EAGLE: Elevating Geometric Reasoning through LLM-empowered Visual Instruction Tuning

• URL: http://arxiv.org/abs/2408.11397v1
• Date: Wed, 21 Aug 2024 07:43:50 GMT
• Title: EAGLE: Elevating Geometric Reasoning through LLM-empowered Visual Instruction Tuning
• Authors: Zhihao Li, Yao Du, Yang Liu, Yan Zhang, Yufang Liu, Mengdi Zhang, Xunliang Cai,
• Abstract summary: Existing MLLMs mostly optimize the LLM backbone to acquire geometric reasoning capabilities, while rarely emphasizing improvements in visual comprehension. Our findings reveal that current MLLMs severely suffer from inaccurate geometric perception and hallucinations. We propose a novel two-stage end-to-end visual enhancement MLLM framework designed to ElevAte Geometric reasoning.
• Score: 16.631783647518706
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multi-modal Large Language Models have recently experienced rapid developments and excel in various multi-modal tasks. However, they still struggle with mathematical geometric problem solving, which requires exceptional visual perception proficiency. Existing MLLMs mostly optimize the LLM backbone to acquire geometric reasoning capabilities, while rarely emphasizing improvements in visual comprehension. In this paper, we first investigate the visual perception performance of MLLMs when facing geometric diagrams. Our findings reveal that current MLLMs severely suffer from inaccurate geometric perception and hallucinations. To address these limitations, we propose EAGLE, a novel two-stage end-to-end visual enhancement MLLM framework designed to ElevAte Geometric reasoning through LLM-Empowered visual instruction tuning. Specifically, in the preliminary stage, we feed geometric image-caption pairs into our MLLM that contains a fully fine-tuning CLIP ViT and a frozen LLM, aiming to endow our model with basic geometric knowledge. In the subsequent advanced stage, we incorporate LoRA modules into the vision encoder and unfreeze the LLM backbone. This enables the model to leverage the inherent CoT rationales within question-answer pairs, guiding the MLLM to focus on nuanced visual cues and enhancing its overall perceptual capacity. Moreover, we optimize the cross-modal projector in both stages to foster adaptive visual-linguistic alignments. After the two-stage visual enhancement, we develop the geometry expert model EAGLE-7B. Extensive experiments on popular benchmarks demonstrate the effectiveness of our model. For example, on the GeoQA benchmark, EAGLE-7B not only surpasses the exemplary G-LLaVA 7B model by 2.9%, but also marginally outperforms the larger G-LLaVA 13B model. On the MathVista benchmark, EAGLE-7B achieves remarkable 3.8% improvements compared with the proprietary model GPT-4V.

