Pith. sign in

REVIEW 5 major objections 5 minor 40 references

Perception of Visual Content: Differences Between Humans and Foundation Models

T0 review · 5 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read On 1,886 Dollar Street images, machine captions beat human labels at region classification, but human annotations still win where context is subtle.

desk verdict Solid empirical comparison of human vs machine annotations on Dollar Street, but the headline income-regression claim rests on a significance test against the wrong baseline. read the letter →

arxiv 2411.18968 v3 pith:N4Q3RHA2 submitted 2024-11-28 cs.CV cs.LG

Nardiena A. Pratama , Shaoyang Fan , Gianluca Demartini This is my paper · ORCID
classification cs.CVcs.LG
keywords humanannotationmachine-generatedfoundationmodelsimagecaptioningobjectdetectionsimilarityalgorithmicbiasDollarStreetdataset
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 asks whether machine-generated image annotations can replace human-written ones, and it answers with a qualified no. Comparing crowdsourced human labels, object-detection labels, and image captions on 1,886 images from the Dollar Street dataset, the paper finds that machine captions and human labels are the most similar pair of annotations, while object labels are more distant. When the annotations are used to train predictive models, machine captions alone achieve the best overall region-classification score, and object labels combined with captions achieve the lowest income-prediction error. Yet human labels remain more effective for non-action categories such as living rooms and places for dinner, so the paper concludes that machine annotations are useful but not yet a full replacement for human ones.

What carries the argument

The central mechanism is a comparison-and-prediction pipeline built on sentence embeddings. All three annotation sets—object labels from Faster R-CNN with a ResNet-50-FPN backbone, captions from BLIP with a Vision Transformer backbone, and crowdsourced human labels—are embedded by a fine-tuned Sentence Transformer model into 384-dimensional vectors; cosine similarity between these vectors measures annotation similarity, and the same vectors train a RUSBoost classifier for region prediction and an AdaBoost.R2 regressor for income prediction.

What would settle it

Re-run both predictive experiments with fresh random splits of the data each time rather than the same 80/20 split, and recompute the significance tests; if the differences between best and worst annotation models no longer reach significance, the central reliability claim fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that human and machine annotations of the same images overlap in vocabulary and sentence structure but differ in what they emphasize, and that neither source is uniformly better for downstream prediction. At the similarity level, ML Captions and Human Labels show the highest mean cosine similarity (0.69), while ML Objects sits lower (0.49 to 0.50); all three annotation types show consistent cross-region patterns, with European images receiving the strongest agreement. At the prediction level, ML Captions yield the best overall region-classification F1 score (0.41), and ML Objects combined with ML Captions yield the best overall income-regression RMSE (1817.49). The paper reports these gaps as statistically significant and interprets the category-level results as evidence that ML annotations work best for action categories, human labels work best for non-action categories, and combining annotation types often gives more consistent predictions.

Load-bearing premise

The reliability claims assume that the 10 runs with different random seeds are independent samples, even though every run reuses the same test images, so the reported significance levels likely overstate how much the results would generalize.

Editorial extensions

If this is right

  • Region-classification models trained on ML captions alone can outperform those trained on human labels or object labels, suggesting captions carry the most region-discriminative signal.
  • For income prediction, combining object labels with captions gives the lowest overall error, but no single annotation type dominates every region or image category.
  • ML annotations are strongest for action-based categories, while human labels remain stronger for static or context-dependent categories such as living rooms and places for dinner.
  • Human and machine annotations show consistent cross-region patterns, so replacing human labels with machine ones does not by itself remove geographic bias.
  • Combining annotation types often yields more consistent predictions, such as the lower variance seen for African samples, than relying on a single source.

