Yanjie Wang

Jinghui Lu, Haiyang Yu, Yanjie Wang et al.

Recently, many studies have demonstrated that exclusively incorporating OCR-derived text and spatial layouts with large language models (LLMs) can be highly effective for document understanding tasks. However, existing methods that integrate spatial layouts with text have limitations, such as producing overly long text sequences or failing to fully leverage the autoregressive traits of LLMs. In this work, we introduce Interleaving Layout and Text in a Large Language Model (LayTextLLM)} for document understanding. LayTextLLM projects each bounding box to a single embedding and interleaves it with text, efficiently avoiding long sequence issues while leveraging autoregressive traits of LLMs. LayTextLLM not only streamlines the interaction of layout and textual data but also shows enhanced performance in KIE and VQA. Comprehensive benchmark evaluations reveal significant improvements of LayTextLLM, with a 15.2% increase on KIE tasks and 10.7% on VQA tasks compared to previous SOTA OCR-based LLMs. All resources are available at https://github.com/LayTextLLM/LayTextLLM.

Qiwen Zhu, Yanjie Wang, Shilv Cai et al.

Training Single-Image Super-Resolution (SISR) models using pixel-based regression losses can achieve high distortion metrics scores (e.g., PSNR and SSIM), but often results in blurry images due to insufficient recovery of high-frequency details. Conversely, using GAN or perceptual losses can produce sharp images with high perceptual metric scores (e.g., LPIPS), but may introduce artifacts and incorrect textures. Balancing these two types of losses can help achieve a trade-off between distortion and perception, but the challenge lies in tuning the loss function weights. To address this issue, we propose a novel method that incorporates Multi-Objective Optimization (MOO) into the training process of SISR models to balance perceptual quality and distortion. We conceptualize the relationship between loss weights and image quality assessment (IQA) metrics as black-box objective functions to be optimized within our Multi-Objective Bayesian Optimization Super-Resolution (MOBOSR) framework. This approach automates the hyperparameter tuning process, reduces overall computational cost, and enables the use of numerous loss functions simultaneously. Extensive experiments demonstrate that MOBOSR outperforms state-of-the-art methods in terms of both perceptual quality and distortion, significantly advancing the perception-distortion Pareto frontier. Our work points towards a new direction for future research on balancing perceptual quality and fidelity in nearly all image restoration tasks. The source code and pretrained models are available at: https://github.com/ZhuKeven/MOBOSR.

Han Wang, Yanjie Wang, Yongjie Ye et al.

Multi-modal Large Language Models (MLLMs) have demonstrated their ability to perceive objects in still images, but their application in video-related tasks, such as object tracking, remains understudied. This lack of exploration is primarily due to two key challenges. Firstly, extensive pretraining on large-scale video datasets is required to equip MLLMs with the capability to perceive objects across multiple frames and understand inter-frame relationships. Secondly, processing a large number of frames within the context window of Large Language Models (LLMs) can impose a significant computational burden. To address the first challenge, we introduce ElysiumTrack-1M, a large-scale video dataset supported for three tasks: Single Object Tracking (SOT), Referring Single Object Tracking (RSOT), and Video Referring Expression Generation (Video-REG). ElysiumTrack-1M contains 1.27 million annotated video frames with corresponding object boxes and descriptions. Leveraging this dataset, we conduct training of MLLMs and propose a token-compression model T-Selector to tackle the second challenge. Our proposed approach, Elysium: Exploring Object-level Perception in Videos via MLLM, is an end-to-end trainable MLLM that attempts to conduct object-level tasks in videos without requiring any additional plug-in or expert models. All codes and datasets are available at https://github.com/Hon-Wong/Elysium.

Junjie Yu, Pengrui Lu, Weiye Si et al.

