Pith. sign in

REVIEW 3 major objections 3 minor 59 references

A static impact-analysis method that combines transformer code embeddings with program dependence graphs beats the strongest conceptual baseline by roughly 10 percentage points on a new 4,405-task benchmark.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-07-31 23:38 UTC pith:H374L43D

load-bearing objection Athena is a well-run empirical paper worth refereeing, but the exact 10-point margins need a held-out validation protocol and an explicit treatment of the co-change-as-impact ground-truth assumption. the 3 major comments →

arxiv 2607.23355 v1 pith:H374L43D submitted 2026-07-25 cs.SE

Enhancing Code Understanding for Impact Analysis by Combining Transformers and Program Dependence Graphs

Yanfu Yan , Nathan Cooper , Kevin Moran , Gabriele Bavota , Denys Poshyvanyk , Steve Rich This is my paper
classification cs.SE
keywords impact analysisconceptual couplingprogram dependence graphstransformer code modelsGraphCodeBERTembedding propagationbenchmark constructiontangled commits
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper tries to establish that impact analysis — predicting which methods in a codebase will need to change when a given method changes — can be done accurately without change histories or test executions, using only one snapshot of source code. The proposed method, Athena, takes embeddings of methods from a transformer-based code model and then propagates those embeddings over a program dependence graph, so each method's representation absorbs the semantics of the methods it calls and shares a class with. On a new benchmark built from 910 manually untangled bug-fix commits in 25 Java projects, the paper reports that Athena beats the best conceptual-coupling baseline (latent semantic indexing) by 10.34% in mean reciprocal rank, 9.55% in mean average precision, and 11.68% in HIT@10, with statistical significance. The claim matters because if correct, it gives developers a practical, history-free tool for a cognitively demanding maintenance task, and it shows that combining structural and semantic code understanding is more effective than either alone.

Core claim

On its own terms, Athena makes two interconnected claims. First, the paper claims that a static, history-free approach can achieve state-of-the-art impact-analysis accuracy: take method embeddings from a transformer code model fine-tuned on code search, propagate them over an undirected program dependence graph whose edges are call dependencies and class-member dependencies, and rank the corpus by cosine similarity to the query. Second, it claims this combination is what drives the gain: ablations show neither the transformer embeddings nor the graph propagation alone matches the full method, and the margin over the best conceptual baseline (latent semantic indexing) is 10.34% in mRR, 9.55%

What carries the argument

Embedding propagation over a program dependence graph is the load-bearing mechanism. A static parser builds an undirected graph whose nodes are methods and whose edges encode two relationships: call dependence (caller–callee) and class-member dependence (methods declared in the same class). Each node starts with an embedding formed by averaging the last-layer hidden states of a transformer code model (the paper's best configuration uses GraphCodeBERT, fine-tuned on code search). The update step is a weight-free graph convolution, M' = (I + w · D^{-1/2}(A_c + A_cm) D^{-1/2}) M, with w = 0.5, aggregating neighbors within two hops. This lets a method's final representation carry both its own lo

Load-bearing premise

The benchmark assumes that every method edited together in a manually untangled bug-fix commit is truly impacted by every other method in that commit; co-change in a bug fix is taken as evidence of causal impact, even though methods may be edited together just to address the same bug without one's change affecting the other.

What would settle it

Run Athena on Alexandria with the dependence graph's edges randomly shuffled while keeping node embeddings identical; if mRR and mAP stay close to the reported 60.32% and 35.19%, the graph structure is not responsible for the gain and the central mechanism claim collapses. A complementary check is to have independent developers judge a random subset of ground-truth pairs for genuine causal impact; low agreement would indict the benchmark's co-change-as-impact assumption.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • If Athena's result holds, impact analysis becomes feasible for codebases with no change history and no execution traces: only the current snapshot is needed, so the method applies to new projects from day one.
  • The reported margins imply that transformer-derived code semantics, which already power code search and clone detection, transfer to impact analysis and beat traditional IR representations (LSI, TF-IDF, doc2vec) by a wide margin.
  • The dependence-graph propagation contributes most when the impacted methods live outside the query's class — the hard case that dominates the whole-project setting.
  • Benchmark construction quality is not neutral: the paper's own comparison shows that tangled commits distort measured accuracy, so future IA evaluation should use untangled ground truth.
  • Because dependency information is extracted statically from one snapshot, the approach sidesteps the brittleness of evolutionary and dynamic IA while retaining the benefits of multiple information sources.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The class-member edge acts as an implicit same-class prior: the paper's own 'reduce cosine distance to same-class methods by 50%' experiment nearly matches Athena's whole-project gain without any graph machinery, suggesting part of the reported improvement may reflect a prior rather than genuinely new semantic understanding.
  • A decisive check of the mechanism: shuffle the dependence-graph edges and re-run Athena. If the mRR/mAP gains over LSI survive random graphs, then the specific call/class structure is not the source of the improvement; if they collapse, the graph information is doing the work claimed.
  • The same design — transformer embeddings plus unweighted graph propagation over static dependency edges — could plausibly transfer to other retrieval-style maintenance tasks such as feature location or test-impact analysis, which share the structure of ranking code entities by relevance to a seed.
  • The benchmark's untangling cost is high and language-specific; extending Alexandria to other languages would require similarly careful line-level annotation, which may limit how quickly the evaluation protocol can spread.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 3 minor

