An Event-based Fast Intensity Reconstruction Scheme for UAV Real-time Perception

Daochun Li; Jinwu Xiang; Xin Dong; Yangjie Cui; Yiwei Zhang; Zhan Tu

REVIEW 3 major objections 2 minor 70 references

An Event-based Fast Intensity Reconstruction Scheme for UAV Real-time Perception

T0 review · 3 major / 2 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that one integration of an event stream, plus an enhanced decay step, reconstructs intensity images at 100 FPS with a small enough computation load to run onboard a UAV and keep visual tracking alive at 2–10 lux.

desk verdict The submitted artifact contains none of the ESI paper—just an unrelated repository-mining manuscript—so the claims are unverifiable and the submission should be returned, not reviewed. read the letter →

arxiv 2508.02238 v1 pith:HFW2BNVK submitted 2025-08-04 cs.CV cs.RO

Xin Dong , Yiwei Zhang , Yangjie Cui , Jinwu Xiang , Daochun Li , Zhan Tu This is my paper · ORCID

classification cs.CVcs.RO

keywords eventcameraintensityreconstructionUAVperceptionvisualtrackinglowilluminationreal-timevisionevent-basedsingleintegrationonboardcomputation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes ESI, event-based single integration, as a minimal recipe for turning an asynchronous event stream into ordinary intensity images: integrate the events once and apply an enhanced decay algorithm. The claim is that this single step is enough to produce usable images in real time at roughly 100 frames per second, with a computation load light enough to run on a UAV. If true, it makes established frame-based computer vision methods portable to event cameras without redesigning them, while keeping event cameras' strengths of wide dynamic range, high temporal resolution, and motion-blur immunity. The reported payoff is that a UAV equipped with ESI can track a target under 2–10 lux illumination, conditions where the comparison algorithms fail because their frame rate, image quality, or real-time performance is insufficient.

What carries the argument

The central object is ESI: one pass over the event stream that accumulates event contributions into an intensity image, plus an enhanced decay algorithm that prevents the accumulation from saturating or going stale. The decay step is what makes a single integration self-sufficient, since it lets old events fade so the reconstructed frame tracks the current scene instead of blurring into an average.

What would settle it

Run ESI and a conventional full reconstruction method on the same event stream from a fast-moving low-light scene; if a standard tracker fed ESI frames fails at 2–10 lux in a flight test where the paper says it succeeds, the central claim is falsified. A more surgical test is to feed the same tracker only every third ESI frame: the claim that real-time 100 FPS reconstruction is what enables tracking predicts a clear drop in tracking success at the reduced rate.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central discovery is that a full reconstruction pipeline—multi-step integration, learning-based inference, or iterative optimization—is unnecessary for downstream perception. A single integration of the event stream, combined with a decay algorithm that refreshes or resets older contributions, yields intensity images whose quality is high enough for onboard visual tracking. The method is evaluated against state-of-the-art reconstruction algorithms and reported to be faster, with better reconstruction quality and a 100 FPS output rate, and it is the only one among the compared methods that keeps working in UAV in-flight tests at 2–10 lux.

Load-bearing premise

The load-bearing premise is that one integration of the event stream, with an enhanced decay algorithm, retains enough photometric information for a downstream tracker, and that the comparison algorithms were exercised in configurations comparable to the paper's own.