Benchmarks within the OpenClaw ecosystem have thus far evaluated exclusively assistant-level tasks, leaving the academic-level capabilities of OpenClaw largely unexamined. We introduce AcademiClaw, a bilingual benchmark of 80 complex, long-horizon tasks sourced directly from university students' real academic workflows -- homework, research projects, competitions, and personal projects -- that they found current AI agents unable to solve effectively. Curated from 230 student-submitted candidates through rigorous expert review, the final task set spans 25+ professional domains, ranging from olympiad-level mathematics and linguistics problems to GPU-intensive reinforcement learning and full-stack system debugging, with 16 tasks requiring CUDA GPU execution. Each task executes in an isolated Docker sandbox and is scored on task completion by multi-dimensional rubrics combining six complementary techniques, with an independent five-category safety audit providing additional behavioral analysis. Experiments on six frontier models show that even the best achieves only a 55\% pass rate. Further analysis uncovers sharp capability boundaries across task domains, divergent behavioral strategies among models, and a disconnect between token consumption and output quality, providing fine-grained diagnostic signals beyond what aggregate metrics reveal. We hope that AcademiClaw and its open-sourced data and code can serve as a useful resource for the OpenClaw community, driving progress toward agents that are more capable and versatile across the full breadth of real-world academic demands. All data and code are available at https://github.com/GAIR-NLP/AcademiClaw.

Han Wang, Yuxiang Nie, Yongjie Ye et al.

The application of Large Vision-Language Models (LVLMs) for analyzing images and videos is an exciting and rapidly evolving field. In recent years, we've seen significant growth in high-quality image-text datasets for fine-tuning image understanding, but there is still a lack of comparable datasets for videos. Additionally, many VideoLLMs are extensions of single-image VLMs, which may not efficiently handle the complexities of longer videos. In this study, we introduce a large-scale synthetic dataset created from proprietary models, using carefully designed prompts to tackle a wide range of questions. We also explore a dynamic visual token compression architecture that strikes a balance between computational efficiency and performance. Our proposed \model{} achieves state-of-the-art results across various video tasks and shows impressive generalization, setting new baselines in multi-image understanding. Notably, \model{} delivers an absolute improvement of 2.7\% over LLaVA-OneVision on VideoMME and 10.7\% on MuirBench. Codes are available at https://github.com/Hon-Wong/ByteVideoLLM

Han Wang, Yongjie Ye, Bingru Li et al.

We introduce Vision as LoRA (VoRA), a novel paradigm for transforming an LLM into an MLLM. Unlike prevalent MLLM architectures that rely on external vision modules for vision encoding, VoRA internalizes visual capabilities by integrating vision-specific LoRA layers directly into the LLM. This design allows the added parameters to be seamlessly merged into the LLM during inference, eliminating structural complexity and minimizing computational overhead. Moreover, inheriting the LLM's ability of handling flexible context, VoRA can process inputs at arbitrary resolutions. To further strengthen VoRA's visual capabilities, we introduce a block-wise distillation method that transfers visual priors from a pre-trained ViT into the LoRA layers, effectively accelerating training by injecting visual knowledge. Additionally, we apply bi-directional attention masks to better capture the context information of an image. We successfully demonstrate that with additional pre-training data, VoRA can perform comparably with conventional encode-based MLLMs. All training data, codes, and model weights will be released at https://github.com/Hon-Wong/VoRA.

Haiyang Yu, Jinghui Lu, Yanjie Wang et al.

The advent of Large Vision-Language Models (LVLMs) has advanced the video-based tasks, such as video captioning and video understanding. Some previous research indicates that taking texts in videos as input can further improve the performance of video understanding. As a type of indispensable information in short videos or movies, subtitles can assist LVLMs to better understand videos. Most existing methods for video subtitle extraction are based on a multi-stage framework, handling each frame independently. They can hardly exploit the temporal information of videos. Although some LVLMs exhibit the robust OCR capability, predicting accurate timestamps for subtitle texts is still challenging. In this paper, we propose an End-to-end Video Subtitle Extraction method, called EVE, which consists of three modules: a vision encoder, an adapter module, and a large language model. To effectively compress the visual tokens from the vision encoder, we propose a novel adapter InterleavedVT to interleave two modalities. It contains a visual compressor and a textual region compressor. The proposed InterleavedVT exploits both the merits of average pooling and Q-Former in token compression. Taking the temporal information of videos into account, we introduce a sliding-window mechanism in the textual region compressor. To benchmark the video subtitle extraction task, we propose a large dataset ViSa including 2.5M videos. Extensive experiments on ViSa demonstrate that the proposed EVE can outperform existing open-sourced tools and LVLMs.

