Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Target-Augmented Shared Fusion-based Multimodal Sarcasm Explanation Generation

T0 review · 3 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read TURBO, a target-augmented shared-fusion model, claims to produce multimodal sarcasm explanations that beat the previous state of the art by an average of +3.3% on the MORE+ dataset.

desk verdict Headline +3.3% over TEAM is inflated by test-time gold targets; the underlying resource and architecture are still worth a look. read the letter →

arxiv 2502.07391 v1 pith:JMFLL2NI submitted 2025-02-11 cs.CL

Palaash Goel , Dushyant Singh Chauhan , Md Shad Akhtar This is my paper · ORCID
classification cs.CL
keywords multimodalsarcasmexplanationtargetofsharedfusionknowledgegraphBARTMORE+datasetgeneration
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

This paper tries to establish that knowing the target of a sarcastic remark—the entity, event, or person being ridiculed—makes a multimodal explanation generator produce better explanations, and that a shared fusion of image and caption features is the right way to exploit that knowledge. The authors introduce TURBO, a model built on BART with a knowledge graph, a graph convolutional network, and a gated shared-fusion module that combines self-attended visual and textual representations. They extend the MORE dataset with manually annotated targets, calling it MORE+, and report that TURBO outperforms the previous state-of-the-art model TEAM by an average margin of +3.3% on standard generation metrics. A human evaluation finds TURBO's explanations preferred over TEAM's, while two larger multimodal LLMs score higher in some human ratings despite having roughly 30 times more parameters.

What carries the argument

The central mechanism is the shared fusion module: self-attention is applied separately to text and image embeddings, then each modality's attention matrix amplifies the other's features, and four gated combinations—two multimodal pairs and two unimodal-versus-multimodal pairs—are mixed by learnable weights into a single fused representation. Around this sit the knowledge graph (ConceptNet relations with relevance-score edge weights, processed by a graph convolutional network) and the target-of-sarcasm string concatenated into the input with a BART separator token. The fused matrix and the graph output are summed and passed to BART for autoregressive explanation generation. The target's role is to guide which of these fused and graph-derived features matter for the intended irony.

What would settle it

Re-run the MORE+ evaluation with the target withheld from TURBO (as in the TURBO−TS ablation) and compare directly against TEAM under identical inputs, with significance testing across multiple seeds; if the target-free TURBO does not exceed TEAM on the primary metrics, the paper's central claim that the model architecture plus target explains the improvement is not supported.

Watch

Extended reading notes

Core claim

On the authors' own account, the central discovery is that the intended target of sarcasm is a load-bearing input for explaining it. TURBO incorporates that target by concatenating it with knowledge-enriched text, computes a weighted knowledge graph from captions, image descriptions, detected objects, and ConceptNet relations, and fuses visual and textual embeddings through a shared-fusion mechanism with learnable gates. Against the MORE+ test set, TURBO improves on TEAM on every reported automatic metric, with the largest average gains on ROUGE and METEOR (+3.33%), and its ablations show that removing the target, the shared fusion, or the knowledge graph each degrades performance. The paper also claims that TURBO's explanations are judged better than TEAM's in a human study, and are competitive with much larger multimodal LLMs despite being about 30 times smaller.

Load-bearing premise

The headline comparison assumes the target of sarcasm is available as a manually annotated input at test time; the prior state-of-the-art model is not given this extra information, so the reported gain rests on that annotation being present or correctly predicted.

Editorial extensions

If this is right

  • If the reported margin holds, target annotation becomes a cheap and effective lever for sarcasm explanation systems: labeling who or what is mocked improves output quality on automatic and human metrics.
  • The ablation results imply that a strong multimodal explanation model needs all three components—visual detail, external knowledge, and target conditioning—and that removing any one measurably hurts.
  • The success of a 234M-parameter model against 7–8B multimodal LLMs suggests the task rewards structured fusion and explicit target information more than raw scale, a point the paper's parameter comparison emphasizes.
  • Since the target is currently a manual input, the approach motivates training a separate target predictor, which the paper names as the natural next step for end-to-end deployment.

