Joya Chen

Kristen Grauman, Andrew Westbury, Lorenzo Torresani et al. · cmu, gatech

We present Ego-Exo4D, a diverse, large-scale multimodal multiview video dataset and benchmark challenge. Ego-Exo4D centers around simultaneously-captured egocentric and exocentric video of skilled human activities (e.g., sports, music, dance, bike repair). 740 participants from 13 cities worldwide performed these activities in 123 different natural scene contexts, yielding long-form captures from 1 to 42 minutes each and 1,286 hours of video combined. The multimodal nature of the dataset is unprecedented: the video is accompanied by multichannel audio, eye gaze, 3D point clouds, camera poses, IMU, and multiple paired language descriptions -- including a novel "expert commentary" done by coaches and teachers and tailored to the skilled-activity domain. To push the frontier of first-person video understanding of skilled human activity, we also present a suite of benchmark tasks and their annotations, including fine-grained activity understanding, proficiency estimation, cross-view translation, and 3D hand/body pose. All resources are open sourced to fuel new research in the community. Project page: http://ego-exo4d-data.org/

Kevin Qinghong Lin, Pengchuan Zhang, Joya Chen et al. · microsoft-research, uw

Video Temporal Grounding (VTG), which aims to ground target clips from videos (such as consecutive intervals or disjoint shots) according to custom language queries (e.g., sentences or words), is key for video browsing on social media. Most methods in this direction develop taskspecific models that are trained with type-specific labels, such as moment retrieval (time interval) and highlight detection (worthiness curve), which limits their abilities to generalize to various VTG tasks and labels. In this paper, we propose to Unify the diverse VTG labels and tasks, dubbed UniVTG, along three directions: Firstly, we revisit a wide range of VTG labels and tasks and define a unified formulation. Based on this, we develop data annotation schemes to create scalable pseudo supervision. Secondly, we develop an effective and flexible grounding model capable of addressing each task and making full use of each label. Lastly, thanks to the unified framework, we are able to unlock temporal grounding pretraining from large-scale diverse labels and develop stronger grounding abilities e.g., zero-shot grounding. Extensive experiments on three tasks (moment retrieval, highlight detection and video summarization) across seven datasets (QVHighlights, Charades-STA, TACoS, Ego4D, YouTube Highlights, TVSum, and QFVS) demonstrate the effectiveness and flexibility of our proposed framework. The codes are available at https://github.com/showlab/UniVTG.

Joya Chen, Difei Gao, Kevin Qinghong Lin et al. · microsoft-research, uw

Humans excel at learning from expert demonstrations and solving their own problems. To equip intelligent robots and assistants, such as AR glasses, with this ability, it is essential to ground human hand interactions (i.e., affordances) from demonstration videos and apply them to a target image like a user's AR glass view. The video-to-image affordance grounding task is challenging due to (1) the need to predict fine-grained affordances, and (2) the limited training data, which inadequately covers video-image discrepancies and negatively impacts grounding. To tackle them, we propose Affordance Transformer (Afformer), which has a fine-grained transformer-based decoder that gradually refines affordance grounding. Moreover, we introduce Mask Affordance Hand (MaskAHand), a self-supervised pre-training technique for synthesizing video-image data and simulating context changes, enhancing affordance grounding across video-image discrepancies. Afformer with MaskAHand pre-training achieves state-of-the-art performance on multiple benchmarks, including a substantial 37% improvement on the OPRA dataset. Code is made available at https://github.com/showlab/afformer.

Zechen Bai, Tong He, Haiyang Mei et al.

We introduce VideoLISA, a video-based multimodal large language model designed to tackle the problem of language-instructed reasoning segmentation in videos. Leveraging the reasoning capabilities and world knowledge of large language models, and augmented by the Segment Anything Model, VideoLISA generates temporally consistent segmentation masks in videos based on language instructions. Existing image-based methods, such as LISA, struggle with video tasks due to the additional temporal dimension, which requires temporal dynamic understanding and consistent segmentation across frames. VideoLISA addresses these challenges by integrating a Sparse Dense Sampling strategy into the video-LLM, which balances temporal context and spatial detail within computational constraints. Additionally, we propose a One-Token-Seg-All approach using a specially designed <TRK> token, enabling the model to segment and track objects across multiple frames. Extensive evaluations on diverse benchmarks, including our newly introduced ReasonVOS benchmark, demonstrate VideoLISA's superior performance in video object segmentation tasks involving complex reasoning, temporal understanding, and object tracking. While optimized for videos, VideoLISA also shows promising generalization to image segmentation, revealing its potential as a unified foundation model for language-instructed object segmentation. Code and model will be available at: https://github.com/showlab/VideoLISA.

