Pith. sign in

REVIEW 3 major objections 5 minor 58 references

GALOT: Generative Active Learning via Optimizable Zero-shot Text-to-image Generation

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read A text-to-image active-learning loop with prompt optimization beats state-of-the-art active learning on three image benchmarks.

desk verdict There's a real, novel combination here, but the paper's marquee claim that text-embedding optimization helps is not supported by its own controlled comparisons. read the letter →

arxiv 2412.16227 v1 pith:SF5RRLVC submitted 2024-12-18 cs.CV cs.LG

classification cs.CVcs.LG
keywords activelearningtext-to-imagegenerationdiffusionmodelspseudo-labelingtext-to-modeltrainingdatasynthesispromptoptimizationimageclassification
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 proposes GALOT, an active-learning loop in which a text-to-image diffusion model generates new training images from text prompts, while an active-learning acquisition function such as entropy, margin, or least confidence is used to optimize those prompts so that the generated images are maximally informative for the current classifier. The generated images carry pseudo-labels taken from the class name in the prompt, so no annotation budget is spent on them. The paper claims that on CIFAR10, CIFAR100, and TinyImageNet, GALOT and its non-optimized basic version consistently outperform twelve state-of-the-art active-learning methods, with average gains of 3.84%, 7.71%, and 3.29%, and that optimized prompts add as much as 5.86% over the basic version on CIFAR10. If true, this turns text-to-image generation into an annotation-free source of informative training data and enables a text-to-model training paradigm.

What carries the argument

The load-bearing object is the optimizable text embedding $s$ of a conditional diffusion model. Starting from a template embedding $s^* = h_\tau(y_i)$ for class $y_i$, GALOT updates $s$ by projected gradient descent, $s_i = s_{i-1} + \alpha\,\mathrm{sgn}[\nabla_s \sigma(x_0, f_\theta)]$ under $\|s_i - s^*\|_2 \le \epsilon$, where $\sigma$ is the active-learning acquisition function. To make this tractable, Proposition 1 collapses the gradient sum over diffusion timesteps to the single term $\nabla_s \sigma(x_0, f_\theta) = T\,\nabla_{x_0}\sigma \cdot J_{x_0,s}$, assuming the forward noising relation $x_t = \sqrt{\alpha_t}x_0 + \sqrt{1-\alpha_t}\epsilon$ also holds across the reverse process, so the gradient is estimated by backpropagating through the last denoising step. The same embedding defines the generated image and its pseudo-label, coupling synthesis and annotation in one loop.

What would settle it

Compare GALOT's one-step gradient updates against random perturbations of the same radius $\epsilon$, and against gradients computed through all $T=50$ denoising steps, at a fixed annotation budget such as 1,000 labels on CIFAR10; if either alternative matches or beats GALOT's accuracy, the Proposition 1 gradient estimate is not the source of the reported gains.

Watch

Extended reading notes

Core claim

The central claim is that the text condition of a pre-trained text-to-image diffusion model can be treated as an optimizable data pool for active learning. GALOT initializes the condition from a task template such as 'a realistic photo of a {label}', scores the images the model generates with an active-learning acquisition function, and moves the text embedding in the direction that maximizes that score, subject to a small distance constraint from the template. The generated image is assigned the class name from the prompt as a pseudo-label, so the synthetic batch costs no annotation budget. On CIFAR10, CIFAR100, and TinyImageNet, the paper reports that this loop outperforms twelve active-learning baselines at every annotation budget, with average gains of 3.84%, 7.71%, and 3.29%, and that the embedding optimization adds up to 5.86% over using the plain template on CIFAR10.

Load-bearing premise

The load-bearing premise is that the gradient of the acquisition function with respect to the text embedding is usefully approximated by backpropagating through only the last reverse diffusion step, with the forward noising equation assumed to describe the reverse process; if that gradient is uninformative, prompt optimization adds little beyond the fixed template.

Editorial extensions

If this is right

  • Active learning can be augmented with unlimited pseudo-labeled synthetic data at zero annotation cost, and Table 1 reports GALOT ahead of every compared active-learning method at every budget.
  • A vision classifier can be trained from text alone: the generated-only baseline reaches 58.63% on CIFAR10, and adding 10,000 real labels reaches 92.46%.
  • The generated data are reusable across architectures; the same pool trained VGG16, DenseNet121, MobileNetV2, and DLA, all above the margin and least-confidence active-learning baselines on average.
  • Pseudo-label accuracy stays effectively perfect for descriptive templates even under text-embedding distortion, so annotation-free training is viable when the prompt template is chosen well.

