Pith. sign in

REVIEW 4 major objections 5 minor 44 references

Multi-aspect Knowledge Distillation with Large Language Model

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Adding yes/no aspect answers distilled from a multimodal LLM's logits improves image classification on every dataset and architecture tested.

desk verdict Simple, broadly effective distillation recipe; the accuracy gains are consistent, but the knowledge-transfer mechanism is under-supported and needs a proper control. read the letter →

arxiv 2501.13341 v4 pith:YY4RM2NJ submitted 2025-01-23 cs.CV

Taegyeong Lee , Jinsik Bang , Soyeong Kwon , Taehwan Kim This is my paper · ORCID
classification cs.CV
keywords multi-aspectknowledgedistillationmultimodallargelanguagemodelsimageclassificationfine-grainedlogitbinarycross-entropylossoutputdimensionexpansionobjectdetection
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

The paper proposes a training-time addition for image classifiers: before training, it asks a large language model to generate yes/no questions about visual, categorical, and environmental aspects of each dataset's classes (e.g., 'Does the car have a convertible roof?'). It then feeds each training image and those questions to a frozen multimodal LLM (InternVL2-8B), extracts the softmax probability of the 'yes' token for each question, and expands the classifier's output head to predict those probabilities alongside its class logits. The total loss is the usual cross-entropy on class logits plus a weighted binary cross-entropy on the aspect probabilities. The paper's central claim is that this multi-aspect knowledge distillation improves classification accuracy over the cross-entropy baseline for every dataset and architecture tested, with gains from +0.72 to +6.24 points, and that the gains grow when training data is scarce. The authors argue this matters because it transfers abstract and complex knowledge — not just labels — into the model at almost no training cost, since the multimodal teacher is frozen and queried offline.

What carries the argument

The engine of the method is the multi-aspect knowledge distillation loss $L_{\mathrm{MaKD}}$ (Eq. 5): a binary cross-entropy loss between the student's aspect-logit predictions and the MLLM's softmaxed yes-probabilities. This loss is attached to an expanded output head of dimension $D = C + Q$ (Eq. 2), so the same backbone that classifies also predicts per-image answers to $Q$ yes/no aspect questions. The targets come from a two-step pipeline: GPT-4o generates $N$ candidate yes/no questions from the class list, an LLM ranks and filters them down to $Q$, and InternVL2-8B, prompted with the image and each question, supplies the yes/no token logits whose softmax $q_i$ (Eq. 1) becomes the target. The paper also tunes a weighting factor $\alpha$ on the MaKD loss and finds performance degrades only at very large values (e.g., 50), showing the auxiliary loss is fairly robust to its weight.

What would settle it

A decisive test is to replace the MLLM's aspect targets with the same targets randomly permuted across classes (or with answers deliberately falsified for a subset of questions) and retrain: if accuracy stays at the same level, the gain is not from the content of the aspect knowledge. A complementary check is to evaluate the method on a fine-grained dataset with human-annotated attributes, such as CUB attributes, and measure how well the MLLM's yes/no probabilities match the attribute ground truth; if the correlation is near zero, the method cannot be transferring accurate aspect knowledge, and any improvement would have to be attributed to the auxiliary task structure alone.

Watch

Extended reading notes

Core claim

The central discovery the paper argues for is that a frozen multimodal LLM's yes/no probabilities on aspect questions are a transferable supervision signal for image classification, even though the same MLLM has poor zero-shot accuracy on fine-grained class labels (e.g., 10.27% on CUB200). The method works by distilling the MLLM's softmaxed yes-logits through an expanded output head: the model outputs $D = C + Q$ values, where the first $C$ are class logits trained with cross-entropy and the remaining $Q$ are aspect logits trained with binary cross-entropy against targets $q_i = e^{z_y}/(e^{z_y}+e^{z_n})$ extracted from the MLLM. The paper reports consistent improvements across four architectures and eight datasets, including fine-grained sets like StanfordCars and CUB200 and coarse-grained sets like Caltech101 and Mini-ImageNet, and shows the aspect head also helps when combined with standard knowledge distillation and when transferred to a Faster R-CNN object detector. It further demonstrates through t-SNE and per-sample probability comparisons that the student's aspect logits track the MLLM's aspect logits, which the authors take as evidence that the distilled aspect knowledge is learned rather than ignored.