Editorial extensions

If this is right

Conventional frame-based trackers and detectors can be reused on event cameras by feeding them ESI output, because ESI outputs standard intensity frames.
Onboard event-camera systems can run at 100 FPS with low computation, enabling real-time perception on power-limited platforms such as UAVs.
Event cameras' advantages—wide dynamic range, no motion blur, high temporal resolution—become usable in low-light settings, such as 2–10 lux, where frame-based cameras struggle.
Reconstruction cost no longer scales with model complexity if a single integration suffices, simplifying onboard deployment and freeing compute for downstream tasks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial caution: the supplied full text of this submission is a different paper's manuscript, so the claims above are drawn from the abstract alone; the method's derivations, ablations, and baseline configurations still need to be verified in the actual paper.
Editorial inference: if single integration plus decay is genuinely sufficient across scene types, then the field's trend toward larger learned reconstruction networks may be over-engineered for robotics tasks that only need downstream task-relevant appearance.
Editorial inference: the enhanced decay algorithm is the part most likely to transfer to other event-based methods; testing ESI's decay rule inside existing reconstruction pipelines would isolate its contribution.
Editorial inference: a direct comparison of tracking performance at 100 FPS versus a temporally decimated ESI stream would separate the value of high frame rate from the value of reconstruction quality in the low-light flight results.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague, the paper you sent over doesn't exist in the artifact. The abstract describes ESI, a low-compute event-based intensity reconstruction method for UAV onboard perception. That is a real problem worth working on: event cameras give high temporal resolution, but you need a cheap way to convert events into intensity-like images so existing frame-based trackers work. Single integration plus a decay term is a plausible starting point, and the 100 FPS claim with 2–10 lux flight tests would be interesting if substantiated. But the full text is an entirely different manuscript about LLM-based mining of software repositories. Not a related paper, not a corrupted draft—a different arXiv paper (2508.02233) with different authors. So none of the supporting material is present: no update equations, no decay formulation, no metrics, no hardware details, no ablation, no comparison baselines. The abstract's claims about 'remarkable runtime efficiency improvements' and 'superior reconstruction quality' are just assertions without data.

Where does that leave us? The idea has face validity. The soft spot is not the method; it is that the submission is unverifiable in the supplied form. I cannot check whether the enhanced decay has hand-tuned parameters, whether the comparisons were fair, or whether the 2–10 lux claim holds. The reader's report is correctly UNVERDICTED.

My recommendation: this should be returned to the authors, not sent to referees. A referee cannot review a ghost. If a correct manuscript surfaces, it deserves a look: the problem is well-motivated and the approach is simple enough to be evaluated rigorously. But as submitted, there is no paper here. I would desk-reject the artifact and ask for the actual PDF. If the authors supply it, I would send it out. If they do not, that ends the matter.

Referee Report

3 major / 2 minor

Summary. The paper claims to introduce ESI (event-based single integration), a scheme that reconstructs intensity images from event streams by a single integration combined with an 'enhanced decay algorithm'. The abstract asserts real-time reconstruction at 100 FPS, low computation load suitable for UAV onboard deployment, superior reconstruction quality and runtime efficiency relative to state-of-the-art algorithms, and successful UAV visual tracking under 2-10 lux conditions where comparative methods fail. The submitted full text, however, is a completely different manuscript—'A Methodological Framework for LLM-Based Mining of Software Repositories' (arXiv:2508.02233)—containing no mention of event cameras, intensity reconstruction, decay algorithms, UAV experiments, or any of the technical content needed to support the abstract's claims. Consequently, the central claims are unsupported by the inspectable material; there are no derivations, equations, experimental protocols, ablation studies, error bars, or comparisons available for review.

Significance. If the method worked as described, ESI would be a practically relevant contribution to event-based vision, because it promises to port conventional frame-based methods to event cameras with minimal computational overhead, enabling onboard perception in low-light conditions. The abstract's claims—single-integration reconstruction, 100 FPS throughput, and success at 2-10 lux—are falsifiable and would be significant if verified. However, the submitted artifact provides no technical basis for evaluating these claims. There is no algorithmic description, no analysis of the decay process, no photometric fidelity argument, no hardware implementation details, and no experimental data. The paper therefore cannot be assessed for correctness, novelty, or completeness in its current form.

major comments (3)