Reading between the lines

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

  • An implication the paper leaves implicit is that the same loop should transfer to any guided diffusion domain, since the condition $s$ is abstract and pseudo-labels can be read from it.
  • The paper's own ablations leave open that most of the gain comes from adding synthetic data rather than from prompt optimization; a direct test would replace the learned gradient with random same-radius perturbations.
  • A practical extension the paper does not discuss: if the one-step gradient is unreliable, black-box prompt search over text embeddings would preserve the method while working with API-only generation services.
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 / 5 minor

Summary. The paper proposes GALOT, an active learning framework that uses a pre-trained text-to-image diffusion model to synthesize training images from text prompts. The prompts are initialized from class names via a template, then 'optimized' by backpropagating an active-learning acquisition score through the diffusion model; the generated images are pseudo-labeled by the prompt class. Synthetic images are combined with conventionally selected real labeled samples. Experiments on CIFAR-10, CIFAR-100, and TinyImageNet compare against 12 AL baselines, with ablations over templates, perturbation radius epsilon, acquisition functions, and generative batch size. The headline claims are consistent improvements over SOTA (average +3.84, +7.71, and +3.29 percentage points, respectively) and that text optimization improves over the basic version by up to 5.86% on CIFAR-10.

Significance. If the central claims are validated, the paper introduces a promising direction: zero-shot text-to-image generation as an annotation-free source of informative training data for active learning, with potential text-to-model transfer. The strengths include the breadth of the benchmark (three datasets, twelve baselines), the human evaluation of pseudo-label accuracy, the dataset-reuse experiments showing that one generated pool helps several architectures, and the detailed hyperparameter table (Table 8). The strongest credible contribution is the basic version's improvement over traditional AL. However, the 'optimizable text' contribution is not supported: the proof of Proposition 1 is invalid, the headline comparison confounds template choice with optimization, and the only controlled optimization comparison shows a negligible effect. The significance of the paper as a text-optimization method is therefore not yet established.

major comments (3)
  1. [§3.3, Eq. (8)–(10)] Proposition 1 is not proved as stated. Equation (8) sums over all reverse-process timesteps t the product ∇xt σ · Jxt,s. Equation (9) factors Jxt,s as Jx0,xt · Jxt,x0 · Jx0,s and Equation (10) collapses the sum to T∇x0σ · Jx0,s by appealing to Eq. (3). But Eq. (3) describes the forward noising process (xt = √αt x0 + ...), not the reverse denoising transition p(xt−1|xt, s), and Jxt,x0 is not a well-defined Jacobian of the reverse generative map. The implemented gradient, which backpropagates only through the last reverse diffusion step (Section 3.3), is therefore a heuristic rather than a consequence of the proposition. Because Eq. (5) is the basis of the optimizable-text claim, the theoretical derivation must be corrected or the claim explicitly reframed as an empirical heuristic.
  2. [§4.1, Tables 1, 3, 4, 8] The headline GALOT-vs-GALOT(basic) gap is confounded. Table 8 shows that GALOT uses the template 'a realistic photo of a {label}' with epsilon growing linearly to 0.5, whereas GALOT(basic) uses the bare template '{label}' with epsilon = 0. Table 3 shows that the template alone changes the 1,000-budget accuracy from 0.5566 ('{label}') to 0.6074 ('a realistic photo of a {label}'), a 5.08-point gap that is nearly the size of the 5.86-point improvement attributed to text refinement. The same-template comparison in Table 4 (epsilon=0.5 vs. epsilon=0) yields only a 0.22-point average gain over 1k–5k budgets (0.7963 vs. 0.7941), with epsilon=0 better at the 1k budget, and no error bars are provided. Thus the evidence does not demonstrate that text-embedding optimization, rather than template choice, drives GALOT's advantage.
  3. [§4.1, Table 1; Appendix A] The main results are reported without variance estimates; the only uncertainty reported in the paper is SEM over 2 runs (Table 3, Figure 5, Figure 6). Several per-budget differences are small: on TinyImageNet, GALOT and GALOT(basic) differ by at most 1.6 points at 2k–10k, and GALOT is slightly worse than basic at 6k, 8k, and 9k. Without multiple seeds and error bars on the central comparison, the claim of 'consistently outperforms SOTA methods in all settings' is not fully supported, and the small differences between methods in the ablation studies cannot be interpreted reliably.
minor comments (5)
  1. [Abstract] The abstract says 'sorely using the text description'; this should be 'solely'.
  2. [Algorithm 1] Line 4 initializes the labeled pool L = ∅ inside the training loop; this would discard previously labeled data at every cycle, contradicting the cumulative budgets reported in the experiments. The initialization should be moved outside the loop.
  3. [§4.4] The human evaluation of pseudo-labels uses only 10 images per class per setting (100 total per setting in Table 9). This is a small sample; the paper should report confidence intervals or use a larger sample.
  4. [Table 1] There are minor typos in Table 1, including 'RandomSamping' and an apparent missing digit in the BALD row for TinyImageNet at 4000 ('0.251').
  5. [§3.3] The notation in Eq. (7)–(10) confuses the number of diffusion steps with a continuous time index: the sum in Eq. (8) runs from t=0 to T, while T is elsewhere the total number of steps, making the factor T in Eq. (7) ambiguous.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the headline results are external test-set accuracies, and the text-optimization loop is self-referential active learning by design rather than a definitional reduction.