Load-bearing premise

The load-bearing premise is that the MLLM's yes/no probability answers to the generated aspect questions are correct and informative enough to supervise the student; the paper manually inspects the questions for hallucination but never checks the MLLM's answers against ground truth, so the measured gains could partly come from regularization rather than from accurate aspect knowledge.

Editorial extensions

If this is right

  • Adding the MaKD loss to cross-entropy improves image classification accuracy over baseline for every tested dataset and architecture, with gains between +0.72 and +6.24 points (Table 1).
  • The improvement is larger when training data is scarce: with 40% of StanfordCars, ResNet18 gains +24.01 points over baseline (Table 5).
  • The aspect head composes with existing logit-distillation methods: MaKD adds gains on top of both KD and DKD (Table 4).
  • The same output-expansion recipe carries over to object detection, giving small but consistent AP improvements on MS-COCO with Faster R-CNN-FPN (Table 6).
  • Distilling aspect probabilities rather than class logits sidesteps the MLLM's poor fine-grained zero-shot classification, since aspect questions are easier to answer than fine-grained class labels.

Reading between the lines

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

  • If the gain is genuine, the softmaxed yes/no probability of an MLLM can serve as a dense 'semantic attribute' target, suggesting the same trick could be extended to other dense prediction tasks such as semantic segmentation or depth estimation by attaching an aspect head.
  • Because the method only needs an offline, frozen MLLM, it acts like a cheap way to inject common-sense priors into any vision backbone; a testable prediction is that the gain should shrink as the student's capacity or the amount of training data grows, since the aspect loss then becomes redundant with the label signal.
  • The paper never verifies the MLLM's answers against ground truth, so an important open question is whether the benefit comes from the correctness of the aspect knowledge or merely from the multi-task auxiliary loss; comparing against human-attribute labels or against targets from a weak captioning model would separate these two explanations.
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

4 major / 5 minor

Summary. The paper proposes Multi-aspect Knowledge Distillation (MaKD), a training procedure for image classifiers that appends Q auxiliary output units to the class logits and trains them with binary cross-entropy against yes/no aspect targets. The aspect questions are generated by GPT-4o from the dataset class list, and the targets are the softmax yes-token probabilities extracted from InternVL2-8B for each image. The total loss combines the usual cross-entropy class loss with an alpha-weighted BCE aspect loss (Eq. 6). Experiments on six fine-grained and two coarse-grained datasets with four architectures report consistent accuracy gains over the cross-entropy baseline (Table 1), along with ablations (Tables 2-3, Figures 2-3), extensions to traditional knowledge distillation and reduced training data (Tables 4-5), an object detection extension (Table 6), and qualitative analyses (Figures 4-6).

Significance. If the gains indeed come from the content of the multi-aspect supervision, the method is a simple and broadly applicable contribution: it requires no change to the student architecture beyond widening the final layer, and the evaluation across eight datasets, four architectures, reduced data, and detection is a genuine strength. The paper also includes an explicit limitation discussion in Sec. 6. However, the current experiments do not yet isolate the proposed mechanism. The Rand ablation in Table 2 changes both the semantic content and the marginal distribution of the targets, and the accuracy of the MLLM's yes/no answers is never measured, so the central claim that aspect knowledge, rather than an auxiliary-training regularizer, drives the improvements is not fully established.

