Just Leaf It: Accelerating Diffusion Classifiers with Hierarchical Class Pruning

Andreas Dengel; Arundhati S. Shanbhag; Brian B. Moser; Federico Raue; Stanislav Frolov; Tobias C. Nauen

REVIEW 6 major objections 6 minor 30 references

Just Leaf It: Accelerating Diffusion Classifiers with Hierarchical Class Pruning

T0 review · 6 major / 6 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 Prune the label tree and diffusion classifiers run 60% faster

desk verdict HDC is a legitimate new pruning scheme for diffusion classifiers with a plausible speed-up, but unreported M and table inconsistencies keep me from trusting the exact numbers. read the letter →

arxiv 2411.12073 v2 pith:UKEUEGME submitted 2024-11-18 cs.CV cs.AIcs.LG

Arundhati S. Shanbhag , Brian B. Moser , Tobias C. Nauen , Stanislav Frolov , Federico Raue , Andreas Dengel This is my paper · ORCID

classification cs.CVcs.AIcs.LG

keywords hierarchicalclassificationdiffusionclassifierszero-shotinferenceaccelerationclasspruningImageNet-1KWordNethierarchyMonteCarlonoiseprediction

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

Zero-shot diffusion classifiers decide an image's label by running a Monte Carlo noise-reconstruction estimate for every candidate class, so cost grows with the number of labels. This paper claims that the same classifier can instead walk a label hierarchy: first score coarse categories such as "animal" or "object," keep only the best-scoring branches, and then run the expensive per-label evaluation only on the surviving leaves. On ImageNet-1K it reports cutting inference time by up to about 60% while roughly preserving accuracy, and with a conservative pruning ratio it reaches 65.16% per-class accuracy versus 64.90% for the baseline while using about 980 seconds instead of 1600. If the claim holds, hierarchical class pruning turns diffusion classification from a brute-force scan into a tunable speed-accuracy dial, which matters because diffusion models otherwise remain too slow for large-scale zero-shot recognition.

What carries the argument

The mechanism is a hierarchical label tree $T_h = (N, E)$ over WordNet synsets, traversed level by level. At each non-leaf level the method computes per-synset errors $\epsilon_n = \mathbb{E}_{t,\varepsilon}\, d(\varepsilon, x_t, c_n)$ with the same noise-prediction distance as the diffusion classifier but a smaller Monte Carlo sample count $M$; then Equation 8 keeps only nodes whose error lies under a threshold set by a pruning ratio $K_d$ (top-k) or within two standard deviations of the minimum. This selected set $S^d_{\text{selected}}$ shrinks as the tree is descended, and the surviving leaves are scored once more with the full classifier to pick $\arg\min_{n \in S^h_{\text{selected}}} \epsilon_n$. The load-bearing component is that coarse synset-level scores are informative enough to rank the true class's ancestors near the top, so greedy pruning does not delete the correct leaf.

What would settle it

Run the pruning stage of HDC on ImageNet-1K with the true class known and record, per image, whether the ancestor path survives to the leaf set; repeat with several Monte Carlo sample counts $M$. If the survival rate of the true class is close to the reported top-1 accuracy, the hierarchy is doing the work; if a large fraction of errors are images whose true synset was pruned early, the speedup is bought by discarding correct answers.

Watch

Extended reading notes

Core claim

The central claim is that the computational bottleneck of the diffusion classifier—scoring every label with epsilon-prediction—can be bypassed without retraining by exploiting tree-structured labels. HDC starts at the root of the WordNet-based ImageNet tree and, at each depth, computes Monte Carlo error scores for the children of the currently selected synsets using fewer samples than the full classifier, then keeps only the top-k (fixed ratio) or the near-minimum-error nodes (dynamic threshold). The surviving leaf nodes are then scored with the standard diffusion classifier, and the minimum-error leaf is the prediction. In the paper's numbers, this preserves or slightly exceeds baseline accuracy: Strategy 1 gives 65.16% per-class accuracy in 980 seconds versus 64.90% in 1600 seconds, and Strategy 2 gives 63.33% in 650 seconds, about a 60% speedup. The paper's further claim is that this is a tunable mechanism for trading precision for speed, and a general method that works with different Stable Diffusion versions and even with a self-generated hierarchy for CIFAR-100.