full rationale

GALOT's claimed results are measured on held-out real test sets (CIFAR10, CIFAR100, TinyImageNet), and the labels used during training are pseudo-labels constructed from public class-name text inputs, not from the model's own predictions or from the test labels. The text-embedding optimization in Eq. (2) and Algorithm 1 is a standard active-learning feedback loop: the current classifier's acquisition function guides which synthetic queries are generated, the model is trained on those queries plus the real labeled pool, and performance is then evaluated externally. This is the defining mechanism of active learning, not an identity between an output and its input. There are no self-citations, so no uniqueness theorem or ansatz is imported from the authors' prior work. The invalid proof of Proposition 1 (Eq. (8)-(10), which invokes the forward noising relation Eq. (3) inside the reverse-process Jacobian chain) and the confounded GALOT-versus-basic comparison (Table 8 varies both the template and epsilon) are genuine scientific weaknesses, but they are correctness and experimental-design concerns rather than circular reductions that make the central claim equivalent to its inputs by construction. The comparison to SOTA active-learning baselines is externally falsifiable and does not collapse to a fitted parameter or a self-referential definition; therefore no circularity is found.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The framework has no invented entities; its results are purchased with pretrained Stable Diffusion and CLIP, pseudo-label fidelity, and several tuned hyperparameters. The most consequential free parameter is epsilon, which balances diversity against pseudo-label accuracy. The most fragile axiom is the gradient approximation, whose proof is flawed and whose empirical benefit is small.

free parameters (7)
  • epsilon (text embedding perturbation bound) = 0 to 0.5, linear growth
    Limits how far optimized embedding can move from class template. Chosen after ablation in Table 4; large eps degrades pseudo-label accuracy (Figure 4).
  • alpha (PGD step size) = epsilon/5
    Step size for text embedding updates in Eq (5); scales with epsilon rather than being tuned independently.
  • k (gradient sampling count) = 6
    Number of generated images used to estimate the expectation in Eq (5); hyperparameter from Table 8.
  • n (text update steps) = 10
    Number of PGD iterations for text embedding per AL cycle; from Table 8.
  • text template = a realistic photo of a {label}
    Selected because it outperforms alternatives in Table 3 and retains high pseudo-label accuracy in Figure 4; GALOT(basic) uses bare {label}, confounding template and optimization.
  • BGAL (synthetic samples per cycle) = equal to |L| (cumulative labeled count)
    Chosen to match current labeled pool size; Figure 6 shows changing the ratio to 0.5x or 1.5x has little effect.
  • diffusion steps T = 50
    Default Stable Diffusion sampling steps; not tied to the gradient claim despite appearing in Prop 1's constant factor.
assumptions (5)
  • domain assumption The class-name template generates images that genuinely depict the class, so pseudo-labels are correct.
    Human evaluation in Section 4.4 covers epsilon=0,5,10, not the 0 to 0.5 schedule used in training; at epsilon=5, some templates fall to 60% accuracy.
  • ad hoc to paper Backpropagating through the last reverse diffusion step provides a useful approximation to the gradient in Eq (5).
    Proposition 1's proof is invalid (Eq (8)-(10)); the paper states the Eq (3) assumption is an empirical observation.
  • domain assumption Synthetic images from Stable Diffusion are close enough to real test images that classifiers trained on them transfer.
    Central to GAL-only and GAL-AL results; supported only indirectly by test accuracy, never measured directly.
  • domain assumption Maximizing the AL acquisition score on generated images yields more informative training samples.
    Figure 5 shows random acquisition is competitive, particularly at later cycles, so this assumption is weakly supported.
  • domain assumption CLIP text embeddings of class labels are a suitable optimization space for the constraint ||s-s*||<epsilon.
    Standard CLIP encoding used in T2I pipelines; the paper provides no analysis of whether the epsilon-ball contains meaningful prompts.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GALOT: Generative Active Learning via Optimizable Zero-shot Text-to-image Generation." pith.science (2026). https://pith.science/paper/SF5RRLVC

