Yifan Xu

Wenbo Hu, Yifan Xu, Yi Li et al. · harvard

Vision Language Models (VLMs), which extend Large Language Models (LLM) by incorporating visual understanding capability, have demonstrated significant advancements in addressing open-ended visual question-answering (VQA) tasks. However, these models cannot accurately interpret images infused with text, a common occurrence in real-world scenarios. Standard procedures for extracting information from images often involve learning a fixed set of query embeddings. These embeddings are designed to encapsulate image contexts and are later used as soft prompt inputs in LLMs. Yet, this process is limited to the token count, potentially curtailing the recognition of scenes with text-rich context. To improve upon them, the present study introduces BLIVA: an augmented version of InstructBLIP with Visual Assistant. BLIVA incorporates the query embeddings from InstructBLIP and also directly projects encoded patch embeddings into the LLM, a technique inspired by LLaVA. This approach assists the model to capture intricate details potentially missed during the query decoding process. Empirical evidence demonstrates that our model, BLIVA, significantly enhances performance in processing text-rich VQA benchmarks (up to 17.76% in OCR-VQA benchmark) and in undertaking general (not particularly text-rich) VQA benchmarks (up to 7.9% in Visual Spatial Reasoning benchmark), and achieved 17.72% overall improvement in a comprehensive multimodal LLM benchmark (MME), comparing to our baseline InstructBLIP. BLIVA demonstrates significant capability in decoding real-world images, irrespective of text presence. To demonstrate the broad industry applications enabled by BLIVA, we evaluate the model using a new dataset comprising YouTube thumbnails paired with question-answer sets across 11 diverse categories. Our code and models are freely accessible at https://github.com/mlpc-ucsd/BLIVA.

Yifan Xu, Mengdan Zhang, Xiaoshan Yang et al.

In this paper, we for the first time explore helpful multi-modal contextual knowledge to understand novel categories for open-vocabulary object detection (OVD). The multi-modal contextual knowledge stands for the joint relationship across regions and words. However, it is challenging to incorporate such multi-modal contextual knowledge into OVD. The reason is that previous detection frameworks fail to jointly model multi-modal contextual knowledge, as object detectors only support vision inputs and no caption description is provided at test time. To this end, we propose a multi-modal contextual knowledge distillation framework, MMC-Det, to transfer the learned contextual knowledge from a teacher fusion transformer with diverse multi-modal masked language modeling (D-MLM) to a student detector. The diverse multi-modal masked language modeling is realized by an object divergence constraint upon traditional multi-modal masked language modeling (MLM), in order to extract fine-grained region-level visual contexts, which are vital to object detection. Extensive experiments performed upon various detection datasets show the effectiveness of our multi-modal context learning strategy, where our approach well outperforms the recent state-of-the-art methods.

Yifan Xu, Qizhang Shen, Vineet Kamat et al.

This paper presents KUKAloha, a general, low-cost, and shared-control teleoperation framework designed for construction robot arms. The proposed system employs a leader-follower paradigm in which a lightweight leading arm enables intuitive human guidance for coarse robot motion, while an autonomous perception module based on AprilTag detection performs precise alignment and grasp execution. By explicitly decoupling human control from fine manipulation, KUKAloha improves safety and repeatability when operating large-scale manipulators. We implement the framework on a KUKA robot arm and conduct a usability study with representative construction manipulation tasks. Experimental results demonstrate that KUKAloha reduces operator workload, improves task completion efficiency, and provides a practical solution for scalable demonstration collection and shared human-robot control in construction environments.