Load-bearing premise

The method assumes that cheap, noisy error scores at coarse category levels put the true class's parent categories near the top often enough that greedy pruning never deletes the correct leaf before the final scoring step.

Editorial extensions

If this is right

Diffusion classifiers can avoid exhaustive label scans on large taxonomies, changing the per-image cost from scoring every class to scoring only the kept branches.
The pruning ratio $K_d$ becomes a user-facing control: smaller ratios give faster predictions, larger ratios give more accurate ones.
Class sets can be extended or reduced by editing leaf nodes, with no retraining of the diffusion model.
The same pipeline transfers to datasets without a natural hierarchy by generating one, as the paper demonstrates for CIFAR-100.

Reading between the lines

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

We infer that the speedup is bounded by how well the label taxonomy aligns with the diffusion model's error geometry: on ImageNet the confusion matrix is concentrated inside synsets like "Animal," which is exactly the condition that makes coarse scores predictive. A hierarchy that groups unrelated classes under the same parent would likely erase the gain.
We infer that the unstated pruning-stage sample count $M$ is the main knife-edge: low $M$ saves time but adds noise to synset scores, and the paper does not report it, so the reported speed-accuracy trade-off cannot yet be reproduced or predicted on a new dataset.
A testable extension would be to run HDC with a deliberately shuffled or random tree of the same depth on ImageNet; if accuracy stays near baseline, the hierarchy is not the carrier, and if it collapses, the method's value depends on taxonomy quality.

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.

Desk Editor's Note

read the letter

Quick take: the core idea is genuinely new — hierarchical label-tree pruning for diffusion classifiers — and the speed-up is mechanically plausible. But the paper's own tables are inconsistent, and the load-bearing assumption about pruning recall is never directly measured. I'd send it to review, not desk reject, and ask for a revision.

What's actually new: prior acceleration work (Li et al., Clark et al.) used weak filters or bandit elimination. Nobody used a semantic hierarchy to prune candidate classes before the expensive per-class evaluation. The method is simple and clearly explained: compute coarse ε-prediction errors at synset level with fewer Monte Carlo samples, keep top-k or a dynamic threshold, descend the tree, then run the full diffusion classifier on survivors. That's a legitimate new application. The reported 38–59% speed-up is believable given the mechanics — fewer class evaluations. The Strategy 1 accuracy gain of 0.26 pp is small but not absurd, and the CIFAR-100 result (+3.3 pp with Kd=0.4) is interesting.

Soft spots, in order of severity:

1. M for the pruning stage is never reported, even though Algorithm 1 takes M as an explicit input. The entire method depends on coarse synset scores ranking the true class's ancestors near the top. If those scores are noisy, one early mis-prune is irreversible. The paper gives no per-level survival rate of the true leaf. This is the biggest gap and it's fixable with a small ablation.

2. Table inconsistencies. Table 1 says baseline 64.70% and Strategy 1 64.90%; Table 5 says 64.90% and 65.16%. Table 2 lists the SD2.0 Strategy 2 time as 980 s, Table 1 says 650 s. I can't verify the headline accuracy claim when the paper contradicts itself. This looks like sloppy reporting, not fabrication, but it has to be cleaned up.

3. No comparison to prior acceleration baselines. The paper compares only to brute-force diffusion classification. Since Li et al. already proposed a weak-model filter and Clark et al. a bandit elimination, the reader can't tell whether hierarchical pruning is actually better than existing accelerations. The 'up to 60%' claim is relative to vanilla, not to the state of the art.

4. Starting at level 3 because level 2 showed no variation is a post hoc structural choice. It might be fine, but it deserves justification rather than a passing concession.

None of this is fatal. The method is clear, the mechanics are sound, and the code is released. I'd give this a conditional accept: require reporting M, variance across seeds, per-level recall, and consistent tables. The central speed-up claim will likely survive; the accuracy gain is less certain.