@misc{pith2026241216227,
  author       = {Pith},
  title        = {Pith review of: GALOT: Generative Active Learning via Optimizable Zero-shot Text-to-image Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SF5RRLVC}},
  note         = {Machine review of arXiv:2412.16227}
}
read the original abstract

Active Learning (AL) represents a crucial methodology within machine learning, emphasizing the identification and utilization of the most informative samples for efficient model training. However, a significant challenge of AL is its dependence on the limited labeled data samples and data distribution, resulting in limited performance. To address this limitation, this paper integrates the zero-shot text-to-image (T2I) synthesis and active learning by designing a novel framework that can efficiently train a machine learning (ML) model sorely using the text description. Specifically, we leverage the AL criteria to optimize the text inputs for generating more informative and diverse data samples, annotated by the pseudo-label crafted from text, then served as a synthetic dataset for active learning. This approach reduces the cost of data collection and annotation while increasing the efficiency of model training by providing informative training samples, enabling a novel end-to-end ML task from text description to vision models. Through comprehensive evaluations, our framework demonstrates consistent and significant improvements over traditional AL methods.

Figures

Figures reproduced from arXiv: 2412.16227 by the authors.

Figure 1
Figure 1. Overview of GALOT. The task-related text is first converted into the text embedding. Then, it iteratively executes 1) optimizing the text embedding according to the AL criteria based on the current model’s output, 2) generating data sam￾ples with optimized text embedding, and 3) train￾ing the model with generated data and pseudo label. GALOT train the vision model from text inputs. Active learning is a pivotal techn… view at source ↗
Figure 2
Figure 2. GALOT Workflow for Each Active Learning Cycle. In each active learning cycle, the data sample is generated using pre-trained T2I models with the embedding s and pseudo label. Optionally combined with the traditional AL with real datasets as a complement, the generated data can be used to train a model. The embedding is then optimized according to the updated model via the gradients of the AL acquisition. Consider a … view at source ↗
Figure 3
Figure 3. Comparison of Different Base￾lines. As is shown in [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Text-to-image Generation Ac￾curacy (Human Evaluated) vs. ϵ with Different Templates. To evaluate the efficacy of pseudo-labels generated by our model, we compute the accuracy of the pseudo-label through human annotation. Specifically, for each tem￾plate, we randomly op…
Figure 6
Figure 6. Figure 6: Accuracy vs BGAL. Sampling Number BGAL vs. Annotation Budgets. In this exper￾iment, we examine the effects of the number of generated samples on the performance of GALOT. We generate 0.5|L|, |L|, and 1.5|L| data samples for GALOT. The results are presented in [PITH_FU…
Figure 7
Figure 7. Figure 7: Visual Comparison of Generated Images VS. Real Images on CIFAR10. [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Generated Images for CIFAR10 Classification [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]
Figure 9
Figure 9. Figure 9: Generated Images for CIFAR100 Classification [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Generated Images for TinyImageNet Classification [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 34 canonical work pages

  1. [1]

    arXiv preprint arXiv:1603.04467 (2016)

    Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G.S., Davis, A., Dean, J., Devin, M., et al.: Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467 (2016)

  2. [2]