Kevin Qinghong Lin, Pengchuan Zhang, Difei Gao et al.

Narrative videos, such as movies, pose significant challenges in video understanding due to their rich contexts (characters, dialogues, storylines) and diverse demands (identify who, relationship, and reason). In this paper, we introduce MovieSeq, a multimodal language model developed to address the wide range of challenges in understanding video contexts. Our core idea is to represent videos as interleaved multimodal sequences (including images, plots, videos, and subtitles), either by linking external knowledge databases or using offline models (such as whisper for subtitles). Through instruction-tuning, this approach empowers the language model to interact with videos using interleaved multimodal instructions. For example, instead of solely relying on video as input, we jointly provide character photos alongside their names and dialogues, allowing the model to associate these elements and generate more comprehensive responses. To demonstrate its effectiveness, we validate MovieSeq's performance on six datasets (LVU, MAD, Movienet, CMD, TVC, MovieQA) across five settings (video classification, audio description, video-text retrieval, video captioning, and video question-answering). The code will be public at https://github.com/showlab/MovieSeq.

Difei Gao, Lei Ji, Luowei Zhou et al.

Recent research on Large Language Models (LLMs) has led to remarkable advancements in general NLP AI assistants. Some studies have further explored the use of LLMs for planning and invoking models or APIs to address more general multi-modal user queries. Despite this progress, complex visual-based tasks still remain challenging due to the diverse nature of visual tasks. This diversity is reflected in two aspects: 1) Reasoning paths. For many real-life applications, it is hard to accurately decompose a query simply by examining the query itself. Planning based on the specific visual content and the results of each step is usually required. 2) Flexible inputs and intermediate results. Input forms could be flexible for in-the-wild cases, and involves not only a single image or video but a mixture of videos and images, e.g., a user-view image with some reference videos. Besides, a complex reasoning process will also generate diverse multimodal intermediate results, e.g., video narrations, segmented video clips, etc. To address such general cases, we propose a multi-modal AI assistant, AssistGPT, with an interleaved code and language reasoning approach called Plan, Execute, Inspect, and Learn (PEIL) to integrate LLMs with various tools. Specifically, the Planner is capable of using natural language to plan which tool in Executor should do next based on the current reasoning progress. Inspector is an efficient memory manager to assist the Planner to feed proper visual information into a specific tool. Finally, since the entire reasoning process is complex and flexible, a Learner is designed to enable the model to autonomously explore and discover the optimal solution. We conducted experiments on A-OKVQA and NExT-QA benchmarks, achieving state-of-the-art results. Moreover, showcases demonstrate the ability of our system to handle questions far more complex than those found in the benchmarks.

Benita Wong, Joya Chen, You Wu et al.

A long-standing goal of intelligent assistants such as AR glasses/robots has been to assist users in affordance-centric real-world scenarios, such as "how can I run the microwave for 1 minute?". However, there is still no clear task definition and suitable benchmarks. In this paper, we define a new task called Affordance-centric Question-driven Task Completion, where the AI assistant should learn from instructional videos to provide step-by-step help in the user's view. To support the task, we constructed AssistQ, a new dataset comprising 531 question-answer samples from 100 newly filmed instructional videos. We also developed a novel Question-to-Actions (Q2A) model to address the AQTC task and validate it on the AssistQ dataset. The results show that our model significantly outperforms several VQA-related baselines while still having large room for improvement. We expect our task and dataset to advance Egocentric AI Assistant's development. Our project page is available at: https://showlab.github.io/assistq/.

Shiwei Wu, Joya Chen, Kevin Qinghong Lin et al.

A well-known dilemma in large vision-language models (e.g., GPT-4, LLaVA) is that while increasing the number of vision tokens generally enhances visual understanding, it also significantly raises memory and computational costs, especially in long-term, dense video frame streaming scenarios. Although learnable approaches like Q-Former and Perceiver Resampler have been developed to reduce the vision token burden, they overlook the context causally modeled by LLMs (i.e., key-value cache), potentially leading to missed visual cues when addressing user queries. In this paper, we introduce a novel approach to reduce vision compute by leveraging redundant vision tokens "skipping layers" rather than decreasing the number of vision tokens. Our method, VideoLLM-MoD, is inspired by mixture-of-depths LLMs and addresses the challenge of numerous vision tokens in long-term or streaming video. Specifically, for each transformer layer, we learn to skip the computation for a high proportion (e.g., 80\%) of vision tokens, passing them directly to the next layer. This approach significantly enhances model efficiency, achieving approximately \textasciitilde42\% time and \textasciitilde30\% memory savings for the entire training. Moreover, our method reduces the computation in the context and avoid decreasing the vision tokens, thus preserving or even improving performance compared to the vanilla model. We conduct extensive experiments to demonstrate the effectiveness of VideoLLM-MoD, showing its state-of-the-art results on multiple benchmarks, including narration, forecasting, and summarization tasks in COIN, Ego4D, and Ego-Exo4D datasets.