Related papers

• GOBench: Benchmarking Geometric Optics Generation and Understanding of MLLMs [66.55945133516776]
  We introduce GOBench, the first benchmark to evaluate MLLMs' ability across two tasks: Generating Optically Authentic Imagery and Understanding Underlying Optical Phenomena.<n>We use MLLMs to construct GOBench-Gen-1k dataset. We then organize subjective experiments to assess the generated imagery based on Optical Authenticity, Aesthetic Quality, and Instruction Fidelity.<n>For the understanding task, we apply crafted evaluation instructions to test optical understanding ability of eleven prominent MLLMs. The experimental results demonstrate that current models face significant challenges in both optical generation and understanding.
  arXiv  Detail & Related papers  (2025-06-01T12:46:14Z)
• Training-Free Reasoning and Reflection in MLLMs [45.134271969594614]
  This paper introduces FRANK Model, a training-FRee ANd r1-liKe MLLM that imbues off-the-shelf MLLMs with reasoning and reflection abilities.<n>Our key insight is to decouple perception and reasoning across MLLM decoder layers.<n>To this end, we propose a layer-wise, Taylor-derived closed-form fusion mechanism that integrates reasoning capacity into deep decoder layers.
  arXiv  Detail & Related papers  (2025-05-22T02:51:12Z)
• LEO: Boosting Mixture of Vision Encoders for Multimodal Large Language Models [9.660892239615364]
  This work explores fusion strategies of visual tokens for hybrid MLLMs, leading to the design of LEO. Leo is a novel MLLM with a dual-branch vision encoder framework that incorporates a post-adaptation fusion strategy and adaptive tiling. We show that LEO can be adapted to the specialized domain of autonomous driving without altering the model architecture or training recipe.
  arXiv  Detail & Related papers  (2025-01-13T00:29:55Z)
• Open Eyes, Then Reason: Fine-grained Visual Mathematical Understanding in MLLMs [62.875934732547435]
  Current large language models (MLLMs) often underperform on mathematical problem-solving tasks that require fine-grained visual understanding. In this paper, we evaluate the visual grounding capabilities of state-of-the-art MLLMs and reveal a significant negative correlation between visual grounding accuracy and problem-solving performance. We propose a novel approach, SVE-Math, featuring a geometric-grounded vision encoder and a feature router that dynamically adjusts the contribution of hierarchical visual feature maps.
  arXiv  Detail & Related papers  (2025-01-11T04:08:44Z)
• OLA-VLM: Elevating Visual Perception in Multimodal LLMs with Auxiliary Embedding Distillation [95.78870389271832]
  The standard practice for developing contemporary MLLMs is to feed features from vision encoder(s) into the LLM and train with natural language supervision. We propose OLA-VLM, the first approach distilling knowledge into the LLM's hidden representations from a set of target visual representations. We show that OLA-VLM boosts performance by an average margin of up to 2.5% on various benchmarks, with a notable improvement of 8.7% on the Depth task in CV-Bench.
  arXiv  Detail & Related papers  (2024-12-12T18:55:18Z)
• AIM: Adaptive Inference of Multi-Modal LLMs via Token Merging and Pruning [19.68349294206012]
  We propose a training-free adaptive inference method for multi-modal LLMs. With a minimalist design, our method can be applied to both video and image LLMs. Under a similar computational cost, our method outperforms the state-of-the-art methods in long video understanding.
  arXiv  Detail & Related papers  (2024-12-04T11:47:57Z)
• LLaVA-KD: A Framework of Distilling Multimodal Large Language Models [70.19607283302712]
  We propose a novel framework to transfer knowledge from l-MLLM to s-MLLM. Specifically, we introduce Multimodal Distillation (MDist) to minimize the divergence between the visual-textual output distributions of l-MLLM and s-MLLM. We also propose a three-stage training scheme to fully exploit the potential of s-MLLM.
  arXiv  Detail & Related papers  (2024-10-21T17:41:28Z)
• Eagle: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders [89.38717274524681]
  This study explores the design space for multimodal large language models (MLLMs) using a mixture of vision encoders and resolutions. Our findings reveal several underlying principles common to various existing strategies, leading to a streamlined yet effective design approach. The resulting family of MLLMs, Eagle, surpasses other leading open-source models on major MLLM benchmarks.
  arXiv  Detail & Related papers  (2024-08-28T17:59:31Z)
• Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs [56.391404083287235]
  We introduce Cambrian-1, a family of multimodal LLMs (MLLMs) designed with a vision-centric approach. Our study uses LLMs and visual instruction tuning as an interface to evaluate various visual representations. We provide model weights, code, supporting tools, datasets, and detailed instruction-tuning and evaluation recipes.
  arXiv  Detail & Related papers  (2024-06-24T17:59:42Z)
• Delta-CoMe: Training-Free Delta-Compression with Mixed-Precision for Large Language Models [79.46938238953916]
  Fine-tuning large language models (LLMs) to diverse applications is crucial to meet complex demands. Recent studies suggest decomposing a fine-tuned LLM into a base model and corresponding delta weights, which are then compressed using low-rank or low-bit approaches to reduce costs. In this work, we observe that existing low-rank and low-bit compression methods can significantly harm the model performance for task-specific fine-tuned LLMs.
  arXiv  Detail & Related papers  (2024-06-13T07:57:27Z)
• Dense Connector for MLLMs [89.50595155217108]
  We introduce the Dense Connector - a plug-and-play vision-language connector that significantly enhances existing MLLMs. Building on this, we also propose the Efficient Dense Connector, which achieves performance comparable to LLaVA-v1.5 with only 25% of the visual tokens. Our model, trained solely on images, showcases remarkable zero-shot capabilities in video understanding as well.
  arXiv  Detail & Related papers  (2024-05-22T16:25:03Z)
• ST-LLM: Large Language Models Are Effective Temporal Learners [58.79456373423189]
  Large Language Models (LLMs) have showcased impressive capabilities in text comprehension and generation. How to effectively encode and understand videos in video-based dialogue systems remains to be solved. We propose ST-LLM, an effective video-LLM baseline with spatial-temporal sequence modeling inside LLM.
  arXiv  Detail & Related papers  (2024-03-30T10:11:26Z)
• Investigating the Catastrophic Forgetting in Multimodal Large Language Models [43.89009178021342]
  We introduce EMT: evaluating MulTimodality for evaluating the catastrophic forgetting in MLLMs. Almost all evaluated MLLMs fail to retain the same performance levels as their vision encoders on standard image classification tasks. As fine-tuning proceeds, the MLLMs begin to hallucinate, resulting in a significant loss of generalizability.
  arXiv  Detail & Related papers  (2023-09-19T04:51:13Z)
• Sight Beyond Text: Multi-Modal Training Enhances LLMs in Truthfulness and Ethics [32.123919380959485]
  Multi-modal large language models (MLLMs) are trained based on large language models (LLM) While they excel in multi-modal tasks, the pure NLP abilities of MLLMs are often underestimated and left untested. We show that visual instruction tuning, a prevailing strategy for transitioning LLMs into MLLMs, unexpectedly and interestingly helps models attain both improved truthfulness and ethical alignment.
  arXiv  Detail & Related papers  (2023-09-13T17:57:21Z)

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.