Papers with MathVerse
Do Multi-Agents Solve Better Than Single? Evaluating Agentic Frameworks for Diagram-Grounded Geometry Problem Solving and Reasoning (2026.eacl-srw)
Copied to clipboard
Mahbub E Sobhani, Md. Faiyaz Abdullah Sayeedi, Mohammad Nehad Alam, Proma Hossain Progga, Swakkhar Shatabda
| Challenge: | Diagram-grounded geometry problem solving is critical for multimodal large language models, but the benefits of multi-agent design over single-aggent remain unclear. |
| Approach: | They compare diagram-grounded geometry problem solving to four visual math benchmarks . they found that multi-agent pipelines provide clear benefits for open-source models . |
| Outcome: | Theorem-based solvers and architectural refinements improve performance on four visual math benchmarks. |
VisualWebInstruct: Scaling up Multimodal Instruction Data through Web Search (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing vision-language models struggle with reasoning-focused tasks due to the lack of high-quality training data. |
| Approach: | They propose a new approach that leverages search engines to create a multimodal multimodal dataset . they use a set of 30,000 seed images to extract HTML data from 700K unique URLs . |
| Outcome: | The proposed model achieves the best known performance on MMMU-Pro (40.7), MathVerse (42.6), and DynaMath (55.7). |
Math-LLaVA: Bootstrapping Mathematical Reasoning for Multimodal Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing image instruction fine-tuning datasets do not fully exploit visual information to enhance multimodal reasoning capabilities of Large language models (LLMs). |
| Approach: | They propose a LLaVA-based model fine-tuned with MathV360K to bridge this gap by collecting 40K high-quality images with question-answer pairs from 24 existing datasets and synthesizing 320K new pairs. |
| Outcome: | The proposed model improves the multimodal reasoning capabilities of LLaVA-1.5 and demonstrates enhanced generalizability on the MMMU benchmark. |
GeoDANO: Geometric VLM with Domain Agnostic Vision Encoder (2025.findings-emnlp)
Copied to clipboard
| Challenge: | GeoDANO is a geometric vision-language model with a domain-agnostic vision encoder . it is currently limited to recognizing geometric features in general-purpose VLMs . |
| Approach: | They propose a geometric vision-language model with a domain-agnostic vision encoder for plane geometry problems. |
| Outcome: | The proposed model outperforms vision encoders in recognizing geometric features . it outperformed specialized methods for plane geometry problems and GPT-4o on MathVerse . |
MAmmoTH-VL: Eliciting Multimodal Reasoning with Instruction Tuning at Scale (2025.acl-long)
Copied to clipboard
Jiawei Guo, Tianyu Zheng, Yizhi Li, Yuelin Bai, Bo Li, Yubo Wang, King Zhu, Graham Neubig, Wenhu Chen, Xiang Yue
| Challenge: | Current instruction-tuning datasets focus on simplistic visual question answering tasks, and provide phrase-level answers without any intermediate rationales. |
| Approach: | They propose to use open-source multimodal large language models to train MLLMs on a dataset with 12M instruction-response pairs to elicit CoT reasoning. |
| Outcome: | The proposed model achieves state-of-the-art performance on benchmarks such as MathVerse, MMMU-Pro, and MuirBench, and gains improvements of up to 4% on non-reasoning-based benchmarks. |
InfiMM-WebMath-40B: Advancing Multimodal Pre-Training for Enhanced Mathematical Reasoning (2025.findings-emnlp)
Copied to clipboard
Xiaotian Han, Yiren Jian, Xuefeng Hu, Haogeng Liu, Yiqi Wang, Qihang Fan, Yuang Ai, Huaibo Huang, Ran He, Zhenheng Yang, Quanzeng You
| Challenge: | InfiMM-WebMath-40B is a dataset of interleaved image-text documents . it consists of 24 million web pages, 85 million image URLs, and 40 billion text tokens . |
| Approach: | InfiMM-WebMath-40B is a high-quality dataset of interleaved image-text documents . it contains 24 million web pages, 85 million image URLs, and 40 billion text tokens . |
| Outcome: | InfiMM-WebMath-40B is a high-quality dataset of interleaved image-text documents . it consists of 24 million web pages, 85 million image URLs, and 40 billion text tokens . |
DRIFT: Transferring Reasoning Priors for Efficient MLLM Fine-Tuning (2026.findings-acl)
Copied to clipboard
Chao Huang, Zeliang Zhang, Jiang Liu, Ximeng Sun, Jialian Wu, Xiaodong Yu, Ze Wang, Chenliang Xu, Emad Barsoum, Zicheng Liu
| Challenge: | Multimodal large language models (MLLMs) have made rapid progress in perception and alignment, but their reasoning ability often lags behind strong text-only LLMs. |
| Approach: | They propose a method that transfers reasoning knowledge in the gradient space while preserving multimodal alignment. |
| Outcome: | Experiments on multimodal reasoning benchmarks show that DRIFT outperforms naive merging and standard SFT. |
InterFeedback: Unveiling Interactive Intelligence of Large Multimodal Models with Human Feedback (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing benchmarks do not test Large Multimodal Models (LMMs) on their interactive intelligence with human users. |
| Approach: | They propose an interactive framework that can be applied to any LMM and assess their interactive intelligence with human users. |
| Outcome: | The proposed framework can be applied to any LMM and dataset to assess interactive intelligence with human users. |
ReFLAIR: Enhancing Multimodal Reasoning via Structured Reflection and Reward-Guided Learning (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing training methods for large models do not address the trade-off between reflection and accuracy. |
| Approach: | a unified framework teaches large models to perform structured reflection via an explicit $think re-think answer $ format and hybrid reward learning. |
| Outcome: | The proposed framework improves model performance on mathematical benchmarks and reduces inference cost by nearly 23%. |