Weijia Wu, Chen Gao, Joya Chen et al.

Recent advances at the intersection of reinforcement learning (RL) and visual intelligence have enabled agents that not only perceive complex visual scenes but also reason, generate, and act within them. This survey offers a critical and up-to-date synthesis of the field. We first formalize visual RL problems and trace the evolution of policy-optimization strategies from RLHF to verifiable reward paradigms, and from Proximal Policy Optimization to Group Relative Policy Optimization. We then organize more than 200 representative works into four thematic pillars: multi-modal large language models, visual generation, unified model frameworks, and vision-language-action models. For each pillar we examine algorithmic design, reward engineering, benchmark progress, and we distill trends such as curriculum-driven training, preference-aligned diffusion, and unified reward modeling. Finally, we review evaluation protocols spanning set-level fidelity, sample-level preference, and state-level stability, and we identify open challenges that include sample efficiency, generalization, and safe deployment. Our goal is to provide researchers and practitioners with a coherent map of the rapidly expanding landscape of visual RL and to highlight promising directions for future inquiry. Resources are available at: https://github.com/weijiawu/Awesome-Visual-Reinforcement-Learning.

Dong Guo, Faming Wu, Feida Zhu et al. · pku

We present Seed1.5-VL, a vision-language foundation model designed to advance general-purpose multimodal understanding and reasoning. Seed1.5-VL is composed with a 532M-parameter vision encoder and a Mixture-of-Experts (MoE) LLM of 20B active parameters. Despite its relatively compact architecture, it delivers strong performance across a wide spectrum of public VLM benchmarks and internal evaluation suites, achieving the state-of-the-art performance on 38 out of 60 public benchmarks. Moreover, in agent-centric tasks such as GUI control and gameplay, Seed1.5-VL outperforms leading multimodal systems, including OpenAI CUA and Claude 3.7. Beyond visual and video understanding, it also demonstrates strong reasoning abilities, making it particularly effective for multimodal reasoning challenges such as visual puzzles. We believe these capabilities will empower broader applications across diverse tasks. In this report, we mainly provide a comprehensive review of our experiences in building Seed1.5-VL across model design, data construction, and training at various stages, hoping that this report can inspire further research. Seed1.5-VL is now accessible at https://www.volcengine.com/ (Volcano Engine Model ID: doubao-1-5-thinking-vision-pro-250428)

Ziteng Gao, Zhan Tong, Kevin Qinghong Lin et al.

The recently proposed SparseFormer architecture provides an alternative approach to visual understanding by utilizing a significantly lower number of visual tokens via adjusting RoIs, greatly reducing computational costs while still achieving promising performance. However, training SparseFormers from scratch is still expensive, and scaling up the number of parameters can be challenging. In this paper, we propose to bootstrap SparseFormers from ViT-based vision foundation models in a simple and efficient way. Since the majority of SparseFormer blocks are the standard transformer ones, we can inherit weights from large-scale pre-trained vision transformers and freeze them as much as possible. Therefore, we only need to train the SparseFormer-specific lightweight focusing transformer to adjust token RoIs and fine-tune a few early pre-trained blocks to align the final token representation. In such a way, we can bootstrap SparseFormer architectures from various large-scale pre-trained models (e.g., IN-21K pre-trained AugRegs or CLIPs) using a rather smaller amount of training samples (e.g., IN-1K) and without labels or captions within just a few hours. As a result, the bootstrapped unimodal SparseFormer (from AugReg-ViT-L/16-384) can reach 84.9% accuracy on IN-1K with only 49 tokens, and the multimodal SparseFormer from CLIPs also demonstrates notable zero-shot performance with highly reduced computational cost without seeing any caption during the bootstrapping procedure. In addition, CLIP-bootstrapped SparseFormers, which align the output space with language without seeing a word, can serve as efficient vision encoders in multimodal large language models. Code and models are available at https://github.com/showlab/sparseformer

Joya Chen, Kai Xu, Yuhui Wang et al.

