Positive-Augmented Contrastive Learning for Vision-and-Language Evaluation and Training
About
Despite significant advancements in caption generation, existing evaluation metrics often fail to capture the full quality or fine-grained details of captions. This is mainly due to their reliance on non-specific human-written references or noisy pre-training data. Still, finding an effective metric is crucial not only for captions evaluation but also for the generation phase. Metrics can indeed play a key role in the fine-tuning stage of captioning models, ultimately enhancing the quality of the generated captions. In this paper, we propose PAC-S++, a learnable metric that leverages the CLIP model, pre-trained on both web-collected and cleaned data and regularized through additional pairs of generated visual and textual positive samples. Exploiting this stronger and curated pre-training, we also apply PAC-S++ as a reward in the Self-Critical Sequence Training (SCST) stage typically employed to fine-tune captioning models. Extensive experiments on different image and video datasets highlight the effectiveness of PAC-S++ compared to popular metrics for the task, including its sensitivity to object hallucinations. Furthermore, we show that integrating PAC-S++ into the fine-tuning stage of a captioning model results in semantically richer captions with fewer repetitions and grammatical errors. Evaluations on out-of-domain benchmarks further demonstrate the efficacy of our fine-tuning approach in enhancing model capabilities. Source code and trained models are publicly available at: https://github.com/aimagelab/pacscore.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Image Captioning Evaluation | Composite | Kendall-c Tau_c62 | 131 | |
| Image Captioning Evaluation | Flickr8K-CF | Kendall-b Correlation (tau_b)38.8 | 99 | |
| Image Captioning Evaluation | Flickr8k Expert | Kendall Tau-c (tau_c)57.9 | 82 | |
| Image Captioning Evaluation | Pascal-50S | Accuracy84.7 | 44 | |
| Hallucination Detection | FOIL | Accuracy (4 Refs)94.1 | 32 | |
| Image Captioning Evaluation | Nebula | Kendall tau_c50.6 | 31 | |
| Multimodal Preference Evaluation | Pascal | P-Acc84.7 | 10 | |
| Multimodal Preference Evaluation | Polaris | tau_c56 | 10 | |
| Multimodal Preference Evaluation | FlickrExp | tau_c55.9 | 10 | |
| Multimodal Preference Evaluation | FlickrCF | Tau-b Score37.6 | 10 |