Yifan Xu, Pourya Shamsolmoali, Jie Yang

Visual place recognition (VPR) is a highly challenging task that has a wide range of applications, including robot navigation and self-driving vehicles. VPR is particularly difficult due to the presence of duplicate regions and the lack of attention to small objects in complex scenes, resulting in recognition deviations. In this paper, we present ClusVPR, a novel approach that tackles the specific issues of redundant information in duplicate regions and representations of small objects. Different from existing methods that rely on Convolutional Neural Networks (CNNs) for feature map generation, ClusVPR introduces a unique paradigm called Clustering-based Weighted Transformer Network (CWTNet). CWTNet leverages the power of clustering-based weighted feature maps and integrates global dependencies to effectively address visual deviations encountered in large-scale VPR problems. We also introduce the optimized-VLAD (OptLAD) layer that significantly reduces the number of parameters and enhances model efficiency. This layer is specifically designed to aggregate the information obtained from scale-wise image patches. Additionally, our pyramid self-supervised strategy focuses on extracting representative and diverse information from scale-wise image patches instead of entire images, which is crucial for capturing representative and diverse information in VPR. Extensive experiments on four VPR datasets show our model's superior performance compared to existing models while being less complex.

Zeju Li, Chao Zhang, Xiaoyan Wang et al.

The remarkable potential of multi-modal large language models (MLLMs) in comprehending both vision and language information has been widely acknowledged. However, the scarcity of 3D scenes-language pairs in comparison to their 2D counterparts, coupled with the inadequacy of existing approaches in understanding of 3D scenes by LLMs, poses a significant challenge. In response, we collect and construct an extensive dataset comprising 75K instruction-response pairs tailored for 3D scenes. This dataset addresses tasks related to 3D VQA, 3D grounding, and 3D conversation. To further enhance the integration of 3D spatial information into LLMs, we introduce a novel and efficient prompt tuning paradigm, 3DMIT. This paradigm eliminates the alignment stage between 3D scenes and language and extends the instruction prompt with the 3D modality information including the entire scene and segmented objects. We evaluate the effectiveness of our method across diverse tasks in the 3D scene domain and find that our approach serves as a strategic means to enrich LLMs' comprehension of the 3D world. Our code is available at https://github.com/staymylove/3DMIT.

Yifan Xu, Xiaoshan Yang, Yaguang Song et al.

In this work, we introduce Libra, a prototype model with a decoupled vision system on a large language model (LLM). The decoupled vision system decouples inner-modal modeling and cross-modal interaction, yielding unique visual information modeling and effective cross-modal comprehension. Libra is trained through discrete auto-regressive modeling on both vision and language inputs. Specifically, we incorporate a routed visual expert with a cross-modal bridge module into a pretrained LLM to route the vision and language flows during attention computing to enable different attention patterns in inner-modal modeling and cross-modal interaction scenarios. Experimental results demonstrate that the dedicated design of Libra achieves a strong MLLM baseline that rivals existing works in the image-to-text scenario with merely 50 million training data, providing a new perspective for future multimodal foundation models. Code is available at https://github.com/YifanXu74/Libra.

Ruilin Yao, Bo Zhang, Jirui Huang et al.

Multimodal Large Language Models (MLLMs) have achieved significant advances in integrating visual and linguistic information, yet their ability to reason about complex and real-world scenarios remains limited. The existing benchmarks are usually constructed in the task-oriented manner without guarantee that different task samples come from the same data distribution, thus they often fall short in evaluating the synergistic effects of lower-level perceptual capabilities on higher-order reasoning. To lift this limitation, we contribute Lens, a multi-level benchmark with 3.4K contemporary images and 60K+ human-authored questions covering eight tasks and 12 daily scenarios, forming three progressive task tiers, i.e., perception, understanding, and reasoning. One feature is that each image is equipped with rich annotations for all tasks. Thus, this dataset intrinsically supports to evaluate MLLMs to handle image-invariable prompts, from basic perception to compositional reasoning. In addition, our images are manully collected from the social media, in which 53% were published later than Jan. 2025. We evaluate 15+ frontier MLLMs such as Qwen2.5-VL-72B, InternVL3-78B, GPT-4o and two reasoning models QVQ-72B-preview and Kimi-VL. These models are released later than Dec. 2024, and none of them achieve an accuracy greater than 60% in the reasoning tasks. Project page: https://github.com/Lens4MLLMs/lens. ICCV 2025 workshop page: https://lens4mllms.github.io/mars2-workshop-iccv2025/