Reading between the lines

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

  • If the target input were replaced by a predicted target, the +3.3% advantage would likely shrink; the paper's own TURBO−TS ablation sits only marginally above TEAM, so the target annotation, not the fusion alone, appears to carry much of the reported edge.
  • The shared-fusion design transfers naturally to other multimodal generation tasks where one modality should amplify the other, such as image-grounded summarization or visual question answering, though the paper does not test this.
  • The human evaluation's mixed ranking—LLMs beating TURBO on semantic accuracy while TURBO beats TEAM—suggests automatic n-gram metrics may understate LLM explanation quality; a fair comparison would need a target-prediction-free setup.
  • A testable extension is to measure whether the target annotation itself is the source of the gain or whether the model is simply benefiting from an extra text clue, by feeding TEAM the same target string and re-running the comparison.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper introduces TURBO, a model for multimodal sarcasm explanation generation on the MORE+ dataset, which extends the MORE dataset with manually annotated target-of-sarcasm labels. TURBO concatenates the gold target into the input text, enriches the caption with BLIP image descriptions, YOLO object labels, and ConceptNet knowledge, builds an undirected weighted knowledge graph, applies a GCN, and combines textual and visual representations through self-attention and a gated shared-fusion mechanism before generating explanations with BART. The authors report automatic-metric gains over TEAM and other baselines, explore zero- and one-shot LLM baselines, run ablations, and provide a human evaluation.

Significance. If the central comparison were on equal footing, the paper would be a solid contribution: it introduces a new target-of-sarcasm annotation layer for MORE, makes code and data available, provides a clearly specified architecture, and includes a thorough ablation study and error analysis. The main issue is that the model is given the gold target at test time while no baseline receives this input, so the headline improvement over TEAM is confounded. The architecture-only variant, TURBO−TS, is close to TEAM on most metrics, with differences that are small and mixed in direction. The contribution may still be valuable in settings where target annotations are available, but the paper's central claim of state-of-the-art superiority needs reframing or additional equal-input evidence.

major comments (3)
  1. [§4.5, Eq. (10); Table 2] The gold target-of-sarcasm token sequence TS is concatenated into the model input at test time, while none of the compared baselines, including TEAM, receive this annotation. The appropriate equal-input comparison is therefore TURBO−TS in Table 2, and it is not compelling: BLEU-1 is 55.37 vs. TEAM's 55.32, BLEU-2 45.09 vs. 45.12, ROUGE-R1 52.25 vs. 51.72, ROUGE-R2 35.41 vs. 34.96, METEOR 52.08 vs. 50.95, BERTScore-F1 91.51 vs. 91.70, and SentBERT 73.62 vs. 72.92. These differences are small and mixed in direction, and no significance tests are reported. Since Section 7 acknowledges that the target is an additional manually provided input and only suggests training a future predictor, the abstract's claim that TURBO outperforms the state of the art by +3.3% is not established on equal footing.
  2. [§5.1, Table 2] The '+3.3% average margin' in the abstract is not an average over the metrics in Table 2 as a whole. It corresponds only to the four relative gains reported for ROUGE-RL (+2.54%), ROUGE-R1 (+3.34%), ROUGE-R2 (+3.20%), and METEOR (+4.22%). The corresponding averages are +1.92% for BLEU-1..4, +0.18% for BERTScore, and +2.83% for SentBERT. The abstract and Section 5.1 should state which metric subset the headline margin refers to, or report a pre-specified aggregate.
  3. [§5.1, §5.2] The paper uses 'significant' and 'significantly' repeatedly (e.g., 'significant disparity', 'each component contributes significantly to the model's performance') without reporting variance, confidence intervals, or paired significance tests. Given the small and mixed differences between TURBO−TS and TEAM in Table 2, a claim of superiority on individual metrics cannot be assessed. Please add significance testing or at least standard errors, and adjust the wording accordingly.
minor comments (4)
  1. [Table 2] The TURBO row has numeric entries typeset with no separators (e.g., '57.0946.9340.28...'), making the row unreadable; please fix the table formatting.
  2. [§5.3, Table 3] The human evaluation covers 20 samples; please report the number of judgments per item, inter-annotator agreement, and significance tests, and avoid calling this 'extensive' in the abstract without qualification.
  3. [§5.1] The phrase 'in line with the existing systems (Desai et al., 2022; Jiang, 2023)' cites 'Jiang, 2023' as an evaluation-metric reference, but the reference list entry is a SemEval-2023 system paper that appears unrelated; please verify and either replace or justify the citation.
  4. [Title and Abstract] The title and abstract do not disclose that TURBO requires a manually annotated target-of-sarcasm at test time; please state this assumption explicitly and position the contribution accordingly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TURBO's gold target input is an evaluation confound, not a derivation loop.