Recommendation: send to a serious referee. This is a solid engineering paper with a real gap in evidence. The right outcome is revision, not rejection.

Referee Report

6 major / 6 minor

Summary. The paper proposes Hierarchical Diffusion Classifier (HDC), a training-free modification of diffusion classifiers that traverses a label tree, pruning candidate synsets level by level using Monte Carlo epsilon-prediction errors with a reduced sample count, and then applies the classical diffusion classifier to the surviving leaf nodes. On ImageNet-1K and CIFAR-100 with Stable Diffusion variants, it reports up to roughly 60% inference speed-up while maintaining or slightly improving accuracy (e.g., 65.16% vs. 64.90% for the baseline). The method is a heuristic inference-time acceleration; no training or learned components are introduced.

Significance. If the empirical claims hold, HDC is a practically useful and conceptually simple acceleration for zero-shot diffusion classifiers, with a tunable speed/accuracy trade-off. The paper contributes a clear algorithm and a public code link, though it does not provide machine-checked proofs or formal guarantees. The principal caveat is that the headline quantitative results are not yet verifiable: the pruning-stage Monte Carlo budget is unspecified, no variance or repeated-run information is given, and key tables contradict one another. The method's core assumption that low-sample synset error scores can safely prune the true class is plausible but untested. Overall, the contribution is an engineering insight with initial evaluations, and its significance is moderate pending the missing support.

major comments (6)

[Section 3.2, Algorithm 1] The number M of Monte Carlo samples used in the pruning stage is never reported; the text only says that 'a smaller number of samples' is used after Eq. (7). This M directly determines both runtime and the noise level of the synset error estimates, so the speed-up and the risk of pruning the true class cannot be assessed. Please report M for each experiment and for each tree level, and provide a sensitivity analysis of accuracy and runtime as a function of M.
[Tables 1, 2, and 5] The tables contradict each other on headline numbers. Table 1 reports baseline 64.70% and HDC Strategy 1 64.90% with 980 s; Table 5 reports baseline 64.90% and HDC Strategy 1 65.16% with 980 s; Table 2 reports the SD 2.0 Strategy 2 time as 980 s, while Tables 1 and 5 report 650 s. Also, Table 2 gives SD 2.0 Strategy 2 accuracy as 63.33/63.20, whereas Table 5 lists 63.33 and Table 1 lists 63.20. These inconsistencies make it impossible to verify the central quantitative claim; please unify the tables or explicitly explain which configuration each row refers to.
[Section 3.2, Eqs. (7)-(9)] The core assumption that low-M synset error scores rank the ancestors of the true leaf highly enough for greedy top-k pruning to be safe is never directly tested. Report the per-level survival rate (recall) of the true class's ancestors and the final leaf for both pruning strategies; without this, an early mis-prune at level d is irreversible because Eq. (9) only searches within S^h_selected. This is the load-bearing premise of the claimed speed-up.
[Section 3.3] The decision to start traversal at level 3 because 'starting at level 2 showed no variation in error scores but increased inference time' is a post hoc structural choice made on the evaluation data. This should be validated on a held-out split or supported by a quantitative comparison of accuracy and per-level recall for level-2 versus level-3 starts; as written, the choice undermines the claim that the hierarchy configuration is principled rather than tuned to the test set.
[Section 3.3] The details of the modified WordNet hierarchy are not reproducible from the text: phrases such as 'choosing synset labels with more definite meanings' and 'we further modify the existing WordNet tree' are not accompanied by a description of the edits, the number of nodes per level, or the final mapping. Please provide the exact hierarchy or a script to generate it, and similarly for the CIFAR-100 self-generated tree, since the speed-up mechanism depends directly on tree structure.
[Section 5] All accuracy figures are reported without variance or number of repeats. The claimed 0.20 to 0.26 percentage-point improvement over baseline is within plausible run-to-run noise for diffusion classification, so it is not established as significant. Report standard deviations across at least three runs, or explicitly state whether these are single-run values.