Summary. The paper proposes Athena, a method-level impact analysis (IA) technique that combines Transformer-based code embeddings (CodeBERT, GraphCodeBERT, UniXcoder) with static program dependence graphs through a GCN-inspired embedding propagation step. It also introduces Alexandria, a benchmark of 4,405 IA tasks derived from 910 manually untangled bug-fix commits across 25 Apache Java projects. On this benchmark, Athena reportedly achieves mRR/mAP/HIT@10 of 60.32/35.19/81.48 in the whole-corpus setting, improving over the best baseline (LSI) by 10.34/9.55/11.68 points with claimed statistical significance. The paper includes ablations over encoders, dependency types, propagation order, and fine-tuning objectives, as well as per-project results.

Significance. If the central claims hold, the paper provides a useful step forward: it is the first application of Transformer-based code models to IA, it makes a new benchmark that is substantially larger than prior ones, and it carefully ablates the contribution of structural propagation. The replication package and the use of manually untangled commits are concrete strengths. However, the headline margins rest on the construct validity of Alexandria's ground-truth labels and on the fairness of the comparison; both need to be strengthened before the quantitative conclusions can be accepted.

major comments (3)
  1. [Section 4.1] The ground-truth construction treats every method in a co-changed set as a query and the remaining methods as its impact set. This equates co-change with mutual causal impact. In a bug-fix commit, methods can be edited together as parallel consequences of the same bug (e.g., changing a shared API and its callers) without one method's change forcing the other. Symmetrization amplifies the issue. Since all models are scored on these labels, the reported 10.34/9.55/11.68 margins may measure co-change retrieval rather than IA as defined in Section 1. Manual untangling removes multi-concern commits but does not validate causality. Please validate a sample with developer judgments or an independent dependency oracle, or restrict to tasks with direct/transitive structural dependence; otherwise reframe the contribution as co-change retrieval.
  2. [Sections 4.4 and 5.1] LSI topic count (1,300), the propagation weight w=0.5, and the 2-hop propagation order were selected after observing results on Alexandria, with no held-out validation described. This makes the reported gains over LSI optimistic: the LSI configuration is tuned, and Athena's configuration is also tuned on the same test tasks. Please provide an evaluation protocol that separates configuration selection from reporting (e.g., per-project cross-validation), or a sensitivity analysis showing that the main conclusions hold across a range of topic counts, w, and propagation orders.
  3. [Section 5.1] The only description of statistical testing is 'Wilcoxon's paired test, p<0.05'. The unit of pairing is unspecified, and the 4,405 tasks are nested within 910 commits and 25 projects, so task-level tests likely inflate significance. Please report the pairing, the number of test units, and a project- or commit-level analysis (e.g., bootstrap or mixed-effects models), together with effect sizes.
minor comments (3)
  1. [Abstract and Section 1] The abstract says 'outperform a simpler baseline' while the conclusion says 'best-performing conceptual IA baseline'; make the reference consistent and precise.
  2. [Section 3.3, Eq. (2)] The definition of D is described as normalization 'with respect to both rows and columns' but the formula uses symmetric normalization D^{-1/2} A D^{-1/2}. Clarify the notation and the role of w.
  3. [Table 4] The header layout of Table 4 is confusing; the encoders, neighbor orders, and ablation variants should be separated into distinct rows or subheadings for readability.

Circularity Check

0 steps flagged

No significant circularity: Athena is an empirical system whose components are trained or computed independently of the Alexandria benchmark labels.

full rationale

The paper's central claim is an empirical comparison on a new benchmark, not a derivation that reduces to its own inputs. The method embeddings come from Transformer models (GraphCodeBERT, CodeBERT, UniXcoder) fine-tuned on the external CodeSearchNet Java split (Section 3.2 and Section 4.4), and the embedding propagation in Eqs. (2)-(3) is a fixed first-order graph filter with constant w=0.5 and no trainable parameters fitted to Alexandria; the graph edges are static call/class-member relations extracted from the parent-commit source. Cosine ranking (Section 3.4) is then applied to held-out co-changed sets. No parameter is fitted to the target benchmark and then reported as a prediction. The benchmark definition in Section 4.1 treats co-changed methods in manually untangled bug-fix commits as mutually impacted; this is a construct-validity assumption shared with prior IA benchmarks and the paper explicitly discusses tangling as a threat, but it is not a circular derivation because the co-change labels are not used as training signal and are not built into the similarity function or propagation equations. The only self-citations (Yan et al. 2024, 2026) point to the online appendix and archived replication package, and play no role in justifying the method's effectiveness. Therefore no circular step meeting the quoting-and-reduction standard is present.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 0 invented entities