    Angluin, D.: Queries and concept learning. Mach. Learn. 2(4), 319–342 (1987). https://doi.org/10.1007/BF00116828, https://doi.org/10.1007/BF00116828

  3. [3]

    arXiv preprint arXiv:1906.03671 (2019)

    Ash, J.T., Zhang, C., Krishnamurthy, A., Langford, J., Agarwal, A.: Deep batch active learning by diverse, uncertain gradient lower bounds. arXiv preprint arXiv:1906.03671 (2019)

  4. [4]

    In: 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Workshop Track Proceedings

    Bachman, P., Sordoni, A., Trischler, A.: Learning algorithms for active learning. In: 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Workshop Track Proceedings. OpenReview.net (2017),https://openreview. net/forum?id=rJj2ZxHtl

  5. [5]

    https://cdn.openai.com/papers/dall-e-3.pdf (2023)

    Betker, J., Goh, G., Jing, L., Brook, T., Wang, J., Li, L., Ouyang, L., Zhuang, J., Lee, J., Guo, Y ., Manassra, W., Dhariwal, P., Chu, C., Jiao, Y ., Ramesh, A.: Improving image generation with better captions. https://cdn.openai.com/papers/dall-e-3.pdf (2023)

  6. [6]

    Cohn, D.A., Atlas, L.E., Ladner, R.E.: Improving generalization with active learning. Mach. Learn. 15(2), 201–221 (1994). https://doi.org/10.1007/BF00993277, https://doi.org/10. 1007/BF00993277

  7. [7]

    IEEE Transactions on Pattern Analysis and Machine Intelligence (2023)

    Croitoru, F.A., Hondru, V ., Ionescu, R.T., Shah, M.: Diffusion models in vision: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence (2023)

  8. [8]

    In: 2009 IEEE conference on computer vision and pattern recognition

    Deng, J., Dong, W., Socher, R., Li, L.J., Li, K., Fei-Fei, L.: Imagenet: A large-scale hierarchical image database. In: 2009 IEEE conference on computer vision and pattern recognition. pp. 248–255. Ieee (2009)

Show all 58 references
  1. [9]

    In: Ranzato, M., Beygelzimer, A., Dauphin, Y .N., Liang, P., Vaughan, J.W

    Ding, M., Yang, Z., Hong, W., Zheng, W., Zhou, C., Yin, D., Lin, J., Zou, X., Shao, Z., Yang, H., Tang, J.: Cogview: Mastering text-to-image generation via transform- ers. In: Ranzato, M., Beygelzimer, A., Dauphin, Y .N., Liang, P., Vaughan, J.W. (eds.) Advances in Neural Info...

  2. [10]

    CoRR abs/1802.09841 (2018), http://arxiv.org/abs/1802.09841

    Ducoffe, M., Precioso, F.: Adversarial active learning for deep networks: a margin based approach. CoRR abs/1802.09841 (2018), http://arxiv.org/abs/1802.09841

  3. [11]

    (No Title) (1965)

    Freeman, L.C.: Elementary applied statistics: for students in behavioral science. (No Title) (1965)

  4. [12]

    In: Precup, D., Teh, Y .W

    Gal, Y ., Islam, R., Ghahramani, Z.: Deep bayesian active learning with image data. In: Precup, D., Teh, Y .W. (eds.) Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, 6-11 August 2017. Proceedings of Machine Learning Rese...

  5. [13]

    In: Pro- ceedings of the 34th International Conference on Machine Learning

    Gal, Y ., Islam, R., Ghahramani, Z.: Deep Bayesian active learning with image data. In: Pro- ceedings of the 34th International Conference on Machine Learning. PMLR (2017)

  6. [14]

    In: Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N.D., Weinberger, K.Q

    Goodfellow, I.J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A.C., Bengio, Y .: Generative adversarial nets. In: Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N.D., Weinberger, K.Q. (eds.) Advances in Neural Information Processing System...

  7. [15]

    In: Larochelle, H., Ranzato, M., Hadsell, R., Balcan, M., Lin, H

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. In: Larochelle, H., Ranzato, M., Hadsell, R., Balcan, M., Lin, H. (eds.) Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020,...