full rationale

The paper's central claim is an empirical comparison, not a formal derivation. The only privileged input, the gold target of sarcasm TS, is concatenated to the input sequence in Eq. 10 and is neither fitted from the reference explanation nor generated by the model; the explanation is produced autoregressively from the multimodal input and the target. The TURBO−TS ablation in Table 2 directly isolates the contribution of this input, and Section 7 explicitly acknowledges that using the annotated target is an extra manually provided input and suggests training a separate target predictor. Giving one system an additional gold annotation while withholding it from all baselines is an evaluation confound that weakens the headline +3.3% comparison, but it does not make the reported improvement equivalent to the input by construction. The dataset's annotation guideline also states that the target phrase must not reveal the underlying sarcastic incongruity, so the target is deliberately distinct from the explanation. The GCN component follows the external TEAM work rather than a self-citation, and no uniqueness theorem or ansatz is imported from the authors' prior work. The human-evaluation metric 'Presence of Target' is naturally correlated with the target being given as input, but that is a stated evaluation criterion rather than a derivation step. Accordingly, no circular step can be exhibited from the paper's equations or argument.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests mainly on modeling assumptions and hand-set hyperparameters rather than new physical entities. No new particles or forces are introduced. The most consequential extra input is the gold target-of-sarcasm annotation, which is a human-labeled resource, not a fitted constant.

free parameters (4)
  • K (max number of detected objects) = 36
    Hand-chosen in Section 5; controls how many YOLOv9 object labels enter the knowledge-enriched text.
  • N (max sequence length) = 256
    Hand-chosen in Section 5; text and image embeddings are padded or truncated to 256 tokens.
  • GCN layer count L = not reported
    The method says L GCN layers are used (Section 4.6), but the experimental setup never states L, so the number is a free choice.
  • Fusion weights alpha1, alpha2, beta1, beta2 = learned
    Learnable combination weights in Equation 17; they are fit during training and the learned values are not reported.
assumptions (5)
  • domain assumption BART base provides suitable contextual representations for explanation generation
    Used as backbone in Section 4.8; the paper does not justify this choice beyond prior use.
  • domain assumption ConceptNet one-hop relevance scores are informative for sarcasm reasoning
    External knowledge concepts and scores from ConceptNet are concatenated and used as edge weights (Sections 4.2 to 4.4); no validation that the scores are reliable.
  • ad hoc to paper The element-wise fusion F_vt = A_t times E_v captures inter-modality relationships
    Equation 13 in Section 4.7; this is a modeling choice, not a derived result, and its validity is taken as given.
  • ad hoc to paper The target-of-sarcasm annotation is an appropriate extra input and does not leak the reference explanation
    The task design in Sections 3 and 4.5 assumes gold targets are available at test time; the guidelines try to avoid revealing the incongruity, but this is not quantitatively verified.
  • domain assumption Graph convolution aggregates knowledge effectively
    Section 4.6 relies on standard GCN propagation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Target-Augmented Shared Fusion-based Multimodal Sarcasm Explanation Generation." pith.science (2026). https://pith.science/paper/JMFLL2NI

@misc{pith2026250207391,
  author       = {Pith},
  title        = {Pith review of: Target-Augmented Shared Fusion-based Multimodal Sarcasm Explanation Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JMFLL2NI}},
  note         = {Machine review of arXiv:2502.07391}
}
abstract