A standard hardware bottleneck when training deep neural networks is GPU memory. The bulk of memory is occupied by caching intermediate tensors for gradient computation in the backward pass. We propose a novel method to reduce this footprint - Dropping Intermediate Tensors (DropIT). DropIT drops min-k elements of the intermediate tensors and approximates gradients from the sparsified tensors in the backward pass. Theoretically, DropIT reduces noise on estimated gradients and therefore has a higher rate of convergence than vanilla-SGD. Experiments show that we can drop up to 90\% of the intermediate tensor elements in fully-connected and convolutional layers while achieving higher testing accuracy for Visual Transformers and Convolutional Neural Networks on various tasks (e.g., classification, object detection, instance segmentation). Our code and models are available at https://github.com/chenjoya/dropit.

Xianfeng Liang, Likang Wu, Joya Chen et al.

Recently, the traffic congestion in modern cities has become a growing worry for the residents. As presented in Baidu traffic report, the commuting stress index has reached surprising 1.973 in Beijing during rush hours, which results in longer trip time and increased vehicular queueing. Previous works have demonstrated that by reasonable scheduling, e.g, rebalancing bike-sharing systems and optimized bus transportation, the traffic efficiency could be significantly improved with little resource consumption. However, there are still two disadvantages that restrict their performance: (1) they only consider single scheduling in a short time, but ignoring the layout after first reposition, and (2) they only focus on the single transport. However, the multi-modal characteristics of urban public transportation are largely under-exploited. In this paper, we propose an efficient and economical multi-modal traffic scheduling scheme named JLRLS based on spatio -temporal prediction, which adopts reinforcement learning to obtain optimal long-term and joint schedule. In JLRLS, we combines multiple transportation to conduct scheduling by their own characteristics, which potentially helps the system to reach the optimal performance. Our implementation of an example by PaddlePaddle is available at https://github.com/bigdata-ustc/Long-term-Joint-Scheduling, with an explaining video at https://youtu.be/t5M2wVPhTyk.

Joya Chen, Dong Liu, Tong Xu et al.

To train accurate deep object detectors under the extreme foreground-background imbalance, heuristic sampling methods are always necessary, which either re-sample a subset of all training samples (hard sampling methods, \eg biased sampling, OHEM), or use all training samples but re-weight them discriminatively (soft sampling methods, \eg Focal Loss, GHM). In this paper, we challenge the necessity of such hard/soft sampling methods for training accurate deep object detectors. While previous studies have shown that training detectors without heuristic sampling methods would significantly degrade accuracy, we reveal that this degradation comes from an unreasonable classification gradient magnitude caused by the imbalance, rather than a lack of re-sampling/re-weighting. Motivated by our discovery, we propose a simple yet effective \emph{Sampling-Free} mechanism to achieve a reasonable classification gradient magnitude by initialization and loss scaling. Unlike heuristic sampling methods with multiple hyperparameters, our Sampling-Free mechanism is fully data diagnostic, without laborious hyperparameters searching. We verify the effectiveness of our method in training anchor-based and anchor-free object detectors, where our method always achieves higher detection accuracy than heuristic sampling methods on COCO and PASCAL VOC datasets. Our Sampling-Free mechanism provides a new perspective to address the foreground-background imbalance. Our code is released at \url{https://github.com/ChenJoya/sampling-free}.

Joya Chen, Ziyun Zeng, Yiqi Lin et al.

Recent video large language models (Video LLMs) often depend on costly human annotations or proprietary model APIs (e.g., GPT-4o) to produce training data, which limits their training at scale. In this paper, we explore large-scale training for Video LLM with cheap automatic speech recognition (ASR) transcripts. Specifically, we propose a novel streaming training approach that densely interleaves the ASR words and video frames according to their timestamps. Compared to previous studies in vision-language representation with ASR, our method naturally fits the streaming characteristics of ASR, thus enabling the model to learn temporally-aligned, fine-grained vision-language modeling. To support the training algorithm, we introduce a data production pipeline to process YouTube videos and their closed captions (CC, same as ASR), resulting in Live-CC-5M dataset for pre-training and Live-WhisperX-526K dataset for high-quality supervised fine-tuning (SFT). Remarkably, even without SFT, the ASR-only pre-trained LiveCC-7B-Base model demonstrates competitive general video QA performance and exhibits a new capability in real-time video commentary. To evaluate this, we carefully design a new LiveSports-3K benchmark, using LLM-as-a-judge to measure the free-form commentary. Experiments show our final LiveCC-7B-Instruct model can surpass advanced 72B models (Qwen2.5-VL-72B-Instruct, LLaVA-Video-72B) in commentary quality even working in a real-time mode. Meanwhile, it achieves state-of-the-art results at the 7B/8B scale on popular video QA benchmarks such as VideoMME and OVOBench, demonstrating the broad generalizability of our approach. All resources of this paper have been released at https://showlab.github.io/livecc.