Reading between the lines

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

  • A testable extension: retrain the captioning and detection models on geographically balanced training data and check whether the European agreement advantage disappears, which would point to training-data geography rather than universal perception.
  • Because every run reuses the same 80/20 test split, the reported significance levels treat correlated samples as independent; re-running with fresh splits would likely widen the error bars and produce more conservative p-values.
  • The category-level pattern suggests a practical division of labor: use machine annotations for scalable coverage and add human labels selectively for categories where ML error is systematically higher.
  • The categories where humans win are exactly those requiring inference about material conditions, so an automated system that switches annotation strategy by category could approximate human quality at lower cost.
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

5 major / 5 minor

Summary. This paper compares human-generated annotations (MTurk labels) with machine-generated annotations (Faster R-CNN object labels and BLIP captions) on 1,886 Dollar Street images spanning eight everyday-life categories. RQ1 measures annotation similarity via fine-tuned Sentence Transformer embeddings, cosine similarity, t-SNE, and annotation lengths. RQ2 trains a RUSBoost classifier to predict geographical region and an AdaBoost.R2 regressor to predict income from the annotation embeddings, using single and combined annotation sets. The main reported findings are that ML Captions and Human Labels are the most similar annotation pair, that ML Captions give the best overall region classification F1 (0.41), that ML Objects plus ML Captions give the lowest income RMSE (1817.49), and that human input is more effective for non-action image categories. The authors conclude that human annotations remain important and are not yet replaceable by machine annotations.

Significance. If the results were fully supported, the paper would provide a useful empirical datapoint on the substitutability of human and machine annotations in a socio-economically diverse image setting, and it would offer a new crowdsourced annotation dataset. Strengths include the release of code and data, the use of a real-world dataset with regional and income diversity, and the coverage of multiple annotation types and combinations. The main limitations are statistical: the income-regression 'best overall' claim is not backed by a significance test against the nearest competitors, the reported tests assume independence across runs that share a single test split, and the predictive tasks lack random or constant baselines. In addition, the text embedding model is fine-tuned on the very annotation texts that are later used for similarity and prediction, which complicates interpretation. These issues are fixable, so the contribution is defensible after revision.

major comments (5)
  1. [Experiment 2] The claim that 'ML Objects and ML Captions performed best overall' for income regression is not statistically supported. The only significance test reported for this task is a Kolmogorov-Smirnov test comparing the best model (ML Objects & ML Captions, RMSE 1817.49 ± 37.13) with the worst model (ML Captions & Human Labels, RMSE 2006.36 ± 58.58). This does not test the difference that matters for the headline claim, namely the gap to the nearest human-inclusive competitors: Human Labels alone (1842.57 ± 14.99) and all three annotations (1842.89 ± 30.83). Using the reported 10-run means and standard deviations, an unpaired Welch t-test of best versus Human Labels gives p ≈ 0.07, and best versus all-annotations gives p ≈ 0.11; neither is significant at the conventional 0.05 level. Please re-run the significance analysis against all pairwise competitors (or report simultaneous confidence intervals) and adjust the claim accordingly.
  2. [Experiment 2] The significance tests in Experiment 2 assume that the 10 runs with different random seeds are independent samples. However, all runs use the same 80/20 train/test split, so the F1 and RMSE values are correlated across runs. Welch's t-test and the Kolmogorov-Smirnov test are therefore anti-conservative, and the reported p-values (1.15e-7 and p < 0.001) overstate the reliability of the performance differences. A paired test across seeds, or better, multiple independent test splits, should be used. At minimum, the correlation should be acknowledged and the conclusions stated with appropriate caveats.
  3. [Experiment 2] The predictive experiments lack baselines. The region classification task has four roughly balanced classes, so a random classifier would achieve about 0.25 F1; the best reported overall F1 is 0.41, which is only modestly above chance. More importantly, some per-category scores fall below chance: for example, the Front Doors category with Human Labels reports F1 = 0.18 in Table 3. Without a random or majority-class baseline, the reader cannot interpret such values. For income regression, the RMSE of a constant predictor (e.g., predicting the mean income in the training set) should be reported as a reference point, because RMSE values around 1800–2000 are meaningful only relative to such a baseline. Please add these baselines and discuss them.
  4. [Data Preparation and Preprocessing] The Sentence Transformer model ('all-MiniLM-L12-v2') is fine-tuned on the three annotation sets before the embeddings are computed. These same embeddings are then used both for the RQ1 similarity analysis and as features for the RQ2 prediction tasks. This is a self-referential step: the representation used for measuring similarity and for prediction has been adapted to the very texts being compared and predicted. Furthermore, because the fine-tuning uses all annotation texts, it includes texts that are later placed in the test split for RQ2, which is a form of leakage at the feature-extractor level. The authors should justify this design choice, compare with off-the-shelf embeddings, and/or fine-tune only on the training split within a cross-validation loop.
  5. [Abstract and Conclusion] The conclusion that 'human-generated annotations are yet to be replaceable' is stronger than what the experiments show. The paper tests whether annotation embeddings are predictive of region and income, not whether human annotations can be replaced in a general annotation pipeline. The results may show complementarity in these two proxy tasks (e.g., human labels help for some non-action categories), but that is not the same as non-replaceability. I recommend softening the claim to something like 'human and machine annotations are complementary in the studied tasks' or explicitly defining what 'replaceable' would mean operationally.
minor comments (5)
  1. [Data Annotation] The enumerated image categories incorrectly list 'washing clothes' twice and omit 'Drying Clothes'; the earlier dataset description should be used to fix this list.
  2. [Results and Discussion, RQ1] The text states that similarity scores are 'generated by the BLIP model's cosine similarity function,' but the methodology clearly uses the Sentence Transformer for embeddings and cosine similarity; this appears to be an error and should be corrected.
  3. [Results and Discussion, RQ2] In the description of Figure 5, please clarify what the line of best fit is fitted against when the x-axis is on a log scale and the y-axis is linear, and state whether the RMSE values in Table 4 are computed on the original or log-transformed income scale.
  4. [Results and Discussion, RQ2] The text refers to 'the ML Captions-Human Model' when describing the worst-performing regression model; this should be 'ML Captions & Human Labels Model' for consistency with Table 4.
  5. [Limitations] The Limitations paragraph contains only a single sentence about generalizability. It should also mention the statistical concerns (shared test split, missing baselines, leakage from fine-tuning) so the reader is aware of these caveats without having to infer them from the experimental details.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's empirical comparisons and predictive evaluations are self-contained, and no fitted parameter is renamed as a prediction.