Sarcasm is a linguistic phenomenon that intends to ridicule a target (e.g., entity, event, or person) in an inherent way. Multimodal Sarcasm Explanation (MuSE) aims at revealing the intended irony in a sarcastic post using a natural language explanation. Though important, existing systems overlooked the significance of the target of sarcasm in generating explanations. In this paper, we propose a Target-aUgmented shaRed fusion-Based sarcasm explanatiOn model, aka. TURBO. We design a novel shared-fusion mechanism to leverage the inter-modality relationships between an image and its caption. TURBO assumes the target of the sarcasm and guides the multimodal shared fusion mechanism in learning intricacies of the intended irony for explanations. We evaluate our proposed TURBO model on the MORE+ dataset. Comparison against multiple baselines and state-of-the-art models signifies the performance improvement of TURBO by an average margin of $+3.3\%$. Moreover, we explore LLMs in zero and one-shot settings for our task and observe that LLM-generated explanation, though remarkable, often fails to capture the critical nuances of the sarcasm. Furthermore, we supplement our study with extensive human evaluation on TURBO's generated explanations and find them out to be comparatively better than other systems.

Figures

Figures reproduced from arXiv: 2502.07391 by the authors.

Figure 1
Figure 1. A sample in the MORE+ dataset. for properly interpreting a sarcastic message. Ex￾isting research suggests significant dependence on cues from multiple sources to interpret sarcastic messages. These can include tone of voice, body language, common sense, etc. Furthermore, the task of identifying and understanding sarcasm is quite relevant in a multimodal scenario where each modality refers to a different source of sa… view at source ↗
Figure 2
Figure 2. A schematic diagram of TURBO. 2024). We keep the top K objects with the highest confidence to ensure that only the most relevant object-level semantics are retained. Y OLOv9(Vi) = Oi = {o (i) 1 , o (i) 2 , . . . o (i) K } (2) where o (i) j is a text label describing the j th object extracted from Vi using YOLOv9. High-Level Detail: Finally, we utilize semantic￾rich representations of visual features. These are extra… view at source ↗
Figure 3
Figure 3. Two samples where: a) TURBO correctly explained the sarcasm and GPT-4o mini missed it; and b) TURBO generated an inaccurate explanation but GPT-4o mini explained the sarcasm correctly. struct one’s view and that people are needlessly complaining about the same. While TURBO is able to understand the sarcasm and correctly points out that the netting does not ruin one’s view of the game, GPT-4o mini instead incorrectly… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Extraction of Irrelevant External Knowl￾edge Concepts: A sample where TURBO extracted irrel￾evant external knowledge concepts for the entities vital to understanding the sarcastic incongruity. Upon extracting the external knowledge con￾cepts for this sample from the mo…
Figure 6
Figure 6. Figure 6: Insufficient OCR Features: A sample con￾taining OCR text where TURBO generated a completely wrong sarcasm explanation. Irrelevant Image Description: We extract low￾level detail from the images in the form of im￾age descriptions generated using BLIP. While this model gi…
Figure 5
Figure 5. Figure 5: Missing External Knowledge Concepts: A sample where TURBO did not extract any knowledge concepts for the entity relevant to sarcasm. Insufficient OCR Features: A notable source of error in our model is that it does not explicitly extract features from the textual entit…
Figure 7
Figure 7. Figure 7: Irrelevant Image Description: A sample where TURBO generated a completely wrong sarcasm explanation due to irrelevant image description. model describes the bar in the foreground. As a result, even though the description is objectively ac￾curate, it does not contribute…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. ViSP: A PPO-Driven Framework for Sarcasm Generation with Contrastive Learning

    cs.CL 2025-07 reject novelty 5.0 of 10

    A new sarcasm-generation dataset and a reward-optimized vision-language model that outperforms LLMs on benchmark metrics, though its main quality metric is the same model used to train it.

Reference graph

Works this paper leans on

35 extracted references · 17 canonical work pages · cited by 1 Pith paper

  1. [1]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Nastaran Babanejad, Heidar Davoudi, Aijun An, and Manos Papagelis. 2020. https://doi.org/10.18653/v1/2020.coling-main.20 Affective and contextual embedding for sarcasm detection . In Proceedings of the 28th International Conference on Computational Linguistics, pages 225--243, Barcelona, Spain (Online). International Committee on Computational Linguistics

    doi:10.18653/v1/2020.coling-main.20 2020

  4. [4]

