arXiv subjects

Zhibin Lan

Publications and source records attributed to Zhibin Lan.

12 recordsLinked to original sources

arXiv2026

Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings

Multimodal Large Language Models (MLLMs) have emerged as a promising foundation for universal multimodal embeddings. Recent studies have shown that reasoning-driven generative multimodal embeddings can outperform discriminative embeddings on several embedding tasks. However, Chain-of-Thought (CoT) reasoning tends to generate redundant thinking steps and introduce semantic ambiguity in the summarized answers in broader retrieval scenarios. To address this limitation, we propose Rewrite-driven Multimodal Embedding (RIME), a unified framework that jointly optimizes generation and embedding through a retrieval-friendly rewrite. Meanwhile, we present the Cross-Mode Alignment (CMA) to bridge the generative and discriminative embedding spaces, enabling flexible mutual retrieval to trade off efficiency and accuracy. Based on this, we also introduce Refine Reinforcement Learning (Refine-RL) that treats discriminative embeddings as stable semantic anchors to guide the rewrite optimization. Extensive experiments on MMEB-V2, MRMR and UVRB demonstrate that RIME substantially outperforms prior generative embedding models while significantly reducing the length of thinking. Code is available at https://github.com/PeppaWu/RIME.

cs.CV↗

arXiv2026

Towards Fine-Grained Code-Switch Speech Translation with Semantic Space Alignment

Code-switching (CS) speech translation (ST) aims to translate speech that alternates between multiple languages into a target language text, posing significant challenges due to the complexity of semantic modeling and the scarcity of CS data. Previous studies mainly rely on the models themselves to implicitly learn semantic representations and resort to costly manual annotations. To mitigate these limitations, we propose enhancing Large Language Models (LLMs) with a Mixture-of-Experts (MoE) speech projector composed of language expert groups, where each group specializes in the semantic space of a specific language for fine-grained speech feature modeling. A language-specific loss and an intra-group load balancing loss are jointly introduced to guide efficient token routing across and within expert groups. Furthermore, we introduce a multi-stage training paradigm that utilizes readily available automatic speech recognition (ASR) and monolingual ST data, facilitating speech-text alignment and improving translation performance. To bridge the data gap for smooth domain transfer, a transition loss is employed to improve adaptation to CS scenarios. Extensive experiments on widely used datasets demonstrate the effectiveness and generality of our approach, achieving average improvements of $0.86$ BLEU and $0.93$ COMET over SeamlessM4T, with maximum improvements of $1.49$ BLEU and $1.41$ COMET across different test sets.

cs.CL↗

arXiv2026

Countering the Over-Reliance Trap: Mitigating Object Hallucination for LVLMs via a Self-Validation Framework

Despite progress in Large Vision Language Models (LVLMs), object hallucination remains a critical issue in image captioning task, where models generate descriptions of non-existent objects, compromising their reliability. Previous work attributes this to LVLMs' over-reliance on language priors and attempts to mitigate it through logits calibration. However, they still lack a thorough analysis of the over-reliance. To gain a deeper understanding of over-reliance, we conduct a series of preliminary experiments, indicating that as the generation length increases, LVLMs' over-reliance on language priors leads to inflated probability of hallucinated object tokens, consequently exacerbating object hallucination. To circumvent this issue, we propose Language-Prior-Free Verification to enable LVLMs to faithfully verify the confidence of object existence. Based on this, we propose a novel training-free Self-Validation Framework to counter the over-reliance trap. It first validates objects' existence in sampled candidate captions and further mitigates object hallucination via caption selection or aggregation. Experiment results demonstrate that our framework mitigates object hallucination significantly in image captioning task (e.g., 65.6% improvement on CHAIRI metric with LLaVA-v1.5-7B), surpassing the previous SOTA methods. This result highlights a novel path towards mitigating hallucination by unlocking the inherent potential within LVLMs themselves.

cs.CV↗

arXiv2026

LLaVE: Large Language and Vision Embedding Models with Hardness-Weighted Contrastive Learning

Universal multimodal embedding models play a critical role in tasks such as interleaved image-text retrieval, multimodal RAG, and multimodal clustering. However, our empirical results indicate that existing LMM-based embedding models trained with the standard InfoNCE loss exhibit a high degree of overlap in similarity distribution between positive and negative pairs, making it challenging to distinguish hard negative pairs effectively. To deal with this issue, we propose a simple yet effective framework that dynamically improves the embedding model's representation learning for negative pairs based on their discriminative difficulty. Within this framework, we train a series of models, named LLaVE, and evaluate them on the MMEB benchmark, which covers 4 meta-tasks and 36 datasets. Experimental results show that LLaVE establishes stronger baselines that achieve state-of-the-art (SOTA) performance while demonstrating strong scalability and efficiency. Specifically, LLaVE-2B surpasses the previous SOTA 7B models, while LLaVE-7B achieves a further performance improvement of 6.2 points. Although LLaVE is trained on image-text data, it can generalize to text-video retrieval tasks in a zero-shot manner and achieve strong performance, demonstrating its remarkable potential for transfer to other embedding tasks.