full rationale

The paper's central claims are empirical comparisons: RQ1 measures similarity between annotation sets via sentence embeddings, and RQ2 measures predictive performance of models trained on annotation embeddings to predict region and income. The target variables (region, income) come from the Dollar Street dataset metadata and are not derived from the annotation embeddings. The predictive models are evaluated on a fixed 80/20 split, and no model parameter is fitted to the test targets; the reported F1 and RMSE values are measured outcomes rather than constructionally forced values. The fine-tuning of the Sentence Transformer on the three annotation texts adapts the representation to the input distribution, which could influence similarity magnitudes, but the RQ1 conclusions are comparative within that representation space and are not definitionally determined; moreover, RQ2 does not depend on any quantity fitted to the targets. The self-citations present (e.g., Fan et al. 2022 and Demartini et al. 2024, both involving the authors) appear in related-work and motivation sections and are not load-bearing for the novel empirical results. Concerns about the significance-test choice (comparing best versus worst models rather than best versus nearest competitor) are statistical-inference issues, not circularity. Accordingly, no circular step can be exhibited from the paper's own equations or construction, and the correct finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new theoretical entities. It relies on a hand-chosen confidence threshold and fine-tuning hyperparameters that affect dataset composition and embeddings, plus domain assumptions about model validity, annotation quality, and statistical independence. These are the main inputs that the central claims depend on beyond the data itself.