The central claim rests on the benchmark's co-change-to-impact identification, the transfer from code-search fine-tuning, and the choice of graph propagation hyperparameters. None of these are independently validated outside the paper's own evaluation, which is why soundness is medium despite the clean writing.

free parameters (3)
  • w (propagation balancing weight) = 0.5
    Balances the method's own embedding against neighbor contributions in Eq. (2); set by hand in Section 4.4 without a held-out search, and the reported results depend on it.
  • Neighbor propagation order = 2 hops
    Orders 1, 2, and 3 were tried; 2 is reported as optimal in Section 5.3 on the same benchmark used for final scores.
  • LSI topic count = 1300
    The baseline LSI was tuned by varying topics from 0 to 2000 and selecting the best performer (1,300) on Alexandria in Section 5.1; this affects the baseline used to compute relative gains.
axioms (4)
  • domain assumption Co-changed methods in a manually-untangled bug-fix commit are mutually impacted.
    The central ground-truth assumption in Section 4.1; every evaluation score in the paper rests on it.
  • domain assumption Transformer embeddings fine-tuned on CodeSearchNet code search transfer to impact analysis.
    Used to justify the fine-tuning proxy in Section 3.2; the only evidence for this transfer is the benchmark being evaluated.
  • domain assumption Dependence edges can be approximated by resolving method calls via name plus argument count, connecting all overloaded candidates.
    The custom graph generator in Section 3.1 uses this approximation; false-positive or false-negative edges propagate into the smoothed embeddings.
  • domain assumption Spectral-style one-step propagation (Eq. 2) is a valid way to merge method semantic embeddings.
    Adapted from GCN first-order approximation but used without training or graph supervision; its validity for IA is asserted by the reported gains.

pith-pipeline@v1.3.0-alltime-deepseek · 27626 in / 10577 out tokens · 108317 ms · 2026-07-31T23:38:48.731544+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Enhancing Code Understanding for Impact Analysis by Combining Transformers and Program Dependence Graphs." pith.science (2026). https://pith.science/paper/H374L43D

@misc{pith2026260723355,
  author       = {Pith},
  title        = {Pith review of: Enhancing Code Understanding for Impact Analysis by Combining Transformers and Program Dependence Graphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/H374L43D}},
  note         = {Machine review of arXiv:2607.23355}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Impact analysis (IA) is a critical software maintenance task that identifies the effects of a given set of code changes on a larger software project with the intention of avoiding potential adverse effects. IA is a cognitively challenging task that involves reasoning about the abstract relationships between various code constructs. Given its difficulty, researchers have worked to automate IA with approaches that primarily use coupling metrics as a measure of the "connectedness" of different parts of a software project. Many of these coupling metrics rely on static, dynamic, or evolutionary information and are based on heuristics that tend to be brittle, require expensive execution analysis, or large histories of co-changes to accurately estimate impact sets. In this paper, we introduce a novel IA approach, called Athena, that combines a software system's dependence graph information with a conceptual coupling approach that uses advances in deep representation learning for code without the need for change histories and execution information. Previous IA benchmarks are small, containing fewer than ten software projects, and suffer from tangled commits, making it difficult to measure accurate results. Therefore, we constructed a large-scale IA benchmark, called Alexandria, from 25 open-source software projects, that utilizes fine-grained commit information from bug fixes. On this new benchmark, our best-performing approach configuration achieves mRR, mAP, and HIT@10 scores of 60.32%, 35.19%, and 81.48%, respectively. Through various ablations and qualitative analyses, we show that Athena's novel combination of program dependence graphs and conceptual coupling information leads it to outperform a simpler baseline by 10.34%, 9.55%, and 11.68% with statistical significance.

Figures

Figures reproduced from arXiv: 2607.23355 by Denys Poshyvanyk, Gabriele Bavota, Kevin Moran, Nathan Cooper, Steve Rich, Yanfu Yan.

Figure 1
Figure 1. Figure 1: Overview of the workflow of the Athena impact analysis approach. also been increasingly applied to learn code representations and automate software engineering tasks [Ahmad et al. 2021; Guo et al. 2022; Wang et al. 2021b,a]. CodeBERT [Feng et al. 2020] was one of the first Transformer-based NL-PL pre-trained models for supporting various code-related tasks. It distinguishes between the PL and the NL modali… view at source ↗
Figure 2
Figure 2. Figure 2: Two qualitative examples illustrating the effectiveness of [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

59 extracted references · 11 canonical work pages

  1. [8]

    In2020 IEEE 27th International Conference on Software Analysis, Evolution and Reengineering (SANER)

    A Reflection on the Predictive Accuracy of Dynamic Impact Analysis. In2020 IEEE 27th International Conference on Software Analysis, Evolution and Reengineering (SANER). 562–566. https://doi.org/10.1109/SANER48275. 2020.9054806 Haipeng Cai and Raul Santelices

    arXiv 2020

  2. [13]

    InProceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation, Dekai Wu, Marine Carpuat, Xavier Carreras, and Eva Maria Vecchi (Eds.)