[Full text (all sections)] The supplied full text is not the paper described in the abstract. It is 'A Methodological Framework for LLM-Based Mining of Software Repositories' (arXiv:2508.02233), which addresses empirical software engineering and contains no reference to event cameras, intensity reconstruction, the 'enhanced decay algorithm', 100 FPS reconstruction, UAV experiments, or the 2-10 lux tracking evaluations. Because every load-bearing technical element of the claimed contribution is absent, the central claims of the abstract are unsupported by any inspectable evidence. This is not a minor formatting issue; it makes review of the method, experiments, and comparisons impossible.
[Abstract (method description)] The abstract introduces 'a single integration of the event streams combined with an enhanced decay algorithm' but provides no equation, pseudo-code, or specification of the decay model. Consequently, it is impossible to audit key questions that would normally determine soundness: whether the decay parameters are hand-tuned or fitted, whether the reconstruction preserves photometric information needed for downstream tracking, and whether the 'enhanced' component is an ad hoc fix rather than a principled mechanism. The reader's concern about free parameters and potential circularity therefore cannot be resolved from the submitted material.
[Abstract (comparisons and experiments)] The abstract reports 'extensive experiments', 'performance comparison of ESI and state-of-the-art algorithms', and 'in-flight tests' demonstrating UAV tracking at 2-10 lux with failure of comparative algorithms. None of these experiments, evaluation metrics, baselines, hardware details, or statistical results appear in the full text. Without the experimental protocol and error bars, the claims of 'remarkable runtime efficiency improvements' and 'superior reconstruction quality' are unverifiable. A reader cannot tell whether the comparisons used fair configurations, how many trials were run, or whether the 2-10 lux conditions were consistently controlled.

minor comments (2)

[Abstract] There are minor typographical issues, such as the missing space in 'conditions(2-10lux)' and the inconsistent hyphenation of 'state-of-the-art'. These are trivial relative to the content problems, but they do need cleanup in any revised submission.
[General] The paper would benefit from a clear statement of the relationship between the event-based intensity reconstruction scheme and the existing literature on event-to-frame conversion, such as references to prior reconstruction methods (e.g., E2VID, FireNet, or other learning-based approaches). No such references appear in the submitted material.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity can be substantiated: the supplied full text is a different arXiv paper, and the ESI derivation is absent.

full rationale

The abstract claims that ESI reconstructs intensity by a single integration of event streams plus an enhanced decay algorithm, but the attached full text is 'A Methodological Framework for LLM-Based Mining of Software Repositories' (arXiv:2508.02233), not the event-camera paper. No ESI update equations, decay definition, hyperparameter fitting procedure, or reconstruction protocol are present, so there is no derivation chain whose steps could be shown to reduce to the method's own inputs. The absence of the technical artifact prevents any quotation-based demonstration of a self-definitional, fitted-input, or self-citation circularity. While the mismatched full text is a serious completeness and verifiability problem, it is not itself evidence of circularity, and the rules require exhibiting a specific reduction rather than speculating about the 'enhanced decay algorithm'. Accordingly, no circular steps are identified and the circularity score is 0.

Assumptions & free parameters 1 free parameters · 2 assumptions · 0 invented entities

From the abstract alone only two background assumptions can be identified; the decay algorithm may contain free parameters, but no values are given. The full-text mismatch means the actual ledger of the ESI paper could not be audited.

free parameters (1)

Enhanced decay parameters
The abstract mentions an 'enhanced decay algorithm' inside ESI, but the decay rate constants, thresholds, or schedules are not stated. Whether these are fixed, hand-tuned, or fitted to data cannot be determined from the abstract.

assumptions (2)

domain assumption Event cameras emit events proportional to temporal contrast of log-brightness, and integrating events approximates scene intensity.
Standard event-camera measurement model underlying all integration-based reconstruction; it is not derived or stated in the abstract but is required for ESI's operation.
domain assumption The benchmark datasets and in-flight test scenes are representative of the target UAV operating conditions.
The abstract claims superiority under 2-10 lux, but the selection of test scenes, lighting, and motion profiles is not described, so the evaluation's generality is assumed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of An Event-based Fast Intensity Reconstruction Scheme for UAV Real-time Perception." pith.science (2026). https://pith.science/paper/HFW2BNVK