Joya Chen, Zhaoyang Lv, Shiwei Wu et al.

Recent Large Language Models have been enhanced with vision capabilities, enabling them to comprehend images, videos, and interleaved vision-language content. However, the learning methods of these large multimodal models typically treat videos as predetermined clips, making them less effective and efficient at handling streaming video inputs. In this paper, we propose a novel Learning-In-Video-Stream (LIVE) framework, which enables temporally aligned, long-context, and real-time conversation within a continuous video stream. Our LIVE framework comprises comprehensive approaches to achieve video streaming dialogue, encompassing: (1) a training objective designed to perform language modeling for continuous streaming inputs, (2) a data generation scheme that converts offline temporal annotations into a streaming dialogue format, and (3) an optimized inference pipeline to speed up the model responses in real-world video streams. With our LIVE framework, we built VideoLLM-online model upon Llama-2/Llama-3 and demonstrate its significant advantages in processing streaming videos. For instance, on average, our model can support streaming dialogue in a 5-minute video clip at over 10 FPS on an A100 GPU. Moreover, it also showcases state-of-the-art performance on public offline video benchmarks, such as recognition, captioning, and forecasting. The code, model, data, and demo have been made available at https://showlab.github.io/videollm-online.

Shiwei Wu, Chao Zhang, Joya Chen et al.

People's social relationships are often manifested through their surroundings, with certain objects or interactions acting as symbols for specific relationships, e.g., wedding rings, roses, hugs, or holding hands. This brings unique challenges to recognizing social relationships, requiring understanding and capturing the essence of these contexts from visual appearances. However, current methods of social relationship understanding rely on the basic classification paradigm of detected persons and objects, which fails to understand the comprehensive context and often overlooks decisive social factors, especially subtle visual cues. To highlight the social-aware context and intricate details, we propose a novel approach that recognizes \textbf{Con}textual \textbf{So}cial \textbf{R}elationships (\textbf{ConSoR}) from a social cognitive perspective. Specifically, to incorporate social-aware semantics, we build a lightweight adapter upon the frozen CLIP to learn social concepts via our novel multi-modal side adapter tuning mechanism. Further, we construct social-aware descriptive language prompts (e.g., scene, activity, objects, emotions) with social relationships for each image, and then compel ConSoR to concentrate more intensively on the decisive visual social factors via visual-linguistic contrasting. Impressively, ConSoR outperforms previous methods with a 12.2\% gain on the People-in-Social-Context (PISC) dataset and a 9.8\% increase on the People-in-Photo-Album (PIPA) benchmark. Furthermore, we observe that ConSoR excels at finding critical visual evidence to reveal social relationships.

Joya Chen, Qi Wu, Dong Liu et al.

Recent years have witnessed the remarkable developments made by deep learning techniques for object detection, a fundamentally challenging problem of computer vision. Nevertheless, there are still difficulties in training accurate deep object detectors, one of which is owing to the foreground-background imbalance problem. In this paper, we survey the recent advances about the solutions to the imbalance problem. First, we analyze the characteristics of the imbalance problem in different kinds of deep detectors, including one-stage and two-stage ones. Second, we divide the existing solutions into two categories: sampling heuristics and non-sampling schemes, and review them in detail. Third, we experimentally compare the performance of some state-of-the-art solutions on the COCO benchmark. Promising directions for future work are also discussed.

Joya Chen, Dong Liu, Bin Luo et al.

For a long time, object detectors have suffered from extreme imbalance between foregrounds and backgrounds. While several sampling/reweighting schemes have been explored to alleviate the imbalance, they are usually heuristic and demand laborious hyper-parameters tuning, which is hard to achieve the optimality. In this paper, we first reveal that such the imbalance could be addressed in a learning-based manner. Guided by this illuminating observation, we propose a novel Residual Objectness (ResObj) mechanism that addresses the imbalance by end-to-end optimization, while no further hand-crafted sampling/reweighting is required. Specifically, by applying multiple cascaded objectness-related modules with residual connections, we formulate an elegant consecutive refinement procedure for distinguishing the foregrounds from backgrounds, thereby progressively addressing the imbalance. Extensive experiments present the effectiveness of our method, as well as its compatibility and adaptivity for both region-based and one-stage detectors, namely, the RetinaNet-ResObj, YOLOv3-ResObj and FasterRCNN-ResObj achieve relative 3.6%, 3.9%, 3.2% Average Precision (AP) improvements compared with their vanilla models on COCO, respectively.