minor comments (6)

[Section 3.3] The text contains a typo: 'sysnet labels' should be 'synset labels'.
[Table 3] The table header does not clearly separate the Strategy 1 columns from the Strategy 2 columns, and the meaning of the three Kd values in the header is not explained in the caption.
[Eq. (5)] The typesetting of Eq. (5), '1 / NCP j=1', is unclear; please use standard summation notation to avoid confusion about whether the normalization includes the number of classes.
[Table 4] The prompt strings in Table 4 are missing spaces inside the angle brackets, e.g., 'A photo of a<class label>' should be 'A photo of a <class label>'.
[Section 4.2] The sentence 'the timesteps uniformly sampled from [1, 1000]' is grammatically incomplete; it should read 'the timesteps are uniformly sampled from [1, 1000]'.
[Section 3.3, References] The hierarchy is attributed to Engstrom et al. [10], which is a software library; please also cite the underlying ImageNet WordNet hierarchy or provide a URL to the exact hierarchy file used, since the robustness library's hierarchy may not exactly match the paper's tree.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: HDC's hierarchical pruning uses the same epsilon-prediction error as the baseline and is not defined in terms of its own outputs.

full rationale

The paper's central claim is that hierarchical pruning over WordNet synsets reduces the number of diffusion classifier evaluations while preserving accuracy. The scoring function used in the pruning stage (Eq. 7) is the same epsilon-prediction error defined in Eq. 3 and estimated via Monte Carlo in Eq. 6; it is not defined in terms of the final predicted class, so the pruning step is not self-definitional. The final prediction (Eq. 9) is also the same error score restricted to the surviving leaf nodes, which is a legitimate inference-time heuristic rather than a circular derivation. The only potentially load-bearing assumptions are that the reduced Monte Carlo sample count M in the pruning stage is sufficient to rank synsets reliably and that the chosen hierarchy and level-3 starting point are appropriate; these are empirical validity and hyperparameter-selection concerns, not circularity. The paper's self-citations ([11], [12], [22], [23]) are background references on generative models and diffusion-based methods and do not supply a uniqueness theorem or the core pruning mechanism, so no self-citation chain carries the argument. No fitted parameter is renamed as a prediction: M and Kd are algorithmic settings, and the accuracy improvements are reported as direct measurements on ImageNet-1K and CIFAR-100 rather than quantities implied by construction. Internal numerical inconsistencies between tables affect verifiability, but they do not constitute a circular step. Therefore, no specific circular reduction can be exhibited, and the appropriate finding is no significant circularity.

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

The method adds no new entities; it assumes the baseline diffusion score is meaningful, that WordNet groupings align with diffusion error geometry, and that cheap Monte Carlo estimates are stable enough for pruning. These assumptions are unvalidated in the paper.

free parameters (5)

Pruning ratio Kd (Strategy 1) = 0.5 for ImageNet; 0.4, 0.5, 0.75 tested on CIFAR-100
Determines how many child nodes survive at each tree level via Eq. 8; no sensitivity analysis on ImageNet, yet headline numbers depend on it.
Dynamic pruning threshold (Strategy 2) = 2 standard deviations from minimum error
Nodes within 2 sigma are kept; hyperparameter chosen without ablations.
Monte Carlo sample count M in pruning stage = not reported
The paper says fewer samples are used than in the classical pipeline, but never states M, so the trade-off cannot be reproduced.
ImageNet tree depth = 7 levels
Depth set by authors to balance speed and accuracy; affects number of evaluation passes.
Starting tree level = level 3
Level 2 was dropped because it showed no variation in error scores; this choice is post hoc and saves runtime.