Zhitong Lai, Haichao Sun, Rui Tian et al.

Skip connections are fundamental units in encoder-decoder networks, which are able to improve the feature propagtion of the neural networks. However, most methods with skip connections just connected features with the same resolution in the encoder and the decoder, which ignored the information loss in the encoder with the layers going deeper. To leverage the information loss of the features in shallower layers of the encoder, we propose a full skip connection network (FSCN) for monocular depth estimation task. In addition, to fuse features within skip connections more closely, we present an adaptive concatenation module (ACM). Further more, we conduct extensive experiments on the ourdoor and indoor datasets (i.e., the KITTI dataste and the NYU Depth V2 dataset) for FSCN and FSCN gets the state-of-the-art results.

Jingqun Tang, Qi Liu, Yongjie Ye et al.

Text-Centric Visual Question Answering (TEC-VQA) in its proper format not only facilitates human-machine interaction in text-centric visual environments but also serves as a de facto gold proxy to evaluate AI models in the domain of text-centric scene understanding. Nonetheless, most existing TEC-VQA benchmarks have focused on high-resource languages like English and Chinese. Despite pioneering works to expand multilingual QA pairs in non-text-centric VQA datasets through translation engines, the translation-based protocol encounters a substantial "visual-textual misalignment" problem when applied to TEC-VQA. Specifically, it prioritizes the text in question-answer pairs while disregarding the visual text present in images. Moreover, it fails to address complexities related to nuanced meaning, contextual distortion, language bias, and question-type diversity. In this work, we tackle multilingual TEC-VQA by introducing MTVQA, the first benchmark featuring high-quality human expert annotations across 9 diverse languages, consisting of 6,778 question-answer pairs across 2,116 images. Further, by comprehensively evaluating numerous state-of-the-art Multimodal Large Language Models~(MLLMs), including Qwen2-VL, GPT-4o, GPT-4V, Claude3, and Gemini, on the MTVQA benchmark, it is evident that there is still a large room for performance improvement (Qwen2-VL scoring 30.9 versus 79.7 for human performance), underscoring the value of MTVQA. Additionally, we supply multilingual training data within the MTVQA dataset, demonstrating that straightforward fine-tuning with this data can substantially enhance multilingual TEC-VQA performance. We aspire that MTVQA will offer the research community fresh insights and stimulate further exploration in multilingual visual text comprehension. The project homepage is available at https://bytedance.github.io/MTVQA/.

Jinghui Lu, Ziwei Yang, Yanjie Wang et al.

In this study, we aim to reduce generation latency for Named Entity Recognition (NER) with Large Language Models (LLMs). The main cause of high latency in LLMs is the sequential decoding process, which autoregressively generates all labels and mentions for NER, significantly increase the sequence length. To this end, we introduce Parallel Decoding in LLM for NE} (PaDeLLM-NER), a approach that integrates seamlessly into existing generative model frameworks without necessitating additional modules or architectural modifications. PaDeLLM-NER allows for the simultaneous decoding of all mentions, thereby reducing generation latency. Experiments reveal that PaDeLLM-NER significantly increases inference speed that is 1.76 to 10.22 times faster than the autoregressive approach for both English and Chinese. Simultaneously it maintains the quality of predictions as evidenced by the performance that is on par with the state-of-the-art across various datasets.

Xiang Fei, Jinghui Lu, Qi Sun et al.

Autoregressive models have become the de facto choice for sequence generation tasks, but standard approaches treat digits as independent tokens and apply cross-entropy loss, overlooking the coherent structure of numerical sequences. This paper introduces Numerical Token Integrity Loss (NTIL) to address this gap. NTIL operates at two levels: (1) token-level, where it extends the Earth Mover's Distance (EMD) to preserve ordinal relationships between numerical values, and (2) sequence-level, where it penalizes the overall discrepancy between the predicted and actual sequences. This dual approach improves numerical prediction and integrates effectively with LLMs/MLLMs. Extensive experiments show significant performance improvements with NTIL.

Han Wang, Yanjie Wang, Yang Li et al.