free parameters (3)
  • ML object confidence threshold = 0.50
    Images were selected only if ML Objects were detected with at least 50% confidence, discarding 1,155 samples. This threshold is chosen by hand and affects the dataset composition and hence all downstream comparisons.
  • Fine-tuning epochs = 5
    The Sentence Transformer was fine-tuned for 5 epochs on the annotation sets; this setting is not justified and influences the embeddings used in all analyses.
  • Fine-tuning batch size = 16
    Batch size 16 used in fine-tuning; not justified and affects embedding quality.
assumptions (4)
  • domain assumption Pre-trained models (Faster R-CNN, BLIP, Sentence Transformer) produce valid annotations and embeddings for this comparison.
    The paper relies on these models without validating their outputs against ground truth for this dataset, yet the conclusions depend on their quality.
  • domain assumption The 80/20 train/test split is representative and the 10 runs provide independent samples for significance testing.
    The test set is the same across runs, so the runs are not independent; this assumption is load-bearing for the statistical claims.
  • domain assumption The crowdsourced human annotations are of sufficient quality to represent human perception.
    The paper post-processed annotations but does not report inter-annotator agreement, so the reliability of the human labels is unmeasured.
  • domain assumption The absence of stop word removal and lemmatization in preprocessing preserves semantic content needed for similarity.
    The authors acknowledge this choice makes syntactic differences more pronounced, which may drive the t-SNE clustering and similarity scores.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Perception of Visual Content: Differences Between Humans and Foundation Models." pith.science (2026). https://pith.science/paper/N4Q3RHA2

@misc{pith2026241118968,
  author       = {Pith},
  title        = {Pith review of: Perception of Visual Content: Differences Between Humans and Foundation Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N4Q3RHA2}},
  note         = {Machine review of arXiv:2411.18968}
}
read the original abstract

Human-annotated content is often used to train machine learning (ML) models. However, recently, language and multi-modal foundational models have been used to replace and scale-up human annotator's efforts. This study explores the similarity between human-generated and ML-generated annotations of images across diverse socio-economic contexts (RQ1) and their impact on ML model performance and bias (RQ2). We aim to understand differences in perception and identify potential biases in content interpretation. Our dataset comprises images of people from various geographical regions and income levels, covering various daily activities and home environments. ML captions and human labels show highest similarity at a low-level, i.e., types of words that appear and sentence structures, but all annotations are consistent in how they perceive images across regions. ML Captions resulted in best overall region classification performance, while ML Objects and ML Captions performed best overall for income regression. ML annotations worked best for action categories, while human input was more effective for non-action categories. These findings highlight the notion that both human and machine annotations are important, and that human-generated annotations are yet to be replaceable.

Figures

Figures reproduced from arXiv: 2411.18968 by the authors.

Figure 1
Figure 1. Example images annotated with the pre-trained Faster R-CNN model. Left: [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Screenshot of the MTurk annotation interface showing the task instructions and image assessment layout. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. t-SNE visualisation of sentence embeddings of [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Pairwise relationships between similarity scores. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Income Regression Model using ML Objects and Captions as Annotations Grouped by Continent. The x-axis repre [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 28 canonical work pages

  1. [1]

    Baeza-Yates, R. 2020. Bias in Search and Recommender Systems. In Proceedings of the 14th ACM Conference on Recommender Systems, RecSys '20, 2. New York, NY, USA: Association for Computing Machinery. ISBN 9781450375832

    work page 2020

  2. [2]

    Y.; Saligrama, V.; and Kalai, A

    Bolukbasi, T.; Chang, K.-W.; Zou, J. Y.; Saligrama, V.; and Kalai, A. T. 2016. Man is to Computer Programmer as Woman is to Homemaker ? Debiasing Word Embeddings . In Advances in Neural Information Processing Systems , volume 29. Curran Associates, Inc

    work page 2016

  3. [3]

    J.; and Narayanan, A

    Caliskan, A.; Bryson, J. J.; and Narayanan, A. 2017. Semantics derived automatically from language corpora contain human-like biases. Science, 356(6334): 183--186. Publisher: American Association for the Advancement of Science

    work page 2017

  4. [4]

    Cao, Z.; Chen, E.; Huang, Y.; Shen, S.; and Huang, Z. 2023. Learning from Crowds with Annotation Reliability. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '23, 2103–2107. New York, NY, USA: Association for Computing Machinery. ISBN 9781450394086

    work page 2023

  5. [5]

    Chen, X.; Fang, H.; Lin, T.; Vedantam, R.; Gupta, S.; Doll \' a r, P.; and Zitnick, C. L. 2015. Microsoft COCO Captions: Data Collection and Evaluation Server. CoRR, abs/1504.00325

    arXiv 2015

  6. [6]

