A Stepwise Distillation Learning Strategy for Non-differentiable Visual Programming Frameworks on Visual Reasoning Tasks

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wan, Wentao, Kang, Nan, Wang, Zeqing, Yang, Zhuojie, Lin, Liang, Wang, Keze
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912240564699136
author Wan, Wentao
Kang, Nan
Wang, Zeqing
Yang, Zhuojie
Lin, Liang
Wang, Keze
author_facet Wan, Wentao
Kang, Nan
Wang, Zeqing
Yang, Zhuojie
Lin, Liang
Wang, Keze
contents Recently, Visual Programming (VProg) has emerged as a significant framework for visual reasoning (VR) tasks due to its interpretability and cross-task generality. However, even with invoking powerful pre-trained Vision-Language models (VLMs) as visual sub-modules, the performance of VProg on specific VR tasks is markedly inferior compared to well-trained task-specific networks. Although invoking task-specific models can further enhance the performance of VProg on specific VR tasks, it greatly diminishes the cross-task generalization ability of VProg. Besides, the non-differentiable nature of VProg prevents direct fine-tuning on specific VR tasks for further performance improvement. Attempt to address these issues, we propose SDVP, a Stepwise Distillation learning strategy for non-differentiable VPorg across various VR tasks. Specifically, our SDVP stepwise distills the capabilities of existing, well-trained small task-specific models for decomposed visual sub-tasks in VProg into the much larger VLMs invoked by corresponding visual sub-modules. Besides, distilling the knowledge of little-size task-specific models into pre-trained larger VLMs rather than replacing them helps keep the cross-task abilities of VProgs. Extensive and comprehensive experimental results on different VProg frameworks demonstrate that our SDVP obtains significant performance gains on specific VR benchmarks, i.e., GQA (+2.4\%) and NLVRv2 (+6.2\%) for VisProg and GQA (+6.5\%) and NLVRv2 (+4.0\%) for ViperGPT, and also maintains a promising performance for VProg on unseen and previous VR tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2309_09809
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle A Stepwise Distillation Learning Strategy for Non-differentiable Visual Programming Frameworks on Visual Reasoning Tasks
Wan, Wentao
Kang, Nan
Wang, Zeqing
Yang, Zhuojie
Lin, Liang
Wang, Keze
Computer Vision and Pattern Recognition
Artificial Intelligence
Recently, Visual Programming (VProg) has emerged as a significant framework for visual reasoning (VR) tasks due to its interpretability and cross-task generality. However, even with invoking powerful pre-trained Vision-Language models (VLMs) as visual sub-modules, the performance of VProg on specific VR tasks is markedly inferior compared to well-trained task-specific networks. Although invoking task-specific models can further enhance the performance of VProg on specific VR tasks, it greatly diminishes the cross-task generalization ability of VProg. Besides, the non-differentiable nature of VProg prevents direct fine-tuning on specific VR tasks for further performance improvement. Attempt to address these issues, we propose SDVP, a Stepwise Distillation learning strategy for non-differentiable VPorg across various VR tasks. Specifically, our SDVP stepwise distills the capabilities of existing, well-trained small task-specific models for decomposed visual sub-tasks in VProg into the much larger VLMs invoked by corresponding visual sub-modules. Besides, distilling the knowledge of little-size task-specific models into pre-trained larger VLMs rather than replacing them helps keep the cross-task abilities of VProgs. Extensive and comprehensive experimental results on different VProg frameworks demonstrate that our SDVP obtains significant performance gains on specific VR benchmarks, i.e., GQA (+2.4\%) and NLVRv2 (+6.2\%) for VisProg and GQA (+6.5\%) and NLVRv2 (+4.0\%) for ViperGPT, and also maintains a promising performance for VProg on unseen and previous VR tasks.
title A Stepwise Distillation Learning Strategy for Non-differentiable Visual Programming Frameworks on Visual Reasoning Tasks
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2309.09809