@misc{pith2026250802238,
  author       = {Pith},
  title        = {Pith review of: An Event-based Fast Intensity Reconstruction Scheme for UAV Real-time Perception},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HFW2BNVK}},
  note         = {Machine review of arXiv:2508.02238}
}

read the original abstract

Event cameras offer significant advantages, including a wide dynamic range, high temporal resolution, and immunity to motion blur, making them highly promising for addressing challenging visual conditions. Extracting and utilizing effective information from asynchronous event streams is essential for the onboard implementation of event cameras. In this paper, we propose a streamlined event-based intensity reconstruction scheme, event-based single integration (ESI), to address such implementation challenges. This method guarantees the portability of conventional frame-based vision methods to event-based scenarios and maintains the intrinsic advantages of event cameras. The ESI approach reconstructs intensity images by performing a single integration of the event streams combined with an enhanced decay algorithm. Such a method enables real-time intensity reconstruction at a high frame rate, typically 100 FPS. Furthermore, the relatively low computation load of ESI fits onboard implementation suitably, such as in UAV-based visual tracking scenarios. Extensive experiments have been conducted to evaluate the performance comparison of ESI and state-of-the-art algorithms. Compared to state-of-the-art algorithms, ESI demonstrates remarkable runtime efficiency improvements, superior reconstruction quality, and a high frame rate. As a result, ESI enhances UAV onboard perception significantly under visual adversary surroundings. In-flight tests, ESI demonstrates effective performance for UAV onboard visual tracking under extremely low illumination conditions(2-10lux), whereas other comparative algorithms fail due to insufficient frame rate, poor image quality, or limited real-time performance.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

70 extracted references · 69 canonical work pages

[1]

Samuel Abedu, Laurine Menneron, SayedHassan Khatoonabadi, and Emad Shihab. 2025. RepoChat: An LLM-Powered Chatbot for GitHub Repository Question-Answering. In �� . IEEE, 255–259. Manuscript submitted to ACM A Methodological Framework for LLM-Based Mining of Software Repositories 43

work page 2025
[2]

Toufique Ahmed, Premkumar Devanbu, Christoph Treude, and Michael Pradel. 2025. Can llms replace manual annotation of software engineering artifacts?. In �� . IEEE, 526–538

work page 2025
[3]

Dharun Anandayuvaraj, Matthew Campbell, Arav Tewari, and James C Davis. 2024. FAIL: Analyzing Software Failures from the News Using LLMs. In �� . 506–518

work page 2024
[4]

Sebastian Baltes and Paul Ralph. 2022. Sampling in software engineering research: A critical review and guidelines. �� 27, 4 (2022), 94

work page 2022
[5]

Abdul Ali Bangash, Hareem Sahar, Shaiful Chowdhury, Alexander William Wong, Abram Hindle, and Karim Ali. 2019. What do developers know about machine learning: a study of ml discussions on stackoverflow. In �� . IEEE, 260–264

work page 2019
[6]

Cauã Ferreira Barros, Bruna Borges Azevedo, Valdemar Vicente Graciano Neto, Mohamad Kassab, Marcos Kalinowski, Hugo Alexandre D Do Nasci- mento, and Michelle CGSP Bandeira. 2025. Large Language Model for Qualitative Research: A Systematic Mapping Study. In �� ...

work page 2025
[7]

João Helis Bernardo, Daniel Alencar Da Costa, Sérgio Queiroz de Medeiros, and Uirá Kulesza. 2024. How do machine learning projects use continuous integration practices? an empirical study on GitHub actions. In �� . 665–676

work page 2024
[8]

Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. �� 33 (2020), 1877–1901

work page 2020

Show all 70 references

[9]