    On the Properties of Neural Machine Translation: Encoder–Decoder Approaches. InProceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation, Dekai Wu, Marine Carpuat, Xavier Carreras, and Eva Maria Vecchi (Eds.). Association for Computational Linguistics, Doha, Qatar, 103–111. https://doi.org/10.3115/v1/W14-4012 Mic...

    doi:10.3115/v1/w14-4012

  3. [15]

    https://doi.org/10

    An Industrial Survey of Safety Evidence Change Impact Analysis Practice.IEEE Transactions on Software Engineering42, 12 (2016), 1095–1117. https://doi.org/10. Proc. ACM Softw. Eng., Vol. 1, No. FSE, Article

    2016

  4. [17]

    InProceedings of the 30th International Conference on Neural Information Processing Systems (Barcelona, Spain)(NIPS’16)

    Convolutional neural networks on graphs with fast localized spectral filtering. InProceedings of the 30th International Conference on Neural Information Processing Systems (Barcelona, Spain)(NIPS’16). Curran Associates Inc., Red Hook, NY, USA, 3844–3852. https://doi.org/10.5555/3157382. 3157527 Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova

    doi:10.5555/3157382

  5. [19]

    Engg.18, 2 (apr 2013), 277–309

    Integrating information retrieval, execution and link analysis algorithms to improve feature location in software.Empirical Softw. Engg.18, 2 (apr 2013), 277–309. https://doi.org/10. 1007/s10664-011-9194-4 Zhangyin Feng, Daya Guo, Duyu Tang, Nan Duan, Xiaocheng Feng, Ming Gong, Linjun Shou, Bing Qin, Ting Liu, Daxin Jiang, and Ming Zhou

    2013

  6. [20]

    InFindings of the Association for Computational Linguistics: EMNLP 2020, Trevor Cohn, Yulan He, and Yang Liu (Eds.)

    CodeBERT: A Pre-Trained Model for Programming and Natural Languages. InFindings of the Association for Computational Linguistics: EMNLP 2020, Trevor Cohn, Yulan He, and Yang Liu (Eds.). Association for Computational Linguistics, Online, 1536–1547. https://doi.org/10.18653/v1/2020.findings-emnlp.139 Stephen Fink and Julian Dolby

    doi:10.18653/v1/2020.findings-emnlp.139 2020

  7. [23]

    arXiv:2009.08366 https://arxiv.org/abs/2009.08366 Alex Gyori, Shuvendu K

    GraphCodeBERT: Pre-training Code Representations with Data Flow.CoRRabs/2009.08366 (2020). arXiv:2009.08366 https://arxiv.org/abs/2009.08366 Alex Gyori, Shuvendu K. Lahiri, and Nimrod Partush

    Pith/arXiv arXiv 2009

  8. [24]

    InProceedings of the 26th ACM SIGSOFT International Symposium on Software Testing and Analysis (Santa Barbara, CA, USA)(ISSTA 2017)

    Refining interprocedural change-impact analysis using equivalence relations. InProceedings of the 26th ACM SIGSOFT International Symposium on Software Testing and Analysis (Santa Barbara, CA, USA)(ISSTA 2017). Association for Computing Machinery, New York, NY, USA, 318–328. https: //doi.org/10.1145/3092703.3092719 David K. Hammond, Pierre Vandergheynst, a...

    arXiv 2017

  9. [25]

    Wavelets on graphs via spectral graph theory.Applied and Computational Harmonic Analysis30, 2 (2011), 129–150. https://doi.org/10.1016/j.acha.2010.04.005 Steffen Herbold, Alexander Trautsch, Benjamin Ledel, Alireza Aghamohammadi, Taher Ahmed Ghaleb, Kuljit Kaur Chahal, Tim Bossenmaier, Bhaveet Nagaria, Philip Makedonski, Matin Nili Ahmadabadi, Kristóf Sza...

    doi:10.1016/j.acha.2010.04.005 2011

  10. [26]

    Empirical Software Engineering27, 6 (2022),

    A Fine-grained Data Set and Analysis of Tangling in Bug Fixing Commits. Empirical Software Engineering27, 6 (2022),

    2022

  11. [28]

    Empirical Software Engineering21 (04 2015)

    The impact of tangled code changes on defect prediction models. Empirical Software Engineering21 (04 2015). https://doi.org/10.1007/s10664-015-9376-6 Kim Herzig and Andreas Zeller

    work page doi:10.1007/s10664-015-9376-6 2015

  12. [29]

    In2013 10th Working Conference on Mining Software Repositories (MSR)

    The impact of tangled code changes. In2013 10th Working Conference on Mining Software Repositories (MSR). 121–130. https://doi.org/10.1109/MSR.2013.6624018 Hamel Husain, Ho-Hsiang Wu, Tiferet Gazit, Miltiadis Allamanis, and Marc Brockschmidt

    arXiv 2013

  13. [30]

    arXiv:1909.09436 http://arxiv.org/abs/1909

    CodeSearchNet Challenge: Evaluating the State of Semantic Code Search.CoRRabs/1909.09436 (2019). arXiv:1909.09436 http://arxiv.org/abs/1909. 09436 Mohammad-Amin Jashki, Reza Zafarani, and Ebrahim Bagheri

