributes and graph topologies. We develop for the first time a novel and robust heterogeneous graph contrastive learning approach, namely HGCL, which introduces two views on respective guidance of node attributes and graph topologies and integrates and enhances them by reciprocally contrastive mechanism to better model heterogeneous graphs. In this new approach, we adopt distinct but most suitable attribute and topology fusion mechanisms in the two views, which are conducive to mining relevant information in attributes and topologies separately. We further use both attribute similarity and topological correlation to construct high-quality contrastive samples. Extensive experiments on three large real-world heterogeneous graphs demonstrate the superiority and robustness of HGCL over state-of-the-art methods.

cs.LG↗

arXiv2023

Remote Collaboration Fuses Fewer Breakthrough Ideas

Theories of innovation emphasize the role of social networks and teams as facilitators of breakthrough discoveries. Around the world, scientists and inventors today are more plentiful and interconnected than ever before. But while there are more people making discoveries, and more ideas that can be reconfigured in novel ways, research suggests that new ideas are getting harder to find-contradicting recombinant growth theory. In this paper, we shed new light on this apparent puzzle. Analyzing 20 million research articles and 4 million patent applications across the globe over the past half-century, we begin by documenting the rise of remote collaboration across cities, underlining the growing interconnectedness of scientists and inventors globally. We further show that across all fields, periods, and team sizes, researchers in these remote teams are consistently less likely to make breakthrough discoveries relative to their onsite counterparts. Creating a dataset that allows us to explore the division of labor in knowledge production within teams and across space, we find that among distributed team members, collaboration centers on late-stage, technical tasks involving more codified knowledge. Yet they are less likely to join forces in conceptual tasks-such as conceiving new ideas and designing research-when knowledge is tacit. We conclude that despite striking improvements in digital technology in recent years, remote teams are less likely to integrate the knowledge of their members to produce new, disruptive ideas.

cs.CY↗

arXiv2023

CP-BCS: Binary Code Summarization Guided by Control Flow Graph and Pseudo Code

Automatically generating function summaries for binaries is an extremely valuable but challenging task, since it involves translating the execution behavior and semantics of the low-level language (assembly code) into human-readable natural language. However, most current works on understanding assembly code are oriented towards generating function names, which involve numerous abbreviations that make them still confusing. To bridge this gap, we focus on generating complete summaries for binary functions, especially for stripped binary (no symbol table and debug information in reality). To fully exploit the semantics of assembly code, we present a control flow graph and pseudo code guided binary code summarization framework called CP-BCS. CP-BCS utilizes a bidirectional instruction-level control flow graph and pseudo code that incorporates expert knowledge to learn the comprehensive binary function execution behavior and logic semantics. We evaluate CP-BCS on 3 different binary optimization levels (O1, O2, and O3) for 3 different computer architectures (X86, X64, and ARM). The evaluation results demonstrate CP-BCS is superior and significantly improves the efficiency of reverse engineering.

cs.PL↗

arXiv2023

Small Teams Propel Fresh Ideas in Science and Technology

The past half-century has seen a dramatic increase in the scale and complexity of scientific research, to which researchers have responded by dedicating more time to education and training, narrowing their areas of specialization, and collaborating in larger teams. A widely held view is that such collaborations, by fostering specialization and encouraging novel combinations of ideas, accelerate scientific innovation. However, recent research challenges this notion, suggesting that small teams and solo researchers consistently disrupt science and technology with fresh ideas and opportunities, while larger teams tend to refine existing ones (Wu et al. 2019). This study, along with other relevant research, has garnered attention for challenging the zeitgeist of our time that views collaboration as the inevitable path forward in scientific and technological advancement. Yet, few studies have re-evaluated its central finding: the innovative advantage of small teams over large ones, using alternative measures. We explore innovation by identifying papers proposing new scientific concepts and patents introducing new technology codes. We analyzed 88 million research articles spanning from 1800 to 2020 and 7 million patent applications from 1976 to 2020 worldwide. Our findings confirm that while large teams contribute to development, small teams play a critical role in innovation by propelling fresh, original ideas in science and technology.

cs.DL↗

arXiv2023

Bridging the Gap between Spatial and Spectral Domains: A Unified Framework for Graph Neural Networks

Deep learning's performance has been extensively recognized recently. Graph neural networks (GNNs) are designed to deal with graph-structural data that classical deep learning does not easily manage. Since most GNNs were created using distinct theories, direct comparisons are impossible. Prior research has primarily concentrated on categorizing existing models, with little attention paid to their intrinsic connections. The purpose of this study is to establish a unified framework that integrates GNNs based on spectral graph and approximation theory. The framework incorporates a strong integration between spatial- and spectral-based GNNs while tightly associating approaches that exist within each respective domain.

cs.LG↗

arXiv2023

SkillQG: Learning to Generate Question for Reading Comprehension Assessment

We present $\textbf{$\texttt{SkillQG}$}$: a question generation framework with controllable comprehension types for assessing and improving machine reading comprehension models. Existing question generation systems widely differentiate questions by $\textit{literal}$ information such as question words and answer types to generate semantically relevant questions for a given context. However, they rarely consider the $\textit{comprehension}$ nature of questions, i.e. the different comprehension capabilities embodied by different questions. In comparison, our $\texttt{SkillQG}$ is able to tailor a fine-grained assessment and improvement to the capabilities of question answering models built on it. Specifically, we first frame the comprehension type of questions based on a hierarchical skill-based schema, then formulate $\texttt{SkillQG}$ as a skill-conditioned question generator. Furthermore, to improve the controllability of generation, we augment the input text with question focus and skill-specific knowledge, which are constructed by iteratively prompting the pre-trained language models. Empirical results demonstrate that $\texttt{SkillQG}$ outperforms baselines in terms of quality, relevance, and skill-controllability while showing a promising performance boost in downstream question answering task.

cs.CL↗

arXiv2023

Toward Subgraph-Guided Knowledge Graph Question Generation with Graph Neural Networks

Knowledge graph (KG) question generation (QG) aims to generate natural language questions from KGs and target answers. Previous works mostly focus on a simple setting which is to generate questions from a single KG triple. In this work, we focus on a more realistic setting where we aim to generate questions from a KG subgraph and target answers. In addition, most of previous works built on either RNN-based or Transformer based models to encode a linearized KG sugraph, which totally discards the explicit structure information of a KG subgraph. To address this issue, we propose to apply a bidirectional Graph2Seq model to encode the KG subgraph. Furthermore, we enhance our RNN decoder with node-level copying mechanism to allow directly copying node attributes from the KG subgraph to the output question. Both automatic and human evaluation results demonstrate that our model achieves new state-of-the-art scores, outperforming existing methods by a significant margin on two QG benchmarks. Experimental results also show that our QG model can consistently benefit the Question Answering (QA) task as a mean of data augmentation.

cs.CL↗