assumptions (5)

domain assumption Uniform prior p(ci)=1/NC over classes, cancelling in the posterior (Eqs. 1-2).
Standard diffusion classifier assumption inherited from Li et al.; needed to turn likelihoods into posteriors.
domain assumption Class likelihood is well approximated by the noise-prediction distance d(epsilon,x,c) via ELBO (Eqs. 3-6).
Baseline score used throughout; HDC does not re-validate it.
standard math Shared Monte Carlo sample set across conditions yields consistent error estimates (Eq. 6).
Fixed sample set makes error rankings comparable across classes.
ad hoc to paper Error scores at coarse synset levels are predictive of leaf-level correctness, so top-down pruning is safe (Eqs. 7-8).
This is the core heuristic of HDC; true only if WordNet grouping aligns with the diffusion model's error geometry. Paper's Limitations section admits dependence on tree quality.
ad hoc to paper Reduced-M estimates in the pruning stage remain reliable enough to retain the true class.
Speed-up relies on cheaper estimation, but M is unreported and stability is not analyzed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Just Leaf It: Accelerating Diffusion Classifiers with Hierarchical Class Pruning." pith.science (2026). https://pith.science/paper/UKEUEGME

@misc{pith2026241112073,
  author       = {Pith},
  title        = {Pith review of: Just Leaf It: Accelerating Diffusion Classifiers with Hierarchical Class Pruning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UKEUEGME}},
  note         = {Machine review of arXiv:2411.12073}
}

read the original abstract

Diffusion models, celebrated for their generative capabilities, have recently demonstrated surprising effectiveness in image classification tasks by using Bayes' theorem. Yet, current diffusion classifiers must evaluate every label candidate for each input, creating high computational costs that impede their use in large-scale applications. To address this limitation, we propose a Hierarchical Diffusion Classifier (HDC) that exploits hierarchical label structures or well-defined parent-child relationships in the dataset. By pruning irrelevant high-level categories and refining predictions only within relevant subcategories (leaf nodes and sub-trees), HDC reduces the total number of class evaluations. As a result, HDC can speed up inference by as much as 60% while preserving and sometimes even improving classification accuracy. In summary, our work provides a tunable control mechanism between speed and precision, making diffusion-based classification more feasible for large-scale applications.

Figures

Figures reproduced from arXiv: 2411.12073 by the authors.

**Figure 2.** Overview of our Hierarchical Diffusion Classifier (HDC). Starting with an input image [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Visualization of the ImageNet1K hierarchy, illustrat [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Example classification of an image in the pruning stage of the HDC using Strategy 1. In this stage, the error scores of each node are used to iteratively prune the tree, narrowing it down to relevant leaf nodes that will undergo further refinement in subsequent stages…

**Figure 5.** Figure 5: Confusion Matrix of HDC (Strategy 1) for the sub [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 18 canonical work pages

[1]

Un- constrained open vocabulary image classification: Zero-shot transfer from text to image via clip inversion

Philipp Allgeuer, Kyra Ahrens, and Stefan Wermter. Un- constrained open vocabulary image classification: Zero-shot transfer from text to image via clip inversion. arXiv preprint arXiv:2407.11211, 2024. 1, 3, 5

arXiv 2024
[2]

A self explanatory review of decision tree classifiers

Anuradha and Gaurav Gupta. A self explanatory review of decision tree classifiers. In International Conference on Recent Advances and Innovations in Engineering (ICRAIE- 2014), pages 1–7, 2014. 4

work page 2014
[3]

Multidiffusion: Fusing diffusion paths for controlled image generation

Omer Bar-Tal, Lior Yariv, Yaron Lipman, and Tali Dekel. Multidiffusion: Fusing diffusion paths for controlled image generation. 2023. 1

2023
[4]

Improving image generation with better captions

James Betker, Gabriel Goh, Li Jing, Tim Brooks, Jianfeng Wang, Linjie Li, Long Ouyang, Juntang Zhuang, Joyce Lee, Yufei Guo, et al. Improving image generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf, 2(3):8, 2023. 1

2023
[5]

Your diffusion model is secretly a certifiably robust classifier

Huanran Chen, Yinpeng Dong, Shitong Shao, Zhongkai Hao, Xiao Yang, Hang Su, and Jun Zhu. Your diffusion model is secretly a certifiably robust classifier. arXiv preprint arXiv:2402.02316, 2024. 1

arXiv 2024
[6]