    Pith/arXiv arXiv 1909

  14. [32]

    https://doi.org/10.1016/j.jss.2019.110441 Proc

    Automatic Method Change Suggestion to Complement Multi-Entity Edits.Journal of Systems and Software159 (10 2019), 110441. https://doi.org/10.1016/j.jss.2019.110441 Proc. ACM Softw. Eng., Vol. 1, No. FSE, Article

    arXiv 2019

  15. [33]

    https://doi.org/10.1007/s10664-012-9233-9 Huzefa Kagdi, Malcom Gethers, Denys Poshyvanyk, and Michael L

    Integrating conceptual and logical couplings for change impact analysis in software.Empirical Software Engineering18 (10 2012). https://doi.org/10.1007/s10664-012-9233-9 Huzefa Kagdi, Malcom Gethers, Denys Poshyvanyk, and Michael L. Collard

    doi:10.1007/s10664-012-9233-9 2012

  16. [34]

    In2010 17th Working Conference on Reverse Engineering

    Blending Conceptual and Evolutionary Couplings to Support Change Impact Analysis in Source Code. In2010 17th Working Conference on Reverse Engineering. 119–128. https://doi.org/10.1109/WCRE.2010.21 Yoon Kim

    work page doi:10.1109/wcre.2010.21 2010

  17. [35]

    InProceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), Alessandro Moschitti, Bo Pang, and Walter Daelemans (Eds.)

    Convolutional Neural Networks for Sentence Classification. InProceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), Alessandro Moschitti, Bo Pang, and Walter Daelemans (Eds.). Association for Computational Linguistics, Doha, Qatar, 1746–1751. https://doi.org/10.3115/v1/D14-1181 Thomas N. Kipf and Max Welling

    doi:10.3115/v1/d14-1181 2014

  18. [36]

    In5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings

    Semi-Supervised Classification with Graph Convolutional Networks. In5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net. https://openreview.net/forum?id=SJU4ayYgl Hiroyuki Kirinuki, Yoshiki Higo, Keisuke Hotta, and Shinji Kusumoto

    2017

  19. [37]

    In Proceedings of the 22nd International Conference on Program Comprehension(Hyderabad, India)(ICPC 2014)

    Hey! are you committing tangled changes?. In Proceedings of the 22nd International Conference on Program Comprehension(Hyderabad, India)(ICPC 2014). Association for Computing Machinery, New York, NY, USA, 262–265. https://doi.org/10.1145/2597008.2597798 Hiroyuki Kirinuki, Yoshiki Higo, Keisuke Hotta, and Shinji Kusumoto

    arXiv 2014

  20. [38]

    In 2016 23rd Asia-Pacific Software Engineering Conference (APSEC)

    Splitting Commits via Past Code Changes. In 2016 23rd Asia-Pacific Software Engineering Conference (APSEC). 129–136. https://doi.org/10.1109/APSEC.2016.028 Pavneet Singh Kochhar, Yuan Tian, and David Lo

    work page doi:10.1109/apsec.2016.028 2016

  21. [39]

    InProceedings of the 29th ACM/IEEE International Conference on Automated Software Engineering(Vasteras, Sweden)(ASE ’14)

    Potential biases in bug localization: do they matter?. InProceedings of the 29th ACM/IEEE International Conference on Automated Software Engineering(Vasteras, Sweden)(ASE ’14). Association for Computing Machinery, New York, NY, USA, 803–814. https://doi.org/10.1145/2642937.2642997 Hongyu Kuang, Patrick Mäder, Hao Hu, Achraf Ghabi, Liguo Huang, Lv Jian, an...

    arXiv

  22. [40]

    https://doi.org/10.1109/ICSM.2012.6405270 Quoc V

    Do data dependencies in source code complement call dependencies for understanding requirements traceability? 181–190. https://doi.org/10.1109/ICSM.2012.6405270 Quoc V. Le and Tomás Mikolov

    arXiv 2012

  23. [41]

    arXiv:1405.4053 http://arxiv.org/abs/1405.4053 Bixin Li, Xiaobing Sun, Hareton Leung, and Sai Zhang

    Distributed Representations of Sentences and Documents.CoRRabs/1405.4053 (2014). arXiv:1405.4053 http://arxiv.org/abs/1405.4053 Bixin Li, Xiaobing Sun, Hareton Leung, and Sai Zhang

    Pith/arXiv arXiv 2014

  24. [42]

    Software Testing, Verification and Reliability23 (12 2013)

    A survey of code-based change impact analysis techniques. Software Testing, Verification and Reliability23 (12 2013). https://doi.org/10.1002/stvr.1475 Ilya Loshchilov and Frank Hutter

    work page doi:10.1002/stvr.1475 2013

  25. [43]

    In7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9,

    Decoupled Weight Decay Regularization. In7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9,

    2019

  26. [44]

    arXiv:2102.04664 https://arxiv.org/abs/2102.04664 Tomás Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean

    CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generation.CoRRabs/2102.04664 (2021). arXiv:2102.04664 https://arxiv.org/abs/2102.04664 Tomás Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean

    Pith/arXiv arXiv 2021

  27. [45]

    In1st International Conference on Learning Representations, ICLR 2013, Scottsdale, Arizona, USA, May 2-4, 2013, Workshop Track Proceedings, Yoshua Bengio and Yann LeCun (Eds.)