major comments (4)
  1. [Sec. 4.3, Table 2] The Rand control replaces the MLLM targets with Gaussian random logits, which simultaneously changes the semantic content and the marginal distribution of the targets. The observed gap between Rand and Ours could therefore be due to differences in target sharpness or entropy rather than to image-specific aspect knowledge. Please add a control that preserves the marginal statistics of the MLLM targets, for example by permuting the MLLM targets across images within each question or by using a fixed set of sharp targets, and report whether the gain persists. Without such a control, the statement that the multi-aspect logits "represent knowledge from various aspects" is not supported by the evidence presented.
  2. [Sec. 4.1, Figures 2-3] The final values of the hyperparameters Q (number of aspects) and alpha (loss weight) are not reported for any dataset or model, even though the text states that Q is tuned per dataset and neural network and Figure 3 shows clear sensitivity to alpha. This is a reproducibility problem and it also leaves open the possibility that part of the reported gains comes from per-dataset hyperparameter selection rather than from the method itself. Please report the chosen Q and alpha values, the search ranges, and the corresponding standard deviations across the three runs.
  3. [Sec. 4.1 and Sec. 5] The paper only manually reviews the generated questions and does not validate the MLLM's yes/no answers against any ground truth. Since InternVL2-8B has low zero-shot classification accuracy on the fine-grained datasets (Table 1), its aspect judgments may be noisy or systematically biased. The t-SNE visualizations and probability comparisons in Figures 5-6 show that the student matches the teacher, but any distillation method would produce such a match, so these figures do not validate the teacher's correctness. Please measure the teacher's aspect-answer accuracy on a labeled sample (or against an oracle), or show that the gains survive when the targets are replaced by an equally sharp but semantically meaningless distribution. This is necessary to support the claimed knowledge-transfer mechanism.
  4. [Tables 1-7] All tables report averages over three runs but no standard deviations or significance tests are reported. Several improvements are small, for example +0.72 accuracy points on FGVC-Aircraft with EfficientNet in Table 1 and +0.23 AP on MS-COCO with MobileNet-V2 in Table 6. Without variance estimates or significance tests, the abstract's claim that "in all experimental results, our method improves the performance of the baselines" cannot be assessed for the smaller entries. Please include standard deviations or confidence intervals, and ideally a paired test for the smallest gains.
minor comments (5)
  1. [Table 5 caption] The caption of Table 5 reads "Extension to class logit distillation with MLLM on Caltech101," but the table shows reduced-data experiments on StanfordCars, OxfordPets, and Caltech101; this appears to be a copy-paste error and should be corrected.
  2. [Sec. 3.4, Eq. (5)] The notation around Eq. (5) is confusing: q is defined as the softmax yes probability from the MLLM in Sec. 3.2, but the text says "where y-hat represents the predicted probability from MLLM," which seems to refer to the student's prediction rather than the teacher's. Please clarify which quantities are teacher targets and which are student outputs.
  3. [Sec. 4.4, object detection] The sentence "we add features to the backbone network of Faster R-CNN-FPN" is vague; please specify where the auxiliary aspect head is attached, how its loss is combined with the detector losses, and how the Q=50 aspect questions were generated for MS-COCO.
  4. [Figure 3] The x-axis labels of Figure 3 are difficult to read (the tick labels appear garbled), and the alpha values used in the experiments should be listed explicitly in the caption or text rather than only appearing in the plot.
  5. [Sec. 6, Limitation] The limitation paragraph acknowledges that performance "may be affected by the multi-aspect questions in MLLMs, making generalization across datasets challenging." This concern is directly relevant to the open control issue in Table 2 and should be addressed with the proposed additional experiments rather than deferred entirely to future work.

Circularity Check

1 steps flagged · score 1.0 of 10

Central accuracy result is grounded in an external MLLM; only a minor tautological validation appears in Sec. 5.3.

  1. self definitional [Section 5.3 (Fig. 5; cf. Eq. 5 and Eq. 6)]
    "As shown in Figure 5, our model demonstrates that the aspect logits of our model exhibit a similar trend to the aspect logits of the MLLM in both fine-grained datasets and coarse-grained datasets. These results indicate that our method can effectively distill various knowledge about the dataset by utilizing the multi-aspect logits extracted from the MLLM."

    Eq. 5 defines L_MaKD as binary cross-entropy between the student's aspect logits and the MLLM's q_i targets, and Eq. 6 adds alpha*L_MaKD to the training loss. Minimizing this loss directly drives the student's aspect outputs toward q_i, so the observed student/MLLM agreement in Fig. 5 is the optimization objective restated rather than an independent confirmation that multi-aspect knowledge was transferred. This tautological validation is not load-bearing for the main classification gains, which are measured externally in Table 1, but it is presented as evidence of the knowledge-transfer mechanism.

full rationale

The central claim that adding the multi-aspect distillation loss L_MaKD to cross-entropy training improves test accuracy is not circular: the aspect targets q_i are produced by a frozen external MLLM (InternVL2-8B) and questions generated by GPT-4o, not by the student or by the authors' equations. The student's expanded output head is trained to match these external targets, and accuracy is measured on held-out test sets against independent baselines. No fitted parameter is renamed as a prediction, and no load-bearing self-citation appears. The only circular-adjacent passage is Sec. 5.3/5.4, where student/MLLM aspect-logit agreement is presented as evidence of knowledge transfer; since Eq. 5 explicitly optimizes that agreement, this particular validation is true by construction. It is not load-bearing for the accuracy result, which stands on Table 1, so the overall circularity score is 1.