Diffusion models are certifiably robust classifiers

Huanran Chen, Yinpeng Dong, Shitong Shao, Hao Zhongkai, Xiao Yang, Hang Su, and Jun Zhu. Diffusion models are certifiably robust classifiers. NeurIPS, 37:50062–50097, 2024. 1

work page 2024
[7]

Text-to-image diffusion mod- els are zero-shot classifiers, 2023

Kevin Clark and Priyank Jaini. Text-to-image diffusion mod- els are zero-shot classifiers, 2023. 1, 2

work page 2023
[8]

Imagenet: A large-scale hierarchical image database

Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 2, 4

2009

Show all 30 references

[9]

Diffusion models beat gans on image synthesis

Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural information processing systems, 34:8780–8794, 2021. 2

2021
[10]

Robustness (python library),

Logan Engstrom, Andrew Ilyas, Hadi Salman, Shibani San- turkar, and Dimitris Tsipras. Robustness (python library),
[11]

Adversarial text-to-image synthesis: A review

Stanislav Frolov, Tobias Hinz, Federico Raue, J ¨orn Hees, and Andreas Dengel. Adversarial text-to-image synthesis: A review. Neural Networks, 144:187–209, 2021. 1, 2

2021
[12]

Spotd- iffusion: A fast approach for seamless panorama generation over time

Stanislav Frolov, Brian B Moser, and Andreas Dengel. Spotd- iffusion: A fast approach for seamless panorama generation over time. arXiv preprint arXiv:2407.15507, 2024. 1

2024 arXiv
[13]

Predictability and surprise in large generative models

Deep Ganguli, Danny Hernandez, Liane Lovitt, Amanda Askell, Yuntao Bai, Anna Chen, Tom Conerly, Nova Das- sarma, Dawn Drain, Nelson Elhage, et al. Predictability and surprise in large generative models. In 2022 ACM Conference on Fairness, Accountability, and Transparency, 2022. 2

2022
[14]

Generative adversarial networks

Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. In NeurIPS,
[15]

To recognize shapes, first learn to generate images

Geoffrey E Hinton. To recognize shapes, first learn to generate images. Progress in brain research, 165:535–547, 2007. 1

2007
[16]

Denoising dif- fusion probabilistic models

Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. NeurIPS, 33:6840–6851, 2020. 1

2020
[17]

Random decision forests

Tin Kam Ho. Random decision forests. In Proceedings of 3rd International Conference on Document Analysis and Recognition, volume 1, pages 278–282 vol.1, 1995. 4

1995
[18]

Learning multiple layers of features from tiny images

Alex Krizhevsky. Learning multiple layers of features from tiny images. Technical report, University of Toronto, 2009. 4

2009
[19]

Your diffusion model is secretly a zero-shot classifier

Alexander C Li, Mihir Prabhudesai, Shivam Duggal, Ellis Brown, and Deepak Pathak. Your diffusion model is secretly a zero-shot classifier. In ICCV, pages 2206–2217, 2023. 1, 2, 6

2023
[20]

Repaint: Inpainting using denoising diffusion probabilistic models

Andreas Lugmayr, Martin Danelljan, Andres Romero, Fisher Yu, Radu Timofte, and Luc Van Gool. Repaint: Inpainting using denoising diffusion probabilistic models. In CVPR,
[21]

A survey of de- cision trees: Concepts, algorithms, and applications

Ibomoiye Domor Mienye and Nobert Jere. A survey of de- cision trees: Concepts, algorithms, and applications. IEEE Access, 12:86716–86727, 2024. 4

2024
[22]

Latent dataset distillation with diffusion models

Brian B Moser, Federico Raue, Sebastian Palacio, Stanislav Frolov, and Andreas Dengel. Latent dataset distillation with diffusion models. arXiv preprint arXiv:2403.03881, 2024. 1, 2

2024
[23]

Diffusion models, image super-resolution and everything: A survey

Brian B Moser, Arundhati S Shanbhag, Federico Raue, Stanislav Frolov, Sebastian Palacio, and Andreas Dengel. Diffusion models, image super-resolution and everything: A survey. arXiv preprint arXiv:2401.00736, 2024. 1