    M. Bedi, S. Kumar, M. Akhtar, and T. Chakraborty. 2023. https://doi.org/10.1109/TAFFC.2021.3083522 Multi-modal sarcasm detection and humor classification in code-mixed conversations . IEEE Transactions on Affective Computing, 14(02):1363--1375

    work page arXiv 2023

  5. [5]

    Steven Bird, Edward Loper, and Ewan Klein. 2009. Natural Language Processing with Python. O'Reilly Media Inc

    work page 2009

  6. [6]

    Mondher Bouazizi and Tomoaki Otsuki Ohtsuki. 2016. https://doi.org/10.1109/ACCESS.2016.2594194 A pattern-based approach for sarcasm detection on twitter . IEEE Access, 4:5477--5488

    work page arXiv 2016

  7. [7]

    Santiago Castro, Devamanyu Hazarika, Ver \'o nica P \'e rez-Rosas, Roger Zimmermann, Rada Mihalcea, and Soujanya Poria. 2019. https://doi.org/10.18653/v1/P19-1455 Towards multimodal sarcasm detection (an \_ O bviously \_ perfect paper) . In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4619--4629, Florence,...

    doi:10.18653/v1/p19-1455 2019

  8. [8]

    Tuhin Chakrabarty, Debanjan Ghosh, Smaranda Muresan, and Nanyun Peng. 2020. https://doi.org/10.18653/v1/2020.acl-main.711 R \^ 3: Reverse, retrieve, and rank for sarcasm generation with commonsense knowledge . In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7976--7986, Online. Association for Computational...

    work page doi:10.18653/v1/2020.acl-main.711 2020

Show all 35 references
  1. [9]

    Poorav Desai, Tanmoy Chakraborty, and Md Shad Akhtar. 2022. https://doi.org/10.1609/aaai.v36i10.21300 Nice perfume. how long did you marinate in it? multimodal sarcasm explanation . Proceedings of the AAAI Conference on Artificial Intelligence, 36(10):10563--10571

    2022 doi

  2. [10]

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. 2021. https://arxiv.org/abs/2010.11929 An image is worth 16x16 words...

    2021 arXiv

  3. [11]

    Abhijeet Dubey, Aditya Joshi, and Pushpak Bhattacharyya. 2019. https://doi.org/10.1145/3297001.3297043 Deep models for converting sarcastic utterances into their non sarcastic interpretation . In Proceedings of the ACM India Joint International Conference on Data Science and M...

    2019

  4. [12]

    Abhimanyu Dubey et al. 2024. https://arxiv.org/abs/2407.21783 The llama 3 herd of models . Preprint, arXiv:2407.21783

    2024 arXiv

  5. [13]

    Bjarke Felbo, Alan Mislove, Anders S gaard, Iyad Rahwan, and Sune Lehmann. 2017. https://doi.org/10.18653/v1/D17-1169 Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion and sarcasm . In Proceedings of the 2017 Conference on...

    2017 doi

  6. [14]

    Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas ...

    2023 arXiv

  7. [15]

    Ye Jiang. 2023. https://doi.org/10.18653/v1/2023.semeval-1.40 Team QUST at S em E val-2023 task 3: A comprehensive study of monolingual and multilingual approaches for detecting online news genre, framing and persuasion techniques . In Proceedings of the 17th International Wor...

    2023 doi

  8. [16]

    Liqiang Jing, Xuemeng Song, Kun Ouyang, Mengzhao Jia, and Liqiang Nie. 2023. https://doi.org/10.18653/v1/2023.acl-long.635 Multi-source semantic graph-based multimodal sarcasm explanation generation . In Proceedings of the 61st Annual Meeting of the Association for Computation...

    2023 doi

  9. [17]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. 2017. https://openreview.net/forum?id=SJU4ayYgl Semi-supervised classification with graph convolutional networks . In International Conference on Learning Representations

    2017

  10. [18]

    Shivani Kumar, Atharva Kulkarni, Md Shad Akhtar, and Tanmoy Chakraborty. 2022. https://doi.org/10.18653/v1/2022.acl-long.411 When did you become so smart, oh wise one?! sarcasm explanation in multi-modal multi-party dialogues . In Proceedings of the 60th Annual Meeting of the ...