  8. [16]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Huang, G., Liu, Z., Van Der Maaten, L., Weinberger, K.Q.: Densely connected convolutional networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 4700–4708 (2017)

  9. [17]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition workshops

    Kampffmeyer, M., Salberg, A.B., Jenssen, R.: Semantic segmentation of small objects and modeling of uncertainty in urban remote sensing images using deep convolutional neural networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition workshops....

  10. [18]

    Advances in Neural Information Processing Systems 34, 22919–22930 (2021)

    Kim, Y .Y ., Song, K., Jang, J., Moon, I.C.: Lada: Look-ahead data acquisition via augmentation for deep active learning. Advances in Neural Information Processing Systems 34, 22919–22930 (2021)

  11. [19]

    Krizhevsky, A., Hinton, G., et al.: Learning multiple layers of features from tiny images (2009)

  12. [20]

    CS 231N7(7), 3 (2015)

    Le, Y ., Yang, X.: Tiny imagenet visual recognition challenge. CS 231N7(7), 3 (2015)

  13. [21]

    In: Proceedings of the 17th Annual International ACM-SIGIR Conference on Research and Development in Information Retrieval

    Lewis, D.D., Gale, W.A.: A sequential algorithm for training text classifiers. In: Proceedings of the 17th Annual International ACM-SIGIR Conference on Research and Development in Information Retrieval. ACM/Springer (1994)

  14. [22]

    In: Wal- lach, H.M., Larochelle, H., Beygelzimer, A., d’Alché-Buc, F., Fox, E.B., Garnett, R

    Li, B., Qi, X., Lukasiewicz, T., Torr, P.H.S.: Controllable text-to-image generation. In: Wal- lach, H.M., Larochelle, H., Beygelzimer, A., d’Alché-Buc, F., Fox, E.B., Garnett, R. (eds.) Advances in Neural Information Processing Systems 32: Annual Conference on Neural In- form...

  15. [23]

    In: ICBDT 2020: 3rd International Conference on Big Data Technologies, Qingdao, China, September, 2020

    Ma, Y ., Lu, S., Xu, E., Yu, T., Zhou, L.: Combining active learning and data aug- mentation for image classification. In: ICBDT 2020: 3rd International Conference on Big Data Technologies, Qingdao, China, September, 2020. pp. 58–62. ACM (2020). https://doi.org/10.1145/3422713...

  16. [24]

    In: International Conference on Learning Representations (2018)

    Madry, A., Makelov, A., Schmidt, L., Tsipras, D., Vladu, A.: Towards deep learning models resistant to adversarial attacks. In: International Conference on Learning Representations (2018)

  17. [25]

    In: IEEE Winter Conference on Applications of Computer Vision, WACV 2020, Snowmass Village, CO, USA, March 1- 5, 2020

    Mayer, C., Timofte, R.: Adversarial sampling for active learning. In: IEEE Winter Conference on Applications of Computer Vision, WACV 2020, Snowmass Village, CO, USA, March 1- 5, 2020. pp. 3060–3068. IEEE (2020). https://doi.org/10.1109/WACV45572.2020.9093556, https://doi.org/...

  18. [26]

    In: Chaudhuri, K., Jegelka, S., Song, L., Szepesvári, C., Niu, G., Sabato, S

    Nichol, A.Q., Dhariwal, P., Ramesh, A., Shyam, P., Mishkin, P., McGrew, B., Sutskever, I., Chen, M.: GLIDE: towards photorealistic image generation and editing with text-guided diffusion models. In: Chaudhuri, K., Jegelka, S., Song, L., Szepesvári, C., Niu, G., Sabato, S. (eds...

  19. [27]

    Advances in neural information processing systems 32 (2019)

    Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al.: Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems 32 (2019)

  20. [28]

    In: The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023

    Poole, B., Jain, A., Barron, J.T., Mildenhall, B.: Dreamfusion: Text-to-3d using 2d diffu- sion. In: The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023. OpenReview.net (2023),https://openreview.net/pdf? id=FjNys5c7VyY 11

  21. [29]

    In: Meila, M., Zhang, T

    Popov, V ., V ovk, I., Gogoryan, V ., Sadekova, T., Kudinov, M.A.: Grad-tts: A diffusion probabilistic model for text-to-speech. In: Meila, M., Zhang, T. (eds.) Proceedings of the 38th International Conference on Machine Learning, ICML 2021, 18-24 July 2021, Virtual Event. Pro...