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

The central claim rests on the reliability of externally generated question and answer signals, not on a mathematical derivation. The main controllable choices (Q, alpha) are tuned per dataset and not fully reported.

free parameters (3)
  • Q (number of multi-aspect questions) = not reported, tuned per dataset and model
    The paper states 'we tune and select the number of multi-aspect questions based on the dataset and neural network' (Sec 4.1). This introduces validation-set selection that affects reported accuracy.
  • alpha (MaKD loss weight) = not reported, swept from 0 to 50 in Fig 3
    The total loss L = L_CE + alpha L_MaKD (Eq. 6). The final alpha per dataset/model is not listed, so the numbers in Table 1 could reflect a tuned alpha.
  • N (initial question count) = 100
    The paper sets N=100 questions at generation time (Sec 4.1). This is a fixed design choice rather than a fitted value, but it is chosen by hand.
assumptions (3)
  • domain assumption The softmax of the yes/no logits of InternVL2-8B provides an accurate probability that the image satisfies each aspect.
    Used to define the distillation targets q_i in Eq. 1. The paper never validates these probabilities against ground-truth aspect labels; it only shows the MLLM has poor zero-shot class accuracy (Table 1).
  • domain assumption The LLM-generated yes/no questions capture aspects that are useful for class discrimination and are free of hallucination.
    Question generation and selection rely on GPT-4o outputs with manual review (Sec 4.1). The later t-SNE and histogram analyses (Figs 4-6) illustrate but do not prove that the chosen questions are uniquely useful.
  • domain assumption The student model has enough capacity in the added Q output heads to distill the aspect knowledge without interfering with class logits.
    The architecture change (Eq. 2) assumes the extra heads can learn the aspects and that this auxiliary task does not hurt class logit learning; the Rand ablation supports this, but the mechanism is not analyzed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-aspect Knowledge Distillation with Large Language Model." pith.science (2026). https://pith.science/paper/YY4RM2NJ

@misc{pith2026250113341,
  author       = {Pith},
  title        = {Pith review of: Multi-aspect Knowledge Distillation with Large Language Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YY4RM2NJ}},
  note         = {Machine review of arXiv:2501.13341}
}
read the original abstract

Recent advancements in deep learning have significantly improved performance on computer vision tasks. Previous image classification methods primarily modify model architectures or add features, and they optimize models using cross-entropy loss on class logits. Since they focus on classifying images with considering class labels, these methods may struggle to learn various \emph{aspects} of classes (e.g., natural positions and shape changes). Rethinking the previous approach from a novel view, we propose a multi-aspect knowledge distillation method using Multimodal Large Language Models (MLLMs). Our approach involves: 1) querying Large Language Model with multi-aspect questions relevant to the knowledge we want to transfer to the model, 2) extracting corresponding logits from MLLM, and 3) expanding the model's output dimensions to distill these multi-aspect logits. We then apply cross-entropy loss to class logits and binary cross-entropy loss to multi-aspect logits. Through our method, the model can learn not only the knowledge about visual aspects but also the abstract and complex aspects that require a deeper understanding. We primarily apply our method to image classification, and to explore the potential for extending our model, such as object detection. In all experimental results, our method improves the performance of the baselines. Additionally, we analyze the effect of multi-aspect knowledge distillation. These results demonstrate that our method can transfer knowledge about various aspects to the model and the aspect knowledge can enhance model performance in computer vision tasks.

Figures

Figures reproduced from arXiv: 2501.13341 by the authors.