    Chhikara, G.; Ghosh, K.; Ghosh, S.; and Chakraborty, A. 2023. Fairness for both Readers and Authors: Evaluating Summaries of User Generated Content. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '23, 1996–2000. New York, NY, USA: Association for Computing Machinery. ISBN 9781450394086

    work page 2023

  7. [7]

    Dash, A.; Shandilya, A.; Biswas, A.; Ghosh, K.; Ghosh, S.; and Chakraborty, A. 2019. Summarizing User-generated Textual Content: Motivation and Methods for Fairness in Algorithmic Summaries. Proc. ACM Hum.-Comput. Interact., 3(CSCW)

    work page 2019

  8. [8]

    Demartini, G.; Roitero, K.; and Mizzaro, S. 2024. Data Bias Management. Communications of the ACM, 67(1): 28–32

    work page 2024

Show all 40 references
  1. [9]

    Devillers, L.; Vidrascu, L.; and Lamel, L. 2005. Challenges in real-life emotion annotation and machine learning based detection. Neural Networks, 18(4): 407--422

    2005

  2. [10]

    Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; Uszkoreit, J.; and Houlsby, N. 2020. An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. CoRR, abs/2010.11929

    2020 arXiv

  3. [11]

    Drucker, H. 1997. Improving Regressors Using Boosting Techniques. Proceedings of the 14th International Conference on Machine Learning

    1997

  4. [12]

    Espejel, O. U. 2022. blog/how-to-train-sentence-transformers.md at main · huggingface/blog --- github.com. https://github.com/huggingface/blog/blob/main/how-to-train-sentence-transformers.md. Accessed: 2024-09-13

    2022

  5. [13]

    Fan, S.; Barlas, P.; Christoforou, E.; Otterbacher, J.; Sadiq, S.; and Demartini, G. 2022. Socio- Economic Diversity in Human Annotations . In Proceedings of the 14th ACM Web Science Conference 2022 , WebSci '22, 98--109. New York, NY, USA: Association for Computing Machinery....

    2022

  6. [14]

    FORCE11 . 2020. The FAIR Data principles. https://force11.org/info/the-fair-data-principles/. Accessed: 2024-09-14

    2020

  7. [15]

    Gapminder. 2021. Dollar Street - photos as data to kill country stereotypes. https://www.gapminder.org/dollar-street. Accessed: 2024-06-10

    2021

  8. [16]

    W.; Wallach, H.; Iii, H

    Gebru, T.; Morgenstern, J.; Vecchione, B.; Vaughan, J. W.; Wallach, H.; Iii, H. D.; and Crawford, K. 2021. Datasheets for datasets. Communications of the ACM, 64(12): 86--92

    2021

  9. [17]

    Ghandi, T.; Pourreza, H.; and Mahyar, H. 2023. Deep learning approaches on image captioning: A review. ACM Computing Surveys, 56(3): 1--39

    2023

  10. [18]

    L.; Choi, Y.; and Smith, N

    Kasai, J.; Sakaguchi, K.; Dunagan, L.; Morrison, J.; Bras, R. L.; Choi, Y.; and Smith, N. A. 2021. Transparent human evaluation for image captioning. arXiv preprint arXiv:2111.08940

    2021 arXiv

  11. [19]

    Li, J.; Li, D.; Xiong, C.; and Hoi, S. 2022. BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation. In ICML

    2022

  12. [20]

    Li, Y.; Xie, S.; Chen, X.; Doll \' a r, P.; He, K.; and Girshick, R. B. 2021. Benchmarking Detection Transfer Learning with Vision Transformers. CoRR, abs/2111.11429

    2021 arXiv

  13. [21]

    J.; Bourdev, L

    Lin, T.; Maire, M.; Belongie, S. J.; Bourdev, L. D.; Girshick, R. B.; Hays, J.; Perona, P.; Ramanan, D.; Doll \' a r, P.; and Zitnick, C. L. 2014. Microsoft COCO: Common Objects in Context. CoRR, abs/1405.0312

    2014 arXiv

  14. [22]

