# SDRT: Enhance Vision-Language Models by Self-Distillation with Diverse Reasoning Traces

arXiv: https://arxiv.org/abs/2503.01754 · HTML: https://mlsanity.com/paper/2503.01754/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 36 of its 41 references that are in the map.
- First posted on arXiv 3 Mar 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)
- Knowledge distillation / [Multimodal distillation](https://mlsanity.com/distill/multimodal-distillation/)

## Builds on (most cited first, 30 of 36)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [Self-Instruct: Aligning Language Models with Self-Generated Instructions](https://mlsanity.com/paper/2212.10560/) (2022)
- [MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models](https://mlsanity.com/paper/2304.10592/) (2023)
- [MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI](https://mlsanity.com/paper/2311.16502/) (2023)
- [PaLM-E: An Embodied Multimodal Language Model](https://mlsanity.com/paper/2303.03378/) (2023)
- [Visual Prompt Tuning](https://mlsanity.com/paper/2203.12119/) (2022)
- [ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning](https://mlsanity.com/paper/2203.10244/) (2022)
- [Conditional Prompt Learning for Vision-Language Models](https://mlsanity.com/paper/2203.05557/) (2022)
- [Show, Attend and Tell: Neural Image Caption Generation with Visual Attention](https://mlsanity.com/paper/1502.03044/) (2015)
- [DocVQA: A Dataset for VQA on Document Images](https://mlsanity.com/paper/2007.00398/) (2020)
- [CLIP-Adapter: Better Vision-Language Models with Feature Adapters](https://mlsanity.com/paper/2110.04544/) (2021)
- [On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes](https://mlsanity.com/paper/2306.13649/) (2023)
- [Vision-Language Models for Vision Tasks: A Survey](https://mlsanity.com/paper/2304.00685/) (2023)
- [Multimodal Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2302.00923/) (2023)
- [Be Your Own Teacher: Improve the Performance of Convolutional Neural Networks via Self Distillation](https://mlsanity.com/paper/1905.08094/) (2019)
- [Tip-Adapter: Training-free Adaption of CLIP for Few-shot Classification](https://mlsanity.com/paper/2207.09519/) (2022)
- [FigureQA: An Annotated Figure Dataset for Visual Reasoning](https://mlsanity.com/paper/1710.07300/) (2017)
- [A Survey of Confidence Estimation and Calibration in Large Language Models](https://mlsanity.com/paper/2311.08298/) (2023)
- [ReFT: Representation Finetuning for Language Models](https://mlsanity.com/paper/2404.03592/) (2024)
- [Navigate through Enigmatic Labyrinth A Survey of Chain of Thought Reasoning: Advances, Frontiers and Future](https://mlsanity.com/paper/2309.15402/) (2023)
- [Self-Distillation Amplifies Regularization in Hilbert Space](https://mlsanity.com/paper/2002.05715/) (2020)
- [Distilling System 2 into System 1](https://mlsanity.com/paper/2407.06023/) (2024)
- [What Makes Good Examples for Visual In-Context Learning?](https://mlsanity.com/paper/2301.13670/) (2023)
- [Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition](https://mlsanity.com/paper/1811.00751/) (2018)

## Built on by (most cited first, 5 of 5)

- [On-Policy Self-Distillation without Any Supervision](https://mlsanity.com/paper/2608.06296/) (2026)
- [Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation](https://mlsanity.com/paper/2605.11651/) (2026)
- [Making VLMs More Robot-Friendly: Self-Critical Distillation of Low-Level Procedural Reasoning](https://mlsanity.com/paper/2507.08224/) (2025)
- [Boosting Omni-Modal Language Models: Staged Post-Training with Visually Debiased Evaluation](https://mlsanity.com/paper/2605.12034/) (2026)
- [PROPA: Toward Process-level Optimization in Visual Reasoning via Reinforcement Learning](https://mlsanity.com/paper/2511.10279/) (2025)

## Abstract (arXiv)

Reasoning is increasingly crucial for various tasks. While chain-of-thought prompting enables large language models to leverage reasoning effectively, harnessing the reasoning capabilities of Vision-Language Models (VLMs) remains challenging. To solve this problem, we propose a novel self-distillation framework that enhances the reasoning capabilities of the model. The proposed framework introduces several key innovations. We start by employing a prompt library tailored to visual reasoning tasks to generate diverse in-context questions and utilize a two-step reasoning procedure to derive reasoning-guided responses. These responses are then used for self-distillation, enabling the model to internalize the reasoning process. Additionally, we improve the model architecture with several innovative components, including an intervention adapter for efficient parameter updates, a cross-modal skip connection to facilitate information exchange between modalities, and an ensemble learning algorithm to integrate diverse reasoning from multiple in-context questions. Extensive experiments show that our method significantly improves the baseline performance across five VQA datasets.