cs.CV↗

arXiv2026

UME-R1: Exploring Reasoning-Driven Generative Multimodal Embeddings

The remarkable success of multimodal large language models (MLLMs) has driven advances in multimodal embeddings, yet existing models remain inherently discriminative, limiting their ability to benefit from reasoning-driven generation paradigm. In this work, we pioneer the exploration of generative embeddings, unifying embedding tasks within a generative paradigm. We propose UME-R1, a universal multimodal embedding framework consisting of a two-stage training strategy: a cold-start supervised fine-tuning equips the model with reasoning capabilities and enables it to generate both discriminative and generative embeddings; a subsequent reinforcement learning enhances reasoning and further optimizes generative embedding quality. This pioneering work reveals four key insights: 1) generative embeddings unlock substantial performance gains over conventional discriminative embeddings by leveraging the powerful generative reasoning capabilities of MLLMs; 2) discriminative and generative embeddings are complementary, whose combined oracle performance far exceeding that of either alone; 3) RL can effectively enhance generative embeddings, establishing a scalable optimization paradigm.; 4) repeated sampling at inference boosts downstream task coverage (pass@k), highlighting the inference-time scalability potential of generative embeddings. Evaluated on the MMEB-V2 benchmark across 78 tasks spanning video, image, and visual documents, UME-R1 significantly outperforms conventional discriminative embedding models and offers a foundation for more interpretable, reasoning-driven generative multimodal embeddings. Our code, models, and datasets will be publicly available at https://github.com/XMUDeepLIT/UME-R1.

cs.LG↗

arXiv2025

PATIMT-Bench: A Multi-Scenario Benchmark for Position-Aware Text Image Machine Translation in Large Vision-Language Models

Text Image Machine Translation (TIMT) aims to translate texts embedded within an image into another language. Current TIMT studies primarily focus on providing translations for all the text within an image, while neglecting to provide bounding boxes and covering limited scenarios. In this work, we extend traditional TIMT into position-aware TIMT (PATIMT), aiming to support fine-grained and layoutpreserving translation, which holds great practical value but remains largely unexplored. This task comprises two key sub-tasks: regionspecific translation and full-image translation with grounding. To support existing models on PATIMT and conduct fair evaluation, we construct the PATIMT benchmark (PATIMTBench), which consists of 10 diverse real-world scenarios. Specifically, we introduce an Adaptive Image OCR Refinement Pipeline, which adaptively selects appropriate OCR tools based on scenario and refines the results of text-rich images. To ensure evaluation reliability, we further construct a test set, which contains 1,200 high-quality instances manually annotated and reviewed by human experts. After fine-tuning on our data, compact Large Vision-Language Models (LVLMs) achieve state-of-the-art performance on both sub-tasks. Experimental results also highlight the scalability and generalizability of our training data

cs.CV↗

arXiv2025

AVG-LLaVA: An Efficient Large Multimodal Model with Adaptive Visual Granularity

Recently, large multimodal models (LMMs) have achieved significant advancements. When dealing with high-resolution images, dominant LMMs typically divide them into multiple local images and a global image, leading to a large number of visual tokens. In this work, we introduce AVG-LLaVA, an LMM that can adaptively select the appropriate visual granularity based on the input image and instruction. Specifically, we first apply the multiple pooling layers to obtain visual tokens at different granularities. Then we propose a visual granularity router, which includes a Transformer layer, an MLP layer, and a voter layer, used to select the appropriate visual granularity based on the image and instruction. Furthermore, we put forward RGLF, a novel training paradigm that aims at aligning the granularity predicted by the router with the preferences of the LMM, without the need for additional manually annotated data. Extensive experiments and analysis show that AVG-LLaVA achieves superior performance across 11 benchmarks, as well as significantly reduces the number of visual tokens and speeds up inference (e.g., an 85.3% reduction in visual tokens and a 2.53$\times$ increase in inference speed on the AI2D benchmark).

cs.CV↗

arXiv2024

"I've Heard of You!": Generate Spoken Named Entity Recognition Data for Unseen Entities