Zhifei Yang, Keyang Lu, Chao Zhang et al. · pku

Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism and controllability in terms of geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for scene generation are constrained to text-based inputs and exhibit insufficient adaptability to flexible user inputs, hindering the ability to precisely control object geometry. To address this issue, we propose MMGDreamer, a dual-branch diffusion model for scene generation that incorporates a novel Mixed-Modality Graph, visual enhancement module, and relation predictor. The mixed-modality graph allows object nodes to integrate textual and visual modalities, with optional relationships between nodes. It enhances adaptability to flexible user inputs and enables meticulous control over the geometry of objects in the generated scenes. The visual enhancement module enriches the visual fidelity of text-only nodes by constructing visual representations using text embeddings. Furthermore, our relation predictor leverages node representations to infer absent relationships between nodes, resulting in more coherent scene layouts. Extensive experimental results demonstrate that MMGDreamer exhibits superior control of object geometry, achieving state-of-the-art scene generation performance. Project page: https://yangzhifeio.github.io/project/MMGDreamer.

Chenjie Hao, Weyl Lu, Yifan Xu et al.

An embodied system must not only model the patterns of the external world but also understand its own motion dynamics. A motion dynamic model is essential for efficient skill acquisition and effective planning. In this work, we introduce the neural motion simulator (MoSim), a world model that predicts the future physical state of an embodied system based on current observations and actions. MoSim achieves state-of-the-art performance in physical state prediction and provides competitive performance across a range of downstream tasks. This works shows that when a world model is accurate enough and performs precise long-horizon predictions, it can facilitate efficient skill acquisition in imagined worlds and even enable zero-shot reinforcement learning. Furthermore, MoSim can transform any model-free reinforcement learning (RL) algorithm into a model-based approach, effectively decoupling physical environment modeling from RL algorithm development. This separation allows for independent advancements in RL algorithms and world modeling, significantly improving sample efficiency and enhancing generalization capabilities. Our findings highlight that world models for motion dynamics is a promising direction for developing more versatile and capable embodied systems.

Huashan Chen, Yifan Xu, Yue Feng et al.

Biometrics authentication has become increasingly popular due to its security and convenience; however, traditional biometrics are becoming less desirable in scenarios such as new mobile devices, Virtual Reality, and Smart Vehicles. For example, while face authentication is widely used, it suffers from significant privacy concerns. The collection of complete facial data makes it less desirable for privacy-sensitive applications. Lip authentication, on the other hand, has emerged as a promising biometrics method. However, existing lip-based authentication methods heavily depend on static lip shape when the mouth is closed, which can be less robust due to lip shape dynamic motion and can barely work when the user is speaking. In this paper, we revisit the nature of lip biometrics and extract shape-independent features from the lips. We study the dynamic characteristics of lip biometrics based on articulator motion. Building on the knowledge, we propose a system for shape-independent continuous authentication via lip articulator dynamics. This system enables robust, shape-independent and continuous authentication, making it particularly suitable for scenarios with high security and privacy requirements. We conducted comprehensive experiments in different environments and attack scenarios and collected a dataset of 50 subjects. The results indicate that our system achieves an overall accuracy of 99.06% and demonstrates robustness under advanced mimic attacks and AI deepfake attacks, making it a viable solution for continuous biometric authentication in various applications.

Yifan Xu, Kekai Sheng, Weiming Dong et al.

Big progress has been achieved in domain adaptation in decades. Existing works are always based on an ideal assumption that testing target domain are i.i.d. with training target domains. However, due to unpredictable corruptions (e.g., noise and blur) in real data like web images, domain adaptation methods are increasingly required to be corruption robust on target domains. In this paper, we investigate a new task, Corruption-agnostic Robust Domain Adaptation (CRDA): to be accurate on original data and robust against unavailable-for-training corruptions on target domains. This task is non-trivial due to large domain discrepancy and unsupervised target domains. We observe that simple combinations of popular methods of domain adaptation and corruption robustness have sub-optimal CRDA results. We propose a new approach based on two technical insights into CRDA: 1) an easy-to-plug module called Domain Discrepancy Generator (DDG) that generates samples that enlarge domain discrepancy to mimic unpredictable corruptions; 2) a simple but effective teacher-student scheme with contrastive loss to enhance the constraints on target domains. Experiments verify that DDG keeps or even improves performance on original data and achieves better corruption robustness that baselines.