Figure 1
Figure 1. Multi-aspect question generation and logit extraction. For multi-aspect question generation (a), we generate various aspect questions from the LLM by using the class and prompt as instructions. For logit extraction about multi-aspect questions (b), we input the generated multi-aspect questions along with the image into the MLLM to extract logits and obtain the probabilities corresponding to yes token. For multi-aspe… view at source ↗
Figure 2
Figure 2. Ablation study on the number of multi-aspect questions. The x-axis represents the number of aspects (0 represents the baseline model), while the y-axis indicates the accuracy. We run each experiment three times and report the average results. Res18 Res34 Mb-N1 EffiNet KL 82.52 82.63 84.94 87.27 Rand 79.36 81.04 83.39 86.65 Ours 83.38 84.33 85.43 88.07 [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. presents the performance of our method with dif￾ferent weights to the multi-aspect logit loss on Stanford￾Cars and Caltech101. The x-axis represents the weights α (0 means the baselines), while the y-axis indicates the accu￾racy. Our method, based on α, demonstrates improvements in the performances of all baseline models. Additionally, we empirically find that the performance decreases when α value reaches 50. Effec… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visualization of the average logit distribution for classes related to aspects. The x-axis represents the classes, and the y-axis represents the mean of the aspect probability distribution from MLLM in the dataset. 1) Aspect : Is the animal's coat thick and woolly? 2) …
Figure 5
Figure 5. Figure 5: Visualization of t-SNE embeddings for the datasets by aspects. Ours is t-SNE visualizations of the aspect logits from our model (ResNet18), while MLLM is t-SNE visualizations of the aspect logits from the MLLM (InternVL2-8B). The yellow points indicate that the probabi…
Figure 6
Figure 6. Figure 6: Comparison of probability values for multi-aspect questions. We compare the probability values of our model with those of the MLLM for multi-aspect questions. Our model shows similar probability values to MLLM across various multi-aspect questions on test datasets. 5.4…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 25 canonical work pages

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ah- mad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 ,

    arXiv

  2. [2]

    Flamingo: a visual language model for few-shot learning

    Jean-Baptiste Alayrac, Jeff Donahue, Pauline Luc, Antoine Miech, Iain Barr, Yana Hasson, Karel Lenc, Arthur Men- sch, Katherine Millican, Malcolm Reynolds, et al. Flamingo: a visual language model for few-shot learning. Advances in neural information processing systems, 35:23716–23736,

  3. [3]

    Towards language models that can see: Computer vision through the lens of natural lan- guage

    William Berrios, Gautam Mittal, Tristan Thrush, Douwe Kiela, and Amanpreet Singh. Towards language models that can see: Computer vision through the lens of natural lan- guage. arXiv preprint arXiv:2306.16410, 2023. 2

    arXiv 2023

  4. [4]

    Internvl: Scaling up vision foundation mod- els and aligning for generic visual-linguistic tasks

    Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, et al. Internvl: Scaling up vision foundation mod- els and aligning for generic visual-linguistic tasks. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24185–24198, 2024. 2, 4, 5, 8

    work page 2024

  5. [5]

    Describing textures in the wild

    Mircea Cimpoi, Subhransu Maji, Iasonas Kokkinos, Sammy Mohamed, and Andrea Vedaldi. Describing textures in the wild. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3606–3613, 2014. 1, 4

    work page 2014

  6. [6]

    Learning gener- ative visual models from few training examples: An incre- mental bayesian approach tested on 101 object categories

    Li Fei-Fei, Rob Fergus, and Pietro Perona. Learning gener- ative visual models from few training examples: An incre- mental bayesian approach tested on 101 object categories. In 2004 conference on computer vision and pattern recognition workshop, pages 178–178. IEEE, 2004. 1, 4, 6

    work page 2004

  7. [7]

    From im- ages to textual prompts: Zero-shot visual question answer- ing with frozen large language models

    Jiaxian Guo, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Boyang Li, Dacheng Tao, and Steven Hoi. From im- ages to textual prompts: Zero-shot visual question answer- ing with frozen large language models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10867–10877, 2023. 1

    work page 2023

  8. [8]

    Class attention transfer based knowledge distillation

    Ziyao Guo, Haonan Yan, Hui Li, and Xiaodong Lin. Class attention transfer based knowledge distillation. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11868–11877, 2023. 4

    work page 2023

Show all 44 references
  1. [9]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 1

    2016

  2. [10]

    3d object representations for fine-grained categorization

    Jonathan Krause, Michael Stark, Jia Deng, and Li Fei-Fei. 3d object representations for fine-grained categorization. In Proceedings of the IEEE international conference on com- puter vision workshops , pages 554–561, 2013. 1, 4, 5, 6, 7

    2013

  3. [11]

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. In In- ternational conference on machine learning , pages 19730– 19742. PMLR, 2023. 1

    2023

  4. [12]

    Feature pyra- mid networks for object detection

    Tsung-Yi Lin, Piotr Doll ´ar, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature pyra- mid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 2117–2125, 2017. 6

    2017

  5. [13]

    Improved baselines with visual instruction tuning

    Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. Improved baselines with visual instruction tuning. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 26296–26306, 2024. 1, 5

    2024

  6. [14]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. Advances in neural information processing systems, 36, 2024. 2

    2024

  7. [15]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, pages 10012–10022, 2021. 1

    2021

  8. [16]

    Annealing- based label-transfer learning for open world object detection

    Yuqing Ma, Hainan Li, Zhange Zhang, Jinyang Guo, Shang- hang Zhang, Ruihao Gong, and Xianglong Liu. Annealing- based label-transfer learning for open world object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11454–11463...