Victor R Basili1 Gianluigi Caldiera and H Dieter Rombach. 1994. The goal question metric approach. �� (1994), 528–532

1994
[10]

Fabio Calefato, Filippo Lanubile, and Luigi Quaranta. 2022. A preliminary investigation of MLOps practices in GitHub. In �� . 283–288

2022
[11]

Bruno Cartaxo, Gustavo Pinto, and Sergio Soares. 2020. Rapid reviews in software engineering. �� (2020), 357–384

2020
[12]

Joel Castaño, Rafael Cabañas, Antonio Salmerón, David Lo, and Silverio Martínez-Fernández. 2024. How do machine learning models change? �� (2024)

2024
[13]

Joel Castaño, Silverio Martínez-Fernández, Xavier Franch, and Justus Bogner. 2023. Exploring the carbon footprint of hugging face’s ml models: A repository mining study. In �� . IEEE, 1–12

2023
[14]

Preetha Chatterjee, Mia Mohammad Imran, and Kostadin Damevski. 2024. Shedding Light on Software Engineering-specific Metaphors and Idioms. In �� . Association for Computing Machinery, 1–13

2024
[15]

Zhenpeng Chen, Yanbin Cao, Yuanqiang Liu, Haoyu Wang, Tao Xie, and Xuanzhe Liu. 2020. A comprehensive study on challenges in deploying deep learning based software. In �� ...

2020
[16]

Jacob Cohen. 1960. A coefficient of agreement for nominal scales. �� 20, 1 (1960), 37–46

1960
[17]

Giuseppe Colavito, Filippo Lanubile, and Nicole Novielli. 2025. Benchmarking large language models for automated labeling: The case of issue report classification. �� (2025), 107758

2025
[18]

Marcelo Costalonga, Bianca Minetto Napoleão, Maria Teresa Baldassarre, Katia Romero Felizardo, Igor Steinmacher, and Marcos Kalinowski. 2025. Can Machine Learning Support the Selection of Studies for Systematic Literature Review Updates? �� (2025)

2025
[19]

Daniela S Cruzes and Tore Dybå. 2011. Recommended steps for thematic synthesis in software engineering. �� (2011), 275–284

2011
[20]

Vincenzo De Martino, Joel Castaño, Fabio Palomba, Xavier Franch, and Silverio Martínez-Fernández. 2025. A framework for using llms for repository mining studies in empirical software engineering. In �� ...

2025
[21]

Vincenzo De Martino, Silverio Martínez-Fernández, and Fabio Palomba. 2024. Do developers adopt green architectural tactics for ml-enabled systems? a mining software repository study. �� (2024)

2024
[22]

Vincenzo De Martino and Fabio Palomba. 2025. Classification and challenges of non-functional requirements in ML-enabled systems: A systematic literature review. �� (2025), 107678

2025
[23]

Vincenzo De Martino, Gilberto Recupito, Giammaria Giordano, Filomena Ferrucci, Dario Di Nucci, and Fabio Palomba. 2025. Into the ML- Universe: An improved classification and characterization of machine-learning projects. �� 230 (2025), 112471. https...

2025
[24]

Don A Dillman, Jolene D Smyth, and Leah Melani Christian. 2014. �� . John Wiley & Sons

2014
[25]

Yihong Dong, Xue Jiang, Zhi Jin, and Ge Li. 2024. Self-collaboration code generation via chatgpt. �� 33, 7 (2024), 1–38

2024
[26]

Angela Fan, Beliz Gokkaya, Mark Harman, Mitya Lyubarskiy, Shubho Sengupta, Shin Yoo, and Jie M Zhang. 2023. Large language models for software engineering: Survey and open problems. In �� ...

2023
[27]

Katia Romero Felizardo, Anderson Deizepe, Daniel Coutinho, Genildo Gomes, Maria Meireles, Marco Gerosa, and Igor Steinmacher. 2025. On the Difficulties of Conducting and Replicating Systematic Literature Reviews Studies Using LLMs in Software Engineering. In �� ...