  22. [30]

    In: International conference on machine learning

    Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al.: Learning transferable visual models from natural language supervision. In: International conference on machine learning. pp. 8748–8763. PMLR (2021)

  23. [31]

    arXiv preprint arXiv:2204.06125 (2022)

    Ramesh, A., Dhariwal, P., Nichol, A., Chu, C., Chen, M.: Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125 (2022)

  24. [32]

    In: Meila, M., Zhang, T

    Ramesh, A., Pavlov, M., Goh, G., Gray, S., V oss, C., Radford, A., Chen, M., Sutskever, I.: Zero-shot text-to-image generation. In: Meila, M., Zhang, T. (eds.) Proceedings of the 38th International Conference on Machine Learning, ICML 2021, 18-24 July 2021, Virtual Event. Proc...

  25. [33]

    In: Balcan, M., Weinberger, K.Q

    Reed, S.E., Akata, Z., Yan, X., Logeswaran, L., Schiele, B., Lee, H.: Generative adversarial text to image synthesis. In: Balcan, M., Weinberger, K.Q. (eds.) Proceedings of the 33nd International Conference on Machine Learning, ICML 2016, New York City, NY , USA, June 19-24, 2...

  26. [34]

    In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022, New Orleans, LA, USA, June 18-24, 2022

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image synthesis with latent diffusion models. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022, New Orleans, LA, USA, June 18-24, 2022. pp. 10674–10685. IEEE (2022). http...

  27. [35]

    In: NeurIPS (2022), http://papers.nips.cc/paper_files/paper/2022/hash/ ec795aeadae0b7d230fa35cbaf04c041-Abstract-Conference.html

    Saharia, C., Chan, W., Saxena, S., Li, L., Whang, J., Denton, E.L., Ghasemipour, S.K.S., Lopes, R.G., Ayan, B.K., Salimans, T., Ho, J., Fleet, D.J., Norouzi, M.: Photorealistic text-to-image diffusion models with deep language understand- ing. In: NeurIPS (2022), http://papers...

  28. [36]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., Chen, L.C.: Mobilenetv2: Inverted residuals and linear bottlenecks. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 4510–4520 (2018)

  29. [37]

    In: Proceedings of the International Symposium on Intelligent Data Analysis (2001)

    Scheffer, T., Decomain, C., Wrobel, S.: Active hidden Markov models for information extraction. In: Proceedings of the International Symposium on Intelligent Data Analysis (2001)

  30. [38]

    In: 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings

    Sener, O., Savarese, S.: Active learning for convolutional neural networks: A core-set approach. In: 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings. OpenReview.net (2018), https:/...

  31. [39]

    Settles, B.: Active learning literature survey (2009)

  32. [40]

    ACM SIGMOBILE mobile comput- ing and communications review 5(1), 3–55 (2001)

    Shannon, C.E.: A mathematical theory of communication. ACM SIGMOBILE mobile comput- ing and communications review 5(1), 3–55 (2001)

  33. [41]

    arXiv preprint arXiv:1409.1556 (2014)

    Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recogni- tion. arXiv preprint arXiv:1409.1556 (2014)

  34. [42]

    In: Bach, F.R., Blei, D.M

    Sohl-Dickstein, J., Weiss, E.A., Maheswaranathan, N., Ganguli, S.: Deep unsupervised learning using nonequilibrium thermodynamics. In: Bach, F.R., Blei, D.M. (eds.) Proceedings of the 32nd International Conference on Machine Learning, ICML 2015, Lille, France, 6-11 July

  35. [43]

    arXiv preprint arXiv:2302.07944 (2023)

    Trabucco, B., Doherty, K., Gurinas, M., Salakhutdinov, R.: Effective data augmentation with diffusion models. arXiv preprint arXiv:2302.07944 (2023)

  36. [44]

    In: Chaudhuri, K., Salakhutdinov, R

    Tran, T., Do, T., Reid, I.D., Carneiro, G.: Bayesian generative active deep learning. In: Chaudhuri, K., Salakhutdinov, R. (eds.) Proceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA. Pro- ceedings of Mach...