    2022 doi

  11. [19]

    Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Veselin Stoyanov, and Luke Zettlemoyer. 2020. https://doi.org/10.18653/v1/2020.acl-main.703 BART : Denoising sequence-to-sequence pre-training for natural language generation, translatio...

    2020 doi

  12. [20]

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. https://arxiv.org/abs/2201.12086 Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation . Preprint, arXiv:2201.12086

    2022 arXiv

  13. [21]

    Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. 2024. https://arxiv.org/abs/2310.03744 Improved baselines with visual instruction tuning . Preprint, arXiv:2310.03744

    2024 arXiv

  14. [22]

    Nayu Liu, Xian Sun, Hongfeng Yu, Wenkai Zhang, and Guangluan Xu. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.144 Multistage fusion with forget gate for multimodal summarization in open-domain videos . In Proceedings of the 2020 Conference on Empirical Methods in Natural ...

    2020 doi

  15. [23]

    Ilya Loshchilov and Frank Hutter. 2017. Fixing weight decay regularization in adam

    2017

  16. [24]

    Lin Ma, Zhengdong Lu, Lifeng Shang, and Hang Li. 2015. https://doi.org/10.1109/ICCV.2015.301 Multimodal convolutional neural networks for matching image and sentence . In 2015 IEEE International Conference on Computer Vision (ICCV), pages 2623--2631

    2015 doi

  17. [25]

    Abhijit Mishra, Tarun Tater, and Karthik Sankaranarayanan. 2019. https://doi.org/10.18653/v1/D19-1636 A modular architecture for unsupervised sarcasm generation . In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th Internationa...

    2019 doi

  18. [26]

    OpenAI, Josh Achiam, et al. 2024. https://arxiv.org/abs/2303.08774 Gpt-4 technical report . Preprint, arXiv:2303.08774

    2024 arXiv

  19. [27]

    Lotem Peled and Roi Reichart. 2017. https://doi.org/10.18653/v1/P17-1155 Sarcasm SIGN : Interpreting sarcasm with sentiment based monolingual machine translation . In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers...

    2017 doi

  20. [28]

    Yang Qiao, Liqiang Jing, Xuemeng Song, Xiaolin Chen, Lei Zhu, and Liqiang Nie. 2023. https://doi.org/10.1609/aaai.v37i8.26138 Mutual-enhanced incongruity learning network for multi-modal sarcasm detection . Proceedings of the AAAI Conference on Artificial Intelligence, 37(8):9...

    2023 doi

  21. [29]

    Rossano Schifanella, Paloma de Juan, Joel Tetreault, and LiangLiang Cao. 2016. https://doi.org/10.1145/2964284.2964321 Detecting sarcasm in multimodal social platforms . In Proceedings of the 24th ACM International Conference on Multimedia, MM '16, page 1136–1145, New York, NY...

    2016

  22. [30]

    Liu, and Christopher D

    Abigail See, Peter J. Liu, and Christopher D. Manning. 2017. https://doi.org/10.18653/v1/P17-1099 Get to the point: Summarization with pointer-generator networks . In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers...

    2017 doi

  23. [31]

    Robyn Speer, Joshua Chin, and Catherine Havasi. 2017. http://aaai.org/ocs/index.php/AAAI/AAAI17/paper/view/14972 Conceptnet 5.5: An open multilingual graph of general knowledge

    2017

  24. [32]

    Yi Tay, Anh Tuan Luu, Siu Cheung Hui, and Jian Su. 2018. https://doi.org/10.18653/v1/P18-1093 Reasoning with sarcasm by reading in-between . In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1010--1020, Me...

    2018 doi

  25. [33]

    Gomez, ukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, ukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS'17, page 6000–6010, ...

    2017

  26. [34]

    Chien-Yao Wang, I-Hau Yeh, and Hong-Yuan Mark Liao. 2024. https://arxiv.org/abs/2402.13616 Yolov9: Learning what you want to learn using programmable gradient information . Preprint, arXiv:2402.13616

    2024 arXiv

  27. [35]

    Shaowei Yao and Xiaojun Wan. 2020. https://doi.org/10.18653/v1/2020.acl-main.400 Multimodal transformer for multimodal machine translation . In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4346--4350, Online. Association for Co...

    2020 doi

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.