    2023

  9. [17]

    Fine-grained visual classi- fication of aircraft

    Subhransu Maji, Esa Rahtu, Juho Kannala, Matthew Blaschko, and Andrea Vedaldi. Fine-grained visual classi- fication of aircraft. arXiv preprint arXiv:1306.5151 , 2013. 4

    2013 arXiv

  10. [18]

    Visual classification via description from large language models

    Sachit Menon and Carl V ondrick. Visual classification via description from large language models. arXiv preprint arXiv:2210.07183, 2022. 2

    2022 arXiv

  11. [19]

    Automated flower classification over a large number of classes

    Maria-Elena Nilsback and Andrew Zisserman. Automated flower classification over a large number of classes. In 2008 Sixth Indian conference on computer vision, graphics & im- age processing, pages 722–729. IEEE, 2008. 4, 7

    2008

  12. [20]

    Chils: Zero-shot image classification with hierarchical label sets

    Zachary Novack, Julian McAuley, Zachary Chase Lipton, and Saurabh Garg. Chils: Zero-shot image classification with hierarchical label sets. In International Conference on Machine Learning, pages 26342–26362. PMLR, 2023. 1

    2023

  13. [21]

    Dinov2: Learning robust visual features with- out supervision, 2024

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mah- moud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michae...

    2024

  14. [22]

    Cats and dogs

    Omkar M Parkhi, Andrea Vedaldi, Andrew Zisserman, and CV Jawahar. Cats and dogs. In 2012 IEEE conference on computer vision and pattern recognition, pages 3498–3505. IEEE, 2012. 1, 4, 7

    2012

  15. [23]

    What does a platypus look like? generating customized prompts for zero-shot image classification

    Sarah Pratt, Ian Covert, Rosanne Liu, and Ali Farhadi. What does a platypus look like? generating customized prompts for zero-shot image classification. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 15691–15701, 2023. 2

    2023

  16. [24]

    Optimization as a model for few-shot learning

    Sachin Ravi and Hugo Larochelle. Optimization as a model for few-shot learning. In International conference on learn- ing representations, 2016. 4

    2016

  17. [25]

    Faster r-cnn: Towards real-time object detection with region proposal networks

    Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster r-cnn: Towards real-time object detection with region proposal networks. IEEE transactions on pattern analysis and machine intelligence, 39(6):1137–1149, 2016. 6

    2016

  18. [26]

    Human attention in fine-grained classification

    Yao Rong, Wenjia Xu, Zeynep Akata, and Enkelejda Kas- neci. Human attention in fine-grained classification. arXiv preprint arXiv:2111.01628, 2021. 1

    2021 arXiv

  19. [27]

    In-context impersonation reveals large language models’ strengths and biases

    Leonard Salewski, Stephan Alaniz, Isabel Rio-Torto, Eric Schulz, and Zeynep Akata. In-context impersonation reveals large language models’ strengths and biases. Advances in Neural Information Processing Systems, 36, 2024. 2

    2024

  20. [28]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pages 6105–6114. PMLR,

  21. [29]

    Maria Tsimpoukelli, Jacob L Menick, Serkan Cabi, S. M. Ali Eslami, Oriol Vinyals, and Felix Hill. Multimodal few-shot learning with frozen language models. In Advances in Neu- ral Information Processing Systems, pages 200–212. Curran Associates, Inc., 2021. 1

    2021

  22. [30]

    Fastvit: A fast hybrid vision transformer using structural reparameterization

    Pavan Kumar Anasosalu Vasu, James Gabriel, Jeff Zhu, On- cel Tuzel, and Anurag Ranjan. Fastvit: A fast hybrid vision transformer using structural reparameterization. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision, pages 5785–5795, 2023. 1

    2023

  23. [31]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017. 1

    2017

  24. [32]

    The caltech-ucsd birds-200-2011 dataset

    Catherine Wah, Steve Branson, Peter Welinder, Pietro Per- ona, and Serge Belongie. The caltech-ucsd birds-200-2011 dataset. 2011. 1, 4, 6

    2011

  25. [33]

    Bi-lrfusion: Bi-directional lidar-radar fusion for 3d dynamic object detection

    Yingjie Wang, Jiajun Deng, Yao Li, Jinshui Hu, Cong Liu, Yu Zhang, Jianmin Ji, Wanli Ouyang, and Yanyong Zhang. Bi-lrfusion: Bi-directional lidar-radar fusion for 3d dynamic object detection. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognitio...