2024 arXiv
[24]

Learning transferable visual models from natural language supervision, 2021

Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision, 2021. 6

2021
[25]

Variational inference with normalizing flows

Danilo Rezende and Shakir Mohamed. Variational inference with normalizing flows. pages 1530–1538. PMLR, 2015. 1 9

2015
[26]

High-resolution image synthesis with latent diffusion models

Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. InCVPR, pages 10684– 10695, 2022. 2, 3, 6

2022
[27]

Pho- torealistic text-to-image diffusion models with deep language understanding

Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L Denton, Kamyar Ghasemipour, Raphael Gontijo Lopes, Burcu Karagol Ayan, Tim Salimans, et al. Pho- torealistic text-to-image diffusion models with deep language understanding. NeurIPS, 35:36479–36494, 2022. 2

2022
[28]

Laion- 5b: An open large-scale dataset for training next generation image-text models, 2022

Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Wortsman, Patrick Schramowski, Srivatsa Kundurthy, Katherine Crowson, Lud- wig Schmidt, Robert Kaczmarczyk, and Jenia Jitsev. Lai...

2022
[29]

Uncovering the disentanglement capability in text-to-image diffusion models

Qiucheng Wu, Yujian Liu, Handong Zhao, Ajinkya Kale, Trung Bui, Tong Yu, Zhe Lin, Yang Zhang, and Shiyu Chang. Uncovering the disentanglement capability in text-to-image diffusion models. In CVPR, pages 1900–1910, 2023. 1

1900
[30]

Adding conditional control to text-to-image diffusion models

Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In ICCV, pages 3836–3847, 2023. 1 10

2023

Pith tools

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

[1] [1]

Un- constrained open vocabulary image classification: Zero-shot transfer from text to image via clip inversion

Philipp Allgeuer, Kyra Ahrens, and Stefan Wermter. Un- constrained open vocabulary image classification: Zero-shot transfer from text to image via clip inversion. arXiv preprint arXiv:2407.11211, 2024. 1, 3, 5

arXiv 2024

[2] [2]

A self explanatory review of decision tree classifiers

Anuradha and Gaurav Gupta. A self explanatory review of decision tree classifiers. In International Conference on Recent Advances and Innovations in Engineering (ICRAIE- 2014), pages 1–7, 2014. 4

work page 2014

[3] [3]

Multidiffusion: Fusing diffusion paths for controlled image generation

Omer Bar-Tal, Lior Yariv, Yaron Lipman, and Tali Dekel. Multidiffusion: Fusing diffusion paths for controlled image generation. 2023. 1

2023

[4] [4]

Improving image generation with better captions

James Betker, Gabriel Goh, Li Jing, Tim Brooks, Jianfeng Wang, Linjie Li, Long Ouyang, Juntang Zhuang, Joyce Lee, Yufei Guo, et al. Improving image generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf, 2(3):8, 2023. 1

2023

[5] [5]

Your diffusion model is secretly a certifiably robust classifier

Huanran Chen, Yinpeng Dong, Shitong Shao, Zhongkai Hao, Xiao Yang, Hang Su, and Jun Zhu. Your diffusion model is secretly a certifiably robust classifier. arXiv preprint arXiv:2402.02316, 2024. 1

arXiv 2024

[6] [6]

Diffusion models are certifiably robust classifiers

Huanran Chen, Yinpeng Dong, Shitong Shao, Hao Zhongkai, Xiao Yang, Hang Su, and Jun Zhu. Diffusion models are certifiably robust classifiers. NeurIPS, 37:50062–50097, 2024. 1

work page 2024

[7] [7]

Text-to-image diffusion mod- els are zero-shot classifiers, 2023

Kevin Clark and Priyank Jaini. Text-to-image diffusion mod- els are zero-shot classifiers, 2023. 1, 2

work page 2023

[8] [8]

Imagenet: A large-scale hierarchical image database

Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 2, 4

2009