Spoken named entity recognition (NER) aims to identify named entities from speech, playing an important role in speech processing. New named entities appear every day, however, annotating their Spoken NER data is costly. In this paper, we demonstrate that existing Spoken NER systems perform poorly when dealing with previously unseen named entities. To tackle this challenge, we propose a method for generating Spoken NER data based on a named entity dictionary (NED) to reduce costs. Specifically, we first use a large language model (LLM) to generate sentences from the sampled named entities and then use a text-to-speech (TTS) system to generate the speech. Furthermore, we introduce a noise metric to filter out noisy data. To evaluate our approach, we release a novel Spoken NER benchmark along with a corresponding NED containing 8,853 entities. Experiment results show that our method achieves state-of-the-art (SOTA) performance in the in-domain, zero-shot domain adaptation, and fully zero-shot settings. Our data will be available at https://github.com/DeepLearnXMU/HeardU.

cs.CL↗

arXiv2024

Empowering Backbone Models for Visual Text Generation with Input Granularity Control and Glyph-Aware Training

Diffusion-based text-to-image models have demonstrated impressive achievements in diversity and aesthetics but struggle to generate images with legible visual texts. Existing backbone models have limitations such as misspelling, failing to generate texts, and lack of support for Chinese text, but their development shows promising potential. In this paper, we propose a series of methods, aiming to empower backbone models to generate visual texts in English and Chinese. We first conduct a preliminary study revealing that Byte Pair Encoding (BPE) tokenization and the insufficient learning of cross-attention modules restrict the performance of the backbone models. Based on these observations, we make the following improvements: (1) We design a mixed granularity input strategy to provide more suitable text representations; (2) We propose to augment the conventional training objective with three glyph-aware training losses, which enhance the learning of cross-attention modules and encourage the model to focus on visual texts. Through experiments, we demonstrate that our methods can effectively empower backbone models to generate semantic relevant, aesthetically appealing, and accurate visual text images, while maintaining their fundamental image generation quality.

cs.CV↗

arXiv2024

Translatotron-V(ison): An End-to-End Model for In-Image Machine Translation

In-image machine translation (IIMT) aims to translate an image containing texts in source language into an image containing translations in target language. In this regard, conventional cascaded methods suffer from issues such as error propagation, massive parameters, and difficulties in deployment and retaining visual characteristics of the input image. Thus, constructing end-to-end models has become an option, which, however, faces two main challenges: 1) the huge modeling burden, as it is required to simultaneously learn alignment across languages and preserve the visual characteristics of the input image; 2) the difficulties of directly predicting excessively lengthy pixel sequences. In this paper, we propose \textit{Translatotron-V(ision)}, an end-to-end IIMT model consisting of four modules. In addition to an image encoder, and an image decoder, our model contains a target text decoder and an image tokenizer. Among them, the target text decoder is used to alleviate the language alignment burden, and the image tokenizer converts long sequences of pixels into shorter sequences of visual tokens, preventing the model from focusing on low-level visual features. Besides, we present a two-stage training framework for our model to assist the model in learning alignment across modalities and languages. Finally, we propose a location-aware evaluation metric called Structure-BLEU to assess the translation quality of the generated images. Experimental results demonstrate that our model achieves competitive performance compared to cascaded models with only 70.9\% of parameters, and significantly outperforms the pixel-level end-to-end IIMT model.

cs.CL↗

arXiv2024

A Survey on Multi-modal Machine Translation: Tasks, Methods and Challenges

In recent years, multi-modal machine translation has attracted significant interest in both academia and industry due to its superior performance. It takes both textual and visual modalities as inputs, leveraging visual context to tackle the ambiguities in source texts. In this paper, we begin by offering an exhaustive overview of 99 prior works, comprehensively summarizing representative studies from the perspectives of dominant models, datasets, and evaluation metrics. Afterwards, we analyze the impact of various factors on model performance and finally discuss the possible research directions for this task in the future. Over time, multi-modal machine translation has developed more types to meet diverse needs. Unlike previous surveys confined to the early stage of multi-modal machine translation, our survey thoroughly concludes these emerging types from different aspects, so as to provide researchers with a better understanding of its current state.

cs.CL↗

arXiv2023

Exploring Better Text Image Translation with Multimodal Codebook

Text image translation (TIT) aims to translate the source texts embedded in the image to target translations, which has a wide range of applications and thus has important research value. However, current studies on TIT are confronted with two main bottlenecks: 1) this task lacks a publicly available TIT dataset, 2) dominant models are constructed in a cascaded manner, which tends to suffer from the error propagation of optical character recognition (OCR). In this work, we first annotate a Chinese-English TIT dataset named OCRMT30K, providing convenience for subsequent studies. Then, we propose a TIT model with a multimodal codebook, which is able to associate the image with relevant texts, providing useful supplementary information for translation. Moreover, we present a multi-stage training framework involving text machine translation, image-text alignment, and TIT tasks, which fully exploits additional bilingual texts, OCR dataset and our OCRMT30K dataset to train our model. Extensive experiments and in-depth analyses strongly demonstrate the effectiveness of our proposed model and training framework.

cs.CL↗