Large Language Models (LLMs) have achieved remarkable success and have been applied across various scientific fields, including chemistry. However, many chemical tasks require the processing of visual information, which cannot be successfully handled by existing chemical LLMs. This brings a growing need for models capable of integrating multimodal information in the chemical domain. In this paper, we introduce ChemVLM, an open-source chemical multimodal large language model specifically designed for chemical applications. ChemVLM is trained on a carefully curated bilingual multimodal dataset that enhances its ability to understand both textual and visual chemical information, including molecular structures, reactions, and chemistry examination questions. We develop three datasets for comprehensive evaluation, tailored to Chemical Optical Character Recognition (OCR), Multimodal Chemical Reasoning (MMCR), and Multimodal Molecule Understanding tasks. We benchmark ChemVLM against a range of open-source and proprietary multimodal large language models on various tasks. Experimental results demonstrate that ChemVLM achieves competitive performance across all evaluated tasks.

Multimodal aspect-based sentiment analysis (MABSA) integrates text and images to perform fine-grained sentiment analysis on specific aspects, enhancing the understanding of user opinions in various applications. Existing methods use modality alignment for information interaction and fusion between images and text, but an inherent gap between these two modalities necessitates a more direct bridging mechanism to effectively connect image understanding with text content. For this, we propose the Descriptions Enhanced Question-Answering Framework (DEQA), which generates descriptions of images using GPT-4, leveraging the multimodal large language model to provide more direct semantic context of images. In DEQA, to help the model better understand the task's purpose, we frame MABSA as a multi-turn question-answering problem to add semantic guidance and hints. We input text, image, and description into separate experts in various combinations, allowing each expert to focus on different features and thereby improving the comprehensive utilization of input information. By integrating these expert outputs within a multi-turn question-answering format, we employ a multi-expert ensemble decision-making approach to produce the final prediction results. Experimental results on two widely-used datasets demonstrate that our method achieves state-of-the-art performance. Furthermore, our framework substantially outperforms GPT-4o and other multimodal large language models, showcasing its superior effectiveness in multimodal sentiment analysis.

Jinfeng Zhou, Yuxuan Chen, Yihan Shi, Xuanming Zhang, Leqi Lei, Yi Feng, Zexuan Xiong, Miao Yan, Xunzhi Wang, Yaru Cao, Jianing Yin, Shuai Wang, Quanyu Dai, Zhenhua Dong, Hongning Wang, Minlie Huang. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Vision-language models (VLMs) have shown remarkable advancements in multimodal reasoning tasks. However, they still often generate inaccurate or irrelevant responses due to issues like hallucinated image understandings or unrefined reasoning paths. To address these challenges, we introduce Critic-V, a novel framework inspired by the Actor-Critic paradigm to boost the reasoning capability of VLMs. This framework decouples the reasoning process and critic process by integrating two independent components: the Reasoner, which generates reasoning paths based on visual and textual inputs, and the Critic, which provides constructive critique to refine these paths. In this approach, the Reasoner generates reasoning responses according to text prompts, which can evolve iteratively as a policy based on feedback from the Critic. This interaction process was theoretically driven by a reinforcement learning framework where the Critic offers natural language critiques instead of scalar rewards, enabling more nuanced feedback to boost the Reasoner’s capability on complex reasoning tasks. The Critic model is trained using Direct Preference Optimization (DPO), leveraging a preference dataset of critiques ranked by Rule-based Reward (RBR) to enhance its critic capabilities. Evaluation results show that the Critic-V framework significantly outperforms existing methods, including GPT-4V, on 5 out of 8 benchmarks, especially regarding reasoning accuracy and efficiency. Combining a dynamic text-based policy for the Reasoner and constructive feedback from the preference-optimized Critic enables a more reliable and context-sensitive multimodal reasoning process. Our approach provides a promising solution to enhance the reliability of VLMs, improving their performance in real-world reasoning-heavy multimodal applications such as autonomous driving and embodied intelligence. Our data and code are released at https://github.com/kyrieLei/Critic-V.

Deep learning models usually require large labeled datasets to generalize well, but this is computationally and financially costly. Cold-start few-shot data selection enables fast model generalization by selecting a few diverse, representative samples from an unlabeled data pool. To achieve this goal, previous work usually divides the training data into several clusters and performs sampling from these clusters. Yet, such a way tends to have two issues. First, imbalanced data distribution in the training data pool still exists in the selected subset, causing models' performance biases and suboptimal generalization ability. Second, these methods improve sample diversity in each cluster by considering either the feature dissimilarity among instances, or model uncertainty for individual instance. They ignore the entire representativeness of samples within a cluster. To tackle these challenges, we propose a novel framework HCDS : Hierarchical Clustering for Cold-Start Few-Shot Data Selection. Specifically, we first perform class-level clustering, using pseudo-labels for class supervision and applying contrastive clustering to derive class-rich features. We then refine these features within the class-level clusters into semantically meaningful features and perform representation-level clustering. Finally, we sample data from the representation-level clusters based on global similarity to ensure representativeness. Experimental results on six public datasets, including both balanced and imbalanced ones, show that HCDS achieves state-of-the-art performance, particularly with limited and imbalanced data.