    2023

  26. [34]

    Simvlm: Simple visual language model pretraining with weak supervision

    Zirui Wang, Jiahui Yu, Adams Wei Yu, Zihang Dai, Yulia Tsvetkov, and Yuan Cao. Simvlm: Simple visual language model pretraining with weak supervision. arXiv preprint arXiv:2108.10904, 2021. 1

    2021 arXiv

  27. [35]

    Fine-grained image analysis with deep learning: A survey

    Xiu-Shen Wei, Yi-Zhe Song, Oisin Mac Aodha, Jianxin Wu, Yuxin Peng, Jinhui Tang, Jian Yang, and Serge Belongie. Fine-grained image analysis with deep learning: A survey. IEEE transactions on pattern analysis and machine intelli- gence, 44(12):8927–8948, 2021. 1

    2021

  28. [36]

    Aligning bag of regions for open- vocabulary object detection

    Size Wu, Wenwei Zhang, Sheng Jin, Wentao Liu, and Chen Change Loy. Aligning bag of regions for open- vocabulary object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15254–15264, 2023. 1

    2023

  29. [37]

    Learning concise and descriptive attributes for visual recognition

    An Yan, Yu Wang, Yiwu Zhong, Chengyu Dong, Zexue He, Yujie Lu, William Yang Wang, Jingbo Shang, and Julian McAuley. Learning concise and descriptive attributes for visual recognition. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 3090–3100,

  30. [38]

    Language in a bottle: Language model guided concept bottlenecks for interpretable image classification

    Yue Yang, Artemis Panagopoulou, Shenghao Zhou, Daniel Jin, Chris Callison-Burch, and Mark Yatskar. Language in a bottle: Language model guided concept bottlenecks for interpretable image classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern R...

    2023

  31. [39]

    An empirical study of gpt-3 for few-shot knowledge-based vqa

    Zhengyuan Yang, Zhe Gan, Jianfeng Wang, Xiaowei Hu, Yumao Lu, Zicheng Liu, and Lijuan Wang. An empirical study of gpt-3 for few-shot knowledge-based vqa. In Pro- ceedings of the AAAI Conference on Artificial Intelligence , pages 3081–3089, 2022. 1

    2022

  32. [40]

    A survey on multimodal large language models

    Shukang Yin, Chaoyou Fu, Sirui Zhao, Ke Li, Xing Sun, Tong Xu, and Enhong Chen. A survey on multimodal large language models. arXiv preprint arXiv:2306.13549, 2023. 2

    2023 arXiv

  33. [41]

    Mm-llms: Recent ad- vances in multimodal large language models

    Duzhen Zhang, Yahan Yu, Chenxing Li, Jiahua Dong, Dan Su, Chenhui Chu, and Dong Yu. Mm-llms: Recent ad- vances in multimodal large language models. arXiv preprint arXiv:2401.13601, 2024. 2

    2024 arXiv

  34. [42]

    Vinvl: Revisiting visual representations in vision-language models

    Pengchuan Zhang, Xiujun Li, Xiaowei Hu, Jianwei Yang, Lei Zhang, Lijuan Wang, Yejin Choi, and Jianfeng Gao. Vinvl: Revisiting visual representations in vision-language models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5579–5588,

  35. [43]

    Decoupled knowledge distillation

    Borui Zhao, Quan Cui, Renjie Song, Yiyu Qiu, and Jiajun Liang. Decoupled knowledge distillation. In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pages 11953–11962, 2022. 4, 6

    2022

  36. [44]

    Biformer: Vision transformer with bi-level routing attention

    Lei Zhu, Xinjiang Wang, Zhanghan Ke, Wayne Zhang, and Rynson WH Lau. Biformer: Vision transformer with bi-level routing attention. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 10323–10333, 2023. 1

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.