Video Text Spotting (VTS) is a fundamental visual task that aims to predict the trajectories and content of texts in a video. Previous works usually conduct local associations and apply IoU-based distance and complex post-processing procedures to boost performance, ignoring the abundant temporal information and the morphological characteristics in VTS. In this paper, we propose a novel Global Video Text Spotting Transformer GloTSFormer to model the tracking problem as global associations and utilize the Gaussian Wasserstein distance to guide the morphological correlation between frames. Our main contributions can be summarized as three folds. 1). We propose a Transformer-based global tracking method GloTSFormer for VTS and associate multiple frames simultaneously. 2). We introduce a Wasserstein distance-based method to conduct positional associations between frames. 3). We conduct extensive experiments on public datasets. On the ICDAR2015 video dataset, GloTSFormer achieves 56.0 MOTA with 4.6 absolute improvement compared with the previous SOTA method and outperforms the previous Transformer-based method by a significant 8.3 MOTA.

Yanjie Wang, Xu Zou, Zhijun Zhang et al.

Detecting oriented objects along with estimating their rotation information is one crucial step for analyzing remote sensing images. Despite that many methods proposed recently have achieved remarkable performance, most of them directly learn to predict object directions under the supervision of only one (e.g. the rotation angle) or a few (e.g. several coordinates) groundtruth values individually. Oriented object detection would be more accurate and robust if extra constraints, with respect to proposal and rotation information regression, are adopted for joint supervision during training. To this end, we innovatively propose a mechanism that simultaneously learns the regression of horizontal proposals, oriented proposals, and rotation angles of objects in a consistent manner, via naive geometric computing, as one additional steady constraint (see Figure 1). An oriented center prior guided label assignment strategy is proposed for further enhancing the quality of proposals, yielding better performance. Extensive experiments demonstrate the model equipped with our idea significantly outperforms the baseline by a large margin to achieve a new state-of-the-art result without any extra computational burden during inference. Our proposed idea is simple and intuitive that can be readily implemented. Source codes and trained models are involved in supplementary files.

Yanjie Wang, Samuel Broscheit, Rainer Gemulla

We propose the Relational Tucker3 (RT) decomposition for multi-relational link prediction in knowledge graphs. We show that many existing knowledge graph embedding models are special cases of the RT decomposition with certain predefined sparsity patterns in its components. In contrast to these prior models, RT decouples the sizes of entity and relation embeddings, allows parameter sharing across relations, and does not make use of a predefined sparsity pattern. We use the RT decomposition as a tool to explore whether it is possible and beneficial to automatically learn sparsity patterns, and whether dense models can outperform sparse models (using the same number of parameters). Our experiments indicate that---depending on the dataset--both questions can be answered affirmatively.

Yanjie Wang, Daniel Ruffinelli, Rainer Gemulla et al.

Knowledge bases contribute to many web search and mining tasks, yet they are often incomplete. To add missing facts to a given knowledge base, various embedding models have been proposed in the recent literature. Perhaps surprisingly, relatively simple models with limited expressiveness often performed remarkably well under today's most commonly used evaluation protocols. In this paper, we explore whether recent models work well for knowledge base completion and argue that the current evaluation protocols are more suited for question answering rather than knowledge base completion. We show that when focusing on a different prediction task for evaluating knowledge base completion, the performance of current embedding models is unsatisfactory even on datasets previously thought to be too easy. This is especially true when embedding models are compared against a simple rule-based baseline. This work indicates the need for more research into the embedding models and evaluation protocols for knowledge base completion.

Yanjie Wang, Rainer Gemulla, Hui Li

We study bilinear embedding models for the task of multi-relational link prediction and knowledge graph completion. Bilinear models belong to the most basic models for this task, they are comparably efficient to train and use, and they can provide good prediction performance. The main goal of this paper is to explore the expressiveness of and the connections between various bilinear models proposed in the literature. In particular, a substantial number of models can be represented as bilinear models with certain additional constraints enforced on the embeddings. We explore whether or not these constraints lead to universal models, which can in principle represent every set of relations, and whether or not there are subsumption relationships between various models. We report results of an independent experimental study that evaluates recent bilinear models in a common experimental setup. Finally, we provide evidence that relation-level ensembles of multiple bilinear models can achieve state-of-the art prediction performance.