2025
[28]

Alexandra González, Xavier Franch, David Lo, and Silverio Martínez-Fernández. 2025. How do Pre-Trained Models Support Software Engineering? An Empirical Study in Hugging Face. �� (2025)

2025
[29]

Joseph F Hair, Arthur H Money, Philip Samouel, and Mike Page. 2007. Research methods for business. �� 49, 4 (2007), 336–337

2007
[30]

Abram Hindle, Daniel M German, and Ric Holt. 2008. What do large commits tell us? a taxonomical study of large commits. In �� . 99–108

2008
[31]

Xinyi Hou, Yanjie Zhao, Yue Liu, Zhou Yang, Kailong Wang, Li Li, Xiapu Luo, David Lo, John Grundy, and Haoyu Wang. 2024. Large language models for software engineering: A systematic literature review. �� 33, 8 (2024), 1–79

2024
[32]

Mia Mohammad Imran, Preetha Chatterjee, and Kostadin Damevski. 2024. Uncovering the causes of emotions in software developer communication using zero-shot llms. In �� . 1–13

2024
[33]

Wenxin Jiang, Jerin Yasmin, Jason Jones, Nicholas Synovic, Jiashen Kuo, Nathaniel Bielanski, Yuan Tian, George K Thiruvathukal, and James C Davis. 2024. PeaTMOSS: A Dataset and Initial Analysis of Pre-Trained Models in Open-Source Software. In �� ...

2024
[34]

R Burke Johnson, Anthony J Onwuegbuzie, and Lisa A Turner. 2007. Toward a definition of mixed methods research. �� 1, 2 (2007), 112–133

2007
[35]

Minseo Kim, Taemin Kim, Thu Hoang Anh Vo, Yugyeong Jung, and Uichin Lee. 2025. Exploring Modular Prompt Design for Emotion and Mental Health Recognition. In �� . 1–18

2025
[36]

Barbara Kitchenham, O Pearl Brereton, David Budgen, Mark Turner, John Bailey, and Stephen Linkman. 2009. Systematic literature reviews in software engineering–a systematic literature review. �� 51, 1 (2009), 7–15

2009
[37]

Barbara A Kitchenham and Shari L Pfleeger. 2008. Personal Opinion Surveys. In �� . Springer, 63–92

2008
[38]

Klaus Krippendorff. 2018. �� (4 ed.). Sage publications

2018
[39]

Patricia Lago, Per Runeson, Qunying Song, and Roberto Verdecchia. 2024. Threats to validity in software engineering–hypocritical paper section or essential analysis?. In �� . 314–324

2024
[40]

Matheus De Morais Leça, Lucas Valença, Reydne Santos, and Ronnie De Souza Santos. 2025. Applications and Implications of Large Language Models in Qualitative Analysis: A New Frontier for Empirical Software Engineering. In �� ...

2025
[41]

Jia Li, Ge Li, Yongmin Li, and Zhi Jin. 2025. Structured chain-of-thought prompting for code generation. �� 34, 2 (2025), 1–23

2025
[42]

Jenny T Liang, Carmen Badea, Christian Bird, Robert DeLine, Denae Ford, Nicole Forsgren, and Thomas Zimmermann. 2024. Can gpt-4 replicate empirical software engineering research? �� 1, FSE (2024), 1330–1353

2024
[43]

Vincenzo De Martino, Joel Castaño, Fabio Palomba, Xavier Franch, and Silverio Martínez-Fernández. 2025. A Methodological Framework for LLM-Based Mining of Software Repositories. (7 2025). https://doi.org/10.6084/m9.figshare.29647169

2025 doi
[44]

Collin McMillan, Mario Linares-Vasquez, Denys Poshyvanyk, and Mark Grechanik. 2011. Categorizing software applications for maintenance. In �� . IEEE, 343–352

2011
[45]