Ke Li, Shijie Wang, Xiang Zhang et al.

In this paper, we present a regression-based pose recognition method using cascade Transformers. One way to categorize the existing approaches in this domain is to separate them into 1). heatmap-based and 2). regression-based. In general, heatmap-based methods achieve higher accuracy but are subject to various heuristic designs (not end-to-end mostly), whereas regression-based approaches attain relatively lower accuracy but they have less intermediate non-differentiable steps. Here we utilize the encoder-decoder structure in Transformers to perform regression-based person and keypoint detection that is general-purpose and requires less heuristic design compared with the existing approaches. We demonstrate the keypoint hypothesis (query) refinement process across different self-attention layers to reveal the recursive self-attention mechanism in Transformers. In the experiments, we report competitive results for pose recognition when compared with the competing regression-based methods.

Weijian Xu, Yifan Xu, Tyler Chang et al.

In this paper, we present Co-scale conv-attentional image Transformers (CoaT), a Transformer-based image classifier equipped with co-scale and conv-attentional mechanisms. First, the co-scale mechanism maintains the integrity of Transformers' encoder branches at individual scales, while allowing representations learned at different scales to effectively communicate with each other; we design a series of serial and parallel blocks to realize the co-scale mechanism. Second, we devise a conv-attentional mechanism by realizing a relative position embedding formulation in the factorized attention module with an efficient convolution-like implementation. CoaT empowers image Transformers with enriched multi-scale and contextual modeling capabilities. On ImageNet, relatively small CoaT models attain superior classification results compared with similar-sized convolutional neural networks and image/vision Transformers. The effectiveness of CoaT's backbone is also illustrated on object detection and instance segmentation, demonstrating its applicability to downstream computer vision tasks.

Yifan Xu, Weijian Xu, David Cheung et al.

In this paper, we present a joint end-to-end line segment detection algorithm using Transformers that is post-processing and heuristics-guided intermediate processing (edge/junction/region detection) free. Our method, named LinE segment TRansformers (LETR), takes advantages of having integrated tokenized queries, a self-attention mechanism, and an encoding-decoding strategy within Transformers by skipping standard heuristic designs for the edge element detection and perceptual grouping processes. We equip Transformers with a multi-scale encoder/decoder strategy to perform fine-grained line segment detection under a direct endpoint distance loss. This loss term is particularly suitable for detecting geometric structures such as line segments that are not conveniently represented by the standard bounding box representations. The Transformers learn to gradually refine line segments through layers of self-attention. In our experiments, we show state-of-the-art results on Wireframe and YorkUrban benchmarks.

Zheng Ding, Yifan Xu, Weijian Xu et al.

We propose an algorithm, guided variational autoencoder (Guided-VAE), that is able to learn a controllable generative model by performing latent representation disentanglement learning. The learning objective is achieved by providing signals to the latent encoding/embedding in VAE without changing its main backbone architecture, hence retaining the desirable properties of the VAE. We design an unsupervised strategy and a supervised strategy in Guided-VAE and observe enhanced modeling and controlling capability over the vanilla VAE. In the unsupervised strategy, we guide the VAE learning by introducing a lightweight decoder that learns latent geometric transformation and principal components; in the supervised strategy, we use an adversarial excitation and inhibition mechanism to encourage the disentanglement of the latent variables. Guided-VAE enjoys its transparency and simplicity for the general representation learning task, as well as disentanglement learning. On a number of experiments for representation learning, improved synthesis/sampling, better disentanglement for classification, and reduced classification errors in meta-learning have been observed.