    Efficient Estimation of Word Representations in Vector Space. In1st International Conference on Learning Representations, ICLR 2013, Scottsdale, Arizona, USA, May 2-4, 2013, Workshop Track Proceedings, Yoshua Bengio and Yann LeCun (Eds.). http://arxiv.org/abs/1301.3781 Chris Mills, Esteban Parra, Jevgenija Pantiuchina, Gabriele Bavota, and Sonia Haiduc

    Pith/arXiv arXiv 2013

  28. [46]

    https: //doi.org/10.1007/s10664-020-09823-w Hoan Anh Nguyen, Anh Tuan Nguyen, and Tien N

    On the relationship between bug reports and queries for text retrieval-based bug localization.Empirical Software Engineering25 (09 2020). https: //doi.org/10.1007/s10664-020-09823-w Hoan Anh Nguyen, Anh Tuan Nguyen, and Tien N. Nguyen

    work page doi:10.1007/s10664-020-09823-w 2020

  29. [47]

    In2013 IEEE 24th International Symposium on Software Reliability Engineering (ISSRE)

    Filtering noise in mixed-purpose fixing commits to improve defect prediction and localization. In2013 IEEE 24th International Symposium on Software Reliability Engineering (ISSRE). 138–147. https://doi.org/10.1109/ISSRE.2013.6698913 Denys Poshyvanyk, Andrian Marcus, Rudolf Ferenc, and Tibor Gyimóthy

    arXiv 2013

  30. [49]

    arXiv:1910.10683 http://arxiv.org/abs/1910.10683 Sable Research Group

    Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer.CoRRabs/1910.10683 (2019). arXiv:1910.10683 http://arxiv.org/abs/1910.10683 Sable Research Group

    Pith/arXiv arXiv 1910

  31. [50]

    In2008 1st International Conference on Software Testing, Verification, and Validation

    Empirical Software Change Impact Analysis using Singular Value Decomposition. In2008 1st International Conference on Software Testing, Verification, and Validation. 268–277. https://doi.org/10.1109/ ICST.2008.25 Yonghee Shin, Andrew Meneely, Laurie Williams, and Jason A. Osborne

    2008

  32. [51]

    https://doi.org/10.1109/TSE.2010.81 Jeffrey Svajlenko, Judith F

    Evaluating Complexity, Code Churn, and Developer Activity Metrics as Indicators of Software Vulnerabilities.IEEE Transactions on Software Engineering37, 6 (2011), 772–787. https://doi.org/10.1109/TSE.2010.81 Jeffrey Svajlenko, Judith F. Islam, Iman Keivanloo, Chanchal K. Roy, and Mohammad Mamun Mia

    work page doi:10.1109/tse.2010.81 2011

  33. [52]

    In2014 IEEE International Conference on Software Maintenance and Evolution

    Towards a Big Data Curated Benchmark of Inter-project Code Clones. In2014 IEEE International Conference on Software Maintenance and Evolution. 476–480. https://doi.org/10.1109/ICSME.2014.77 Proc. ACM Softw. Eng., Vol. 1, No. FSE, Article

    doi:10.1109/icsme.2014.77 2014

  34. [53]

    InProceedings of the ACM SIGSOFT 20th International Symposium on the Foundations of Software Engineering(Cary, North Carolina)(FSE ’12)

    How do software engineers understand code changes? an exploratory study in industry. InProceedings of the ACM SIGSOFT 20th International Symposium on the Foundations of Software Engineering(Cary, North Carolina)(FSE ’12). Association for Computing Machinery, New York, NY, USA, Article 51, 11 pages. https://doi.org/10.1145/2393596.2393656 Marco Torchiano a...

    arXiv

  35. [54]

    InProceedings of the 2010 ACM-IEEE International Symposium on Empirical Software Engineering and Measurement(Bolzano-Bozen, Italy)(ESEM ’10)

    Impact analysis by means of unstructured knowledge in the context of bug repositories. InProceedings of the 2010 ACM-IEEE International Symposium on Empirical Software Engineering and Measurement(Bolzano-Bozen, Italy)(ESEM ’10). Association for Computing Machinery, New York, NY, USA, Article 47, 4 pages. https://doi.org/10.1145/1852786.1852847 Ashish Vasw...

    arXiv 2010

  36. [55]

    https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf Min Wang, Zeqi Lin, Yanzhen Zou, and Bing Xie