Daye Nam, Andrew Macvean, Vincent Hellendoorn, Bogdan Vasilescu, and Brad Myers. 2024. Using an llm to help with code understanding. In �� . 1–13

2024
[46]

Tien Nguyen, Waris Gill, and Muhammad Ali Gulzar. 2025. Are the Majority of Public Computational Notebooks Pathologically Non-Executable? �� (2025)

2025
[47]

Dorin Pomian, Abhiram Bellur, Malinda Dilhara, Zarina Kurbatova, Egor Bogomolov, Timofey Bryksin, and Danny Dig. 2024. Next-generation refactoring: Combining llm insights and ide capabilities for extract method. In �� ...

2024
[48]

Ali Ebrahimi Pourasad and Walid Maalej. 2024. Does GenAI Make Usability Testing Obsolete? �� (2024)

2024
[49]

Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. �� 1, 8 (2019), 9

2019
[50]

Per Runeson and Martin Höst. 2009. Guidelines for conducting and reporting case study research in software engineering. �� 14, 2 (2009), 131–164

2009
[51]

June Sallou, Thomas Durieux, and Annibale Panichella. 2024. Breaking the silence: the threats of using llms in software engineering. In �� . 102–106

2024
[52]

Benjamin Saunders, Julius Sim, Tom Kingstone, Shula Baker, Jackie Waterfield, Bernadette Bartlam, Heather Burroughs, and Clare Jinks. 2018. Saturation in qualitative research: exploring its conceptualization and operationalization. �� 52 (2018), 1893–1907

2018
[53]

Md Shafikuzzaman, Md Rakibul Islam, Alex C Rolli, Sharmin Akhter, and Naeem Seliya. 2024. An Empirical Evaluation of the Zero-Shot, Few-Shot, and Traditional Fine-Tuning Based Pretrained Language Models for Sentiment Analysis in Software Engineering. �� (2024)

2024
[54]

Mohammad Sadegh Sheikhaei, Yuan Tian, Shaowei Wang, and Bowen Xu. 2024. An empirical study on the effectiveness of large language models for SATD identification and classification. �� 29, 6 (2024), 159. Manuscript submitted to ACM A Methodological Fr...

2024
[55]

Luciana Lourdes Silva, Janio Rosa da Silva, João Eduardo Montandon, Marcus Andrade, and Marco Tulio Valente. 2024. Detecting code smells using chatgpt: Initial insights. In �� ...

2024
[56]

Klaas-Jan Stol and Brian Fitzgerald. 2018. Grounded theory in software engineering research: a critical review and guidelines. �� 50, 4 (2018), 1–37

2018
[57]

Bianca Trinkenreich, Fabio Calefato, Geir Hanssen, Kelly Blincoe, Marcos Kalinowski, Mauro Pezzé, Paolo Tell, and Margaret-Anne Storey. 2025. Get on the Train or be Left on the Station: Using LLMs for Software Engineering Research. (2025)

2025
[58]

Michele Tufano, Fabio Palomba, Gabriele Bavota, Rocco Oliveto, Massimiliano Di Penta, Andrea De Lucia, and Denys Poshyvanyk. 2015. When and why your code starts to smell bad. In �� , Vol. 1. IEEE, 403–414

2015
[59]

Secil Ugurel, Robert Krovetz, and C Lee Giles. 2002. What’s the code? automatic classification of source code archives. In �� . 632–638

2002
[60]

Stefan Wagner, Marvin Muñoz Barón, Davide Falessi, and Sebastian Baltes. 2024. Towards Evaluation Guidelines for Empirical Studies involving LLMs. �� (2024)

2024
[61]

Ruiqi Wang, Jiyu Guo, Cuiyun Gao, Guodong Fan, Chun Yong Chong, and Xin Xia. 2025. Can llms replace human evaluators? an empirical study of llm-as-a-judge in software engineering. �� 2, ISSTA (2025), 1955–1977

2025
[62]

Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. �� 35 (2022), 24824–24837

2022
[63]

Claes Wohlin. 2014. Guidelines for Snowballing in Systematic Literature Studies and a Replication in Software Engineering. In �� (London, England, United Kingdom) �� . Asso...

