IG Captioner: Information Gain Captioners are Strong Zero-shot Classifiers

• URL: http://arxiv.org/abs/2311.17072v2
• Date: Wed, 17 Jul 2024 00:10:24 GMT
• Title: IG Captioner: Information Gain Captioners are Strong Zero-shot Classifiers
• Authors: Chenglin Yang, Siyuan Qiao, Yuan Cao, Yu Zhang, Tao Zhu, Alan Yuille, Jiahui Yu,
• Abstract summary: Generative training has been demonstrated to be powerful for building visual-language models. On zero-shot discriminative benchmarks, there is still a performance gap between models trained with generative and discriminative objectives. In this paper, we aim to narrow this gap by improving the efficacy of generative training on classification tasks.
• Score: 31.455819448471157
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Generative training has been demonstrated to be powerful for building visual-language models. However, on zero-shot discriminative benchmarks, there is still a performance gap between models trained with generative and discriminative objectives. In this paper, we aim to narrow this gap by improving the efficacy of generative training on classification tasks, without any finetuning processes or additional modules. Specifically, we focus on narrowing the gap between the generative captioner and the CLIP classifier. We begin by analysing the predictions made by the captioner and classifier and observe that the caption generation inherits the distribution bias from the language model trained with pure text modality, making it less grounded on the visual signal. To tackle this problem, we redesign the scoring objective for the captioner to alleviate the distributional bias and focus on measuring the gain of information brought by the visual inputs. We further design a generative training objective to match the evaluation objective. We name our model trained and evaluated from the novel procedures as Information Gain (IG) captioner. We pretrain the models on the public Laion-5B dataset and perform a series of discriminative evaluations. For the zero-shot classification on ImageNet, IG captioner achieves $> 18\%$ improvements over the standard captioner, achieving comparable performances with the CLIP classifier. IG captioner also demonstrated strong performance on zero-shot image-text retrieval tasks on MSCOCO and Flickr30K. We hope this paper inspires further research towards unifying generative and discriminative training procedures for visual-language models.

Related papers

• SILC: Improving Vision Language Pretraining with Self-Distillation [113.50400246862056]
  We introduce SILC, a novel framework for vision language pretraining. SILC improves image-text contrastive learning with the simple addition of local-to-global correspondence learning by self-distillation. We show that distilling local image features from an exponential moving average (EMA) teacher model significantly improves model performance on dense predictions tasks like detection and segmentation.
  arXiv  Detail & Related papers  (2023-10-20T08:44:47Z)
• Towards Realistic Zero-Shot Classification via Self Structural Semantic Alignment [53.2701026843921]
  Large-scale pre-trained Vision Language Models (VLMs) have proven effective for zero-shot classification. In this paper, we aim at a more challenging setting, Realistic Zero-Shot Classification, which assumes no annotation but instead a broad vocabulary. We propose the Self Structural Semantic Alignment (S3A) framework, which extracts structural semantic information from unlabeled data while simultaneously self-learning.
  arXiv  Detail & Related papers  (2023-08-24T17:56:46Z)
• Zero-shot Visual Question Answering with Language Model Feedback [83.65140324876536]
  We propose a language model guided captioning approach, LAMOC, for knowledge-based visual question answering (VQA) Our approach employs the generated captions by a captioning model as the context of an answer prediction model, which is a Pre-trained Language model (PLM)
  arXiv  Detail & Related papers  (2023-05-26T15:04:20Z)
• Zero-shot Image Captioning by Anchor-augmented Vision-Language Space Alignment [23.072180427273544]
  We discuss that directly employing CLIP for zero-shot image captioning relies more on the textual modality in context and largely ignores the visual information. To address this, we propose Cross-modal Language Models (CLMs) to facilitate unsupervised cross-modal learning. Experiments on MS COCO and Flickr 30K validate the promising performance of proposed approach in both captioning quality and computational efficiency.
  arXiv  Detail & Related papers  (2022-11-14T11:12:19Z)
• Text2Model: Text-based Model Induction for Zero-shot Image Classification [38.704831945753284]
  We address the challenge of building task-agnostic classifiers using only text descriptions. We generate zero-shot classifiers using a hypernetwork that receives class descriptions and outputs a multi-class model. We evaluate this approach in a series of zero-shot classification tasks, for image, point-cloud, and action recognition, using a range of text descriptions.
  arXiv  Detail & Related papers  (2022-10-27T05:19:55Z)
• Masked Unsupervised Self-training for Zero-shot Image Classification [98.23094305347709]
  Masked Unsupervised Self-Training (MUST) is a new approach which leverages two different and complimentary sources of supervision: pseudo-labels and raw images. MUST improves upon CLIP by a large margin and narrows the performance gap between unsupervised and supervised classification.
  arXiv  Detail & Related papers  (2022-06-07T02:03:06Z)
• Prompt-based Learning for Unpaired Image Captioning [86.44188293709307]
  Unpaired Image Captioning (UIC) has been developed to learn image descriptions from unaligned vision-language sample pairs. Recent successes of Vision-Language Pre-Trained Models (VL-PTMs) have triggered the development of prompt-based learning. We present in this paper a novel scheme based on prompt to train the UIC model, making best use of the powerful generalization ability.
  arXiv  Detail & Related papers  (2022-05-26T03:13:43Z)
• Better Captioning with Sequence-Level Exploration [60.57850194028581]
  We show the limitation of the current sequence-level learning objective for captioning tasks. In theory, we show that the current objective is equivalent to only optimizing the precision side of the caption set. Empirical result shows that the model trained by this objective tends to get lower score on the recall side.
  arXiv  Detail & Related papers  (2020-03-08T09:08:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.