    Curran Associates, Inc. https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf Min Wang, Zeqi Lin, Yanzhen Zou, and Bing Xie

    2017

  37. [56]

    In2019 34th IEEE/ACM International Conference on Automated Software Engineering (ASE)

    CoRA: Decomposing and Describing Tangled Code Changes for Reviewer. In2019 34th IEEE/ACM International Conference on Automated Software Engineering (ASE). 1050–1061. https://doi.org/10.1109/ASE.2019.00101 Wei Wang, Yun He, Tong Li, Jiajun Zhu, and Jinzhuo Liu

    arXiv 2019

  38. [59]

    In Proceedings

    Mining version histories to guide software changes. In Proceedings. 26th International Conference on Software Engineering. 563–572. https://doi.org/10.1109/ICSE.2004.1317478 Received 2023-09-29; accepted 2024-01-23 Proc. ACM Softw. Eng., Vol. 1, No. FSE, Article

    Pith/arXiv arXiv 2004

  39. [60]

    Publication date: July 2024

    2024

  40. [125]

    https://doi.org/10.1007/s10664-021-10083-5 Kim Herzig, Sascha Just, and Andreas Zeller

    doi:10.1007/s10664-021-10083-5

  41. [2004]

    https://doi.org/10

    Proceedings.453–457. https://doi.org/10. 1109/ICSM.2004.1357834 Ben Breech, Mike Tegtmeyer, and Lori Pollock

    Pith/arXiv arXiv 2004

  42. [2005]

    In12th Asia-Pacific Software Engineering Conference (APSEC’05)

    Supporting predictive change impact analysis: a control call graph based technique. In12th Asia-Pacific Software Engineering Conference (APSEC’05). 9 pp.–. https://doi.org/10.1109/APSEC. 2005.100 Markus Borg, Krzysztof Wnuk, Björn Regnell, and Per Runeson

    work page doi:10.1109/apsec 2005

  43. [2006]

    In2006 22nd IEEE International Conference on Software Maintenance

    Integrating Influence Mechanisms into Impact Analysis for Increased Precision. In2006 22nd IEEE International Conference on Software Maintenance. 55–65. https://doi.org/10.1109/ICSM.2006. 33 Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Ag...

    work page doi:10.1109/icsm.2006 2006

  44. [2008]

    InProceedings of the 8th ACM SIGPLAN-SIGSOFT Workshop on Program Analysis for Software Tools and Engineering(Atlanta, Georgia)(PASTE ’08)

    Towards a more efficient static software change impact analysis method. InProceedings of the 8th ACM SIGPLAN-SIGSOFT Workshop on Program Analysis for Software Tools and Engineering(Atlanta, Georgia)(PASTE ’08). Association for Computing Machinery, New York, NY, USA, 84–90. https://doi.org/10.1145/1512475.1512493 Zijian Jiang, Ye Wang, Hao Zhong, and Na Meng

    arXiv

  45. [2009]

    https://doi.org/10.1007/s10664-008- 9088-2 Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J

    Using information retrieval based coupling measures for impact analysis.Empirical Software Engineering14 (02 2009), 5–32. https://doi.org/10.1007/s10664-008- 9088-2 Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu

    doi:10.1007/s10664-008- 2009

  46. [2010]

    In2010 IEEE International Conference on Software Maintenance

    Using multivariate time series and association rules to detect logical change coupling: An empirical study. In2010 IEEE International Conference on Software Maintenance. 1–10. https://doi.org/10.1109/ICSM.2010.5609732 Kyunghyun Cho, Bart van Merriënboer, Dzmitry Bahdanau, and Yoshua Bengio

    arXiv 2010

  47. [2011]

    InProceedings of the 33rd International Conference on Software Engineering(Waikiki, Honolulu, HI, USA) (ICSE ’11)

    Practical change impact analysis based on static program slicing for industrial software systems. InProceedings of the 33rd International Conference on Software Engineering(Waikiki, Honolulu, HI, USA) (ICSE ’11). Association for Computing Machinery, New York, NY, USA, 746–755. https://doi.org/10.1145/1985793.1985898 Proc. ACM Softw. Eng., Vol. 1, No. FSE, Article

    arXiv

  48. [2012]

    In2012 34th International Conference on Software Engineering (ICSE)

    Integrated impact analysis for managing software changes. In2012 34th International Conference on Software Engineering (ICSE). 430–440. https://doi.org/10.1109/ICSE. 2012.6227172 Daya Guo, Shuai Lu, Nan Duan, Yanlin Wang, Ming Zhou, and Jian Yin

    arXiv 2012

  49. [2013]

    In2013 IEEE International Conference on Software Maintenance

    srcML: An Infrastructure for the Exploration, Analysis, and Manipulation of Source Code: A Tool Demonstration. In2013 IEEE International Conference on Software Maintenance. 516–519. https://doi.org/10.1109/ICSM.2013.85 Jose Luis de la Vara, Markus Borg, Krzysztof Wnuk, and Leon Moonen

    work page doi:10.1109/icsm.2013.85 2013

  50. [2015]

    https://doi.org/10.1016/j.jss.2015.02.018 Haipeng Cai, Raul Santelices, and Siyuan Jiang