2014
[64]

Claes Wohlin, Emilia Mendes, Katia Romero Felizardo, and Marcos Kalinowski. 2020. Guidelines for the search strategy to update systematic literature reviews in software engineering. �� 127 (2020), 106366

2020
[65]

Claes Wohlin, Per Runeson, Martin Höst, Magnus C Ohlsson, Björn Regnell, and Anders Wesslén. 2012. �� . Springer Science & Business Media

2012
[66]

Cong Wu, Jing Chen, Ziwei Wang, Ruichao Liang, and Ruiying Du. 2024. Semantic Sleuth: Identifying Ponzi Contracts via Large Language Models. In �� . 582–593

2024
[67]

Weiwei Xu, Kai Gao, Hao He, and Minghui Zhou. 2024. LiCoEval: Evaluating LLMs on License Compliance in Code Generation. �� (2024)

2024
[68]

Ting Zhang, Ivana Clairine Irsan, Ferdian Thung, and David Lo. 2025. Revisiting Sentiment Analysis for Software Engineering in the Era of Large Language Models. �� 34, 3 (2025), 1–30

2025
[69]

Yichi Zhang, Zixi Liu, Yang Feng, and Baowen Xu. 2024. Leveraging Large Language Model to Assist Detecting Rust Code Comment Inconsistency. In �� . 356–366

2024
[70]

Zibin Zheng, Kaiwen Ning, Qingyuan Zhong, Jiachi Chen, Wenqing Chen, Lianghong Guo, Weicheng Wang, and Yanlin Wang. 2025. Towards an understanding of large language models in software engineering tasks. �� 30, 2 (2025), 50. Manuscript submitted to ACM

2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Samuel Abedu, Laurine Menneron, SayedHassan Khatoonabadi, and Emad Shihab. 2025. RepoChat: An LLM-Powered Chatbot for GitHub Repository Question-Answering. In �� . IEEE, 255–259. Manuscript submitted to ACM A Methodological Framework for LLM-Based Mining of Software Repositories 43

work page 2025

[2] [2]

Toufique Ahmed, Premkumar Devanbu, Christoph Treude, and Michael Pradel. 2025. Can llms replace manual annotation of software engineering artifacts?. In �� . IEEE, 526–538

work page 2025

[3] [3]

Dharun Anandayuvaraj, Matthew Campbell, Arav Tewari, and James C Davis. 2024. FAIL: Analyzing Software Failures from the News Using LLMs. In �� . 506–518

work page 2024

[4] [4]

Sebastian Baltes and Paul Ralph. 2022. Sampling in software engineering research: A critical review and guidelines. �� 27, 4 (2022), 94

work page 2022

[5] [5]

Abdul Ali Bangash, Hareem Sahar, Shaiful Chowdhury, Alexander William Wong, Abram Hindle, and Karim Ali. 2019. What do developers know about machine learning: a study of ml discussions on stackoverflow. In �� . IEEE, 260–264

work page 2019

[6] [6]

Cauã Ferreira Barros, Bruna Borges Azevedo, Valdemar Vicente Graciano Neto, Mohamad Kassab, Marcos Kalinowski, Hugo Alexandre D Do Nasci- mento, and Michelle CGSP Bandeira. 2025. Large Language Model for Qualitative Research: A Systematic Mapping Study. In �� ...

work page 2025

[7] [7]

João Helis Bernardo, Daniel Alencar Da Costa, Sérgio Queiroz de Medeiros, and Uirá Kulesza. 2024. How do machine learning projects use continuous integration practices? an empirical study on GitHub actions. In �� . 665–676

work page 2024

[8] [8]

Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. �� 33 (2020), 1877–1901

work page 2020