  37. [45]

    IEEE Trans

    Wang, K., Zhang, D., Li, Y ., Zhang, R., Lin, L.: Cost-effective active learning for deep image classification. IEEE Trans. Circuits Syst. Video Technol. 27(12), 2591–2600 (2017). https://doi.org/10.1109/TCSVT.2016.2589879, https://doi.org/10.1109/TCSVT.2016. 2589879

  38. [46]

    In: Avidan, S., Brostow, G.J., Cissé, M., Farinella, G.M., Hassner, T

    Wu, C., Liang, J., Ji, L., Yang, F., Fang, Y ., Jiang, D., Duan, N.: Nüwa: Visual synthesis pre-training for neural visual world creation. In: Avidan, S., Brostow, G.J., Cissé, M., Farinella, G.M., Hassner, T. (eds.) Computer Vision - ECCV 2022 - 17th European Conference, Tel ...

  39. [47]

    In: 2018 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2018, Salt Lake City, UT, USA, June 18- 22, 2018

    Xu, T., Zhang, P., Huang, Q., Zhang, H., Gan, Z., Huang, X., He, X.: Attngan: Fine-grained text to image generation with attentional generative adversarial networks. In: 2018 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2018, Salt Lake City, UT, USA, June 1...

  40. [48]

    ACM Computing Surveys (2022)

    Yang, L., Zhang, Z., Song, Y ., Hong, S., Xu, R., Zhao, Y ., Zhang, W., Cui, B., Yang, M.H.: Diffusion models: A comprehensive survey of methods and applications. ACM Computing Surveys (2022)

  41. [49]

    arXiv preprint arXiv:2304.08821 (2023)

    Yin, Y ., Kaddour, J., Zhang, X., Nie, Y ., Liu, Z., Kong, L., Liu, Q.: Ttida: Controllable generative data augmentation via text-to-text and text-to-image models. arXiv preprint arXiv:2304.08821 (2023)

  42. [50]

    In: IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2019, Long Beach, CA, USA, June 16-20, 2019

    Yoo, D., Kweon, I.S.: Learning loss for active learning. In: IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2019, Long Beach, CA, USA, June 16-20, 2019. pp. 93–

  43. [51]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Yoo, D., Kweon, I.S.: Learning loss for active learning. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 93–102 (2019)

  44. [52]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Yu, F., Wang, D., Shelhamer, E., Darrell, T.: Deep layer aggregation. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 2403–2412 (2018)

  45. [53]

    arXiv preprint arXiv:2203.13450 (2022)

    Zhan, X., Wang, Q., Huang, K.h., Xiong, H., Dou, D., Chan, A.B.: A comparative survey of deep active learning. arXiv preprint arXiv:2203.13450 (2022)

  46. [54]

    In: IEEE International Conference on Computer Vision, ICCV 2017, Venice, Italy, October 22-29, 2017

    Zhang, H., Xu, T., Li, H.: Stackgan: Text to photo-realistic image synthesis with stacked generative adversarial networks. In: IEEE International Conference on Computer Vision, ICCV 2017, Venice, Italy, October 22-29, 2017. pp. 5908–5916. IEEE Computer Society (2017). https://...

  47. [55]

    CoRR abs/1702.07956 (2017), http://arxiv.org/abs/1702.07956

    Zhu, J., Bento, J.: Generative adversarial active learning. CoRR abs/1702.07956 (2017), http://arxiv.org/abs/1702.07956

  48. [56]

    a photo of a {label }

    Zhu, X., Zhang, P., Lin, X., Shi, Y .: Active learning from data streams. In: Proceed- ings of the 7th IEEE International Conference on Data Mining (ICDM 2007), Octo- ber 28-31, 2007, Omaha, Nebraska, USA. pp. 757–762. IEEE Computer Society (2007). https://doi.org/10.1109/ICDM...

  49. [102]

    https://doi.org/10.1109/CVPR.2019.00018, http://openaccess.thecvf.com/content_CVPR_2019/html/Yoo_Learning_Loss_ for_Active_Learning_CVPR_2019_paper.html

    Computer Vision Foundation / IEEE (2019). https://doi.org/10.1109/CVPR.2019.00018, http://openaccess.thecvf.com/content_CVPR_2019/html/Yoo_Learning_Loss_ for_Active_Learning_CVPR_2019_paper.html

  50. [2015]

    JMLR Workshop and Conference Proceedings, vol. 37, pp. 2256–2265. JMLR.org (2015), http://proceedings.mlr.press/v37/sohl-dickstein15.html 12

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.