    Mohammed, A.; and Kora, R. 2023. A comprehensive review on ensemble deep learning: Opportunities and challenges. Journal of King Saud University - Computer and Information Sciences, 35(2): 757--774

    2023

  15. [23]

    Mohammed, H.; Tannouche, A.; and Ounejjar, Y. 2022. Weed Detection in Pea Cultivation with the Faster RCNN ResNet 50 Convolutional Neural Network. Revue d'Intelligence Artificielle, 36(1)

    2022

  16. [24]

    Reimers, N. 2021. sentence-transformers/all- M ini L M - L 12-v2 · H ugging F ace --- huggingface.co. https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2. Accessed: 2024-09-13

    2021

  17. [25]

    Rekabsaz, N.; Kopeinik, S.; and Schedl, M. 2021. Societal Biases in Retrieved Contents: Measurement Framework and Adversarial Mitigation of BERT Rankers. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '...

    2021

  18. [26]

    Ren, S.; He, K.; Girshick, R.; and Sun, J. 2017. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6): 1137--1149

    2017

  19. [27]

    Sarhan, H.; and Hegelich, S. 2023. Understanding and evaluating harms of AI-generated image captions in political images. Frontiers in Political Science, 5: 1245684

    2023

  20. [28]

    M.; Van Hulse, J.; and Napolitano, A

    Seiffert, C.; Khoshgoftaar, T. M.; Van Hulse, J.; and Napolitano, A. 2008. RUSBoost: Improving classification performance when training data is skewed. In 2008 19th International Conference on Pattern Recognition, 1--4

    2008

  21. [29]

    Sharifi Noorian, S.; Qiu, S.; Gadiraju, U.; Yang, J.; and Bozzon, A. 2022. What Should You Know? A Human-In-the-Loop Approach to Unknown Unknowns Characterization in Image Recognition. In Proceedings of the ACM Web Conference 2022, WWW '22, 882–892. New York, NY, USA: Associat...

    2022

  22. [30]

    Sharma, H.; and Padha, D. 2023. A comprehensive survey on image captioning: from handcrafted to deep learning-based techniques, a taxonomy and open research issues. Artificial Intelligence Review, 56(11): 13619--13661

    2023

  23. [31]

    Sun, W.; Nasraoui, O.; and Shafto, P. 2020. Evolution and impact of bias in human and machine learning algorithm interaction. Plos one, 15(8): e0235502

    2020

  24. [32]

    H.; and Hauska, H

    Swain, P. H.; and Hauska, H. 1977. The decision tree classifier: Design and potential. IEEE Transactions on Geoscience Electronics, 15(3): 142--147

    1977

  25. [33]

    Thomas, P.; Spielman, S.; Craswell, N.; and Mitra, B. 2023. Large language models can accurately predict searcher preferences. arXiv preprint arXiv:2309.10621

    2023 arXiv

  26. [34]

    Valizadegan, H.; Nguyen, Q.; and Hauskrecht, M. 2013. Learning classification models from multiple experts. Journal of Biomedical Informatics, 46(6): 1125--1135. Special Section: Social Media Environments

    2013

  27. [35]

    van Atteveldt, W.; van der Velden, M. A. C. G.; and Boukes, M. 2021. The Validity of Sentiment Analysis : Comparing Manual Annotation , Crowd - Coding , Dictionary Approaches , and Machine Learning Algorithms . Communication Methods and Measures, 15(2): 121--140. Publisher: Ro...

    2021

  28. [36]

    van der Maaten, L.; and Hinton, G. 2008. Visualizing Data using t-SNE. Journal of Machine Learning Research, 9(86): 2579--2605

    2008

  29. [37]

    Xu, L.; Tang, Q.; Lv, J.; Zheng, B.; Zeng, X.; and Li, W. 2023. Deep image captioning: A review of methods, trends and future challenges. Neurocomputing, 546: 126287

    2023

  30. [38]

    Zhang, Z.; Singh, J.; Gadiraju, U.; and Anand, A. 2021. Dissonance Between Human and Machine Understanding. CoRR, abs/2101.07337

    2021 arXiv

  31. [39]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  32. [40]

    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 gl...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.