    A comprehensive study of the predictive accuracy of dynamic change-impact analysis.Journal of Systems and Software103 (2015), 248–265. https://doi.org/10.1016/j.jss.2015.02.018 Haipeng Cai, Raul Santelices, and Siyuan Jiang. 2016a. Prioritizing Change-Impact Analysis via Semantic Program- Dependence Quantification.IEEE Transactions on Reliability65, 3 (20...

    work page doi:10.1016/j.jss.2015.02.018 2015

  51. [2016]

    In Proceedings of the 31st IEEE/ACM International Conference on Automated Software Engineering(Singapore, Singapore) (ASE ’16)

    DistIA: a cost-effective dynamic impact analysis for distributed programs. In Proceedings of the 31st IEEE/ACM International Conference on Automated Software Engineering(Singapore, Singapore) (ASE ’16). Association for Computing Machinery, New York, NY, USA, 344–355. https://doi.org/10.1145/2970276.2970352 Gerardo Canfora, Michele Ceccarelli, Luigi Cerulo...

    arXiv

  52. [2017]

    https://doi.org/10.1109/TSE.2016.2620458 Ben Breech, Anthony Danalis, Stacey Shindo, and Lori Pollock

    Supporting Change Impact Analysis Using a Recommendation System: An Industrial Case Study in a Safety-Critical Context.IEEE Transactions on Software Engineering 43, 07 (jul 2017), 675–700. https://doi.org/10.1109/TSE.2016.2620458 Ben Breech, Anthony Danalis, Stacey Shindo, and Lori Pollock

    arXiv 2017

  53. [2018]

    https://doi.org/10.1155/2018/5913634 Xin Wang, Yasheng Wang, Pingyi Zhou, Fei Mi, Meng Xiao, Yadao Wang, Li Li, Xiao Liu, Hao Wu, Jin Liu, and Xin Jiang

    An Integrated Model for Information Retrieval Based Change Impact Analysis.Scientific Programming2018 (03 2018), 1–13. https://doi.org/10.1155/2018/5913634 Xin Wang, Yasheng Wang, Pingyi Zhou, Fei Mi, Meng Xiao, Yadao Wang, Li Li, Xiao Liu, Hao Wu, Jin Liu, and Xin Jiang. 2021b. CLSEBERT: Contrastive Learning for Syntax Enhanced Code Pre-Trained Model.CoR...

    Pith/arXiv arXiv 2018

  54. [2019]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. InProceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2019, Minneapolis, MN, USA, June 2-7, 2019, Volume 1 (Long and Short Papers). Association for Computational Linguistics...

    2019

  55. [2020]

    InProceedings of the 34th International Conference on Neural Information Processing Systems(Vancouver, BC, Canada)(NIPS’20)

    Language models are few-shot learners. InProceedings of the 34th International Conference on Neural Information Processing Systems(Vancouver, BC, Canada)(NIPS’20). Curran Associates Inc., Red Hook, NY, USA, Article 159, 25 pages. https://doi.org/10.5555/3495724.3495883 Max Brunsfeld, Patrick Thomson, Andrew Hlynskyi, Josh Vera, Phil Turnbull, Timothy Clem...

    arXiv 2022

  56. [2021]

    InProceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies

    Unified Pre-training for Program Understanding and Generation. InProceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguistics, Online, 2655–2668. https: //doi.org/10.18653/v1/2021.naacl-main.211 Robert S Arnold. 1996.Software change...

    doi:10.18653/v1/2021.naacl-main.211 2021

  57. [2022]

    InProceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Smaranda Muresan, Preslav Nakov, and Aline Villavicencio (Eds.)

    UniXcoder: Unified Cross-Modal Pre-training for Code Representation. InProceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Smaranda Muresan, Preslav Nakov, and Aline Villavicencio (Eds.). Association for Computational Linguistics, Dublin, Ireland, 7212–7225. https://doi.org/10.18653/v1/2022.acl-...

    doi:10.18653/v1/2022.acl-long.499 2022

  58. [2024]

    44:22 Yanfu Yan, Nathan Cooper, Kevin Moran, Gabriele Bavota, Denys Poshyvanyk, Steve Rich 1109/TSE.2016.2553032 Michaël Defferrard, Xavier Bresson, and Pierre Vandergheynst

    arXiv 2016

  59. [2026]

    https://doi.org/ 10.5281/zenodo.21569282

    Athena: Enhancing Code Understanding for Impact Analysis by Combining Transformers and Program Dependence Graphs. https://doi.org/ 10.5281/zenodo.21569282. https://doi.org/10.5281/zenodo.21569282 Zhilin Yang, Zihang Dai, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, and Quoc V. Le. 2019.XLNet: generalized autoregressive pretraining for language unde...

    work page doi:10.5281/zenodo.21569282 2019