Bimanual manipulation is crucial in robotics, enabling complex tasks in industrial automation and household services. However, it poses significant challenges due to the high-dimensional action space and intricate coordination requirements. While video prediction has been recently studied for representation learning and control, leveraging its ability to capture rich dynamic and behavioral information, its potential for enhancing bimanual coordination remains underexplored. To bridge this gap, we propose a unified diffusion-based framework for the joint optimization of video and action prediction. Specifically, we propose a multi-frame latent prediction strategy that encodes future states in a compressed latent space, preserving task-relevant features. Furthermore, we introduce a unidirectional attention mechanism where video prediction is conditioned on the action, while action prediction remains independent of video prediction. This design allows us to omit video prediction during inference, significantly enhancing efficiency. Experiments on two simulated benchmarks and a real-world setting demonstrate a significant improvement in the success rate over the strong baseline ACT using our method, achieving a \textbf{24.9\%} increase on ALOHA, an \textbf{11.1\%} increase on RoboTwin, and a \textbf{32.5\%} increase in real-world experiments. Our models and code are publicly available at https://github.com/return-sleep/Diffusion_based_imaginative_Coordination.

Multimodal Large Language Models (MLLMs) have shown promising progress in understanding and analyzing video content. However, processing long videos remains a significant challenge constrained by LLM's context size. To address this limitation, we propose \textbf{LongVU}, a spatiotemporal adaptive compression mechanism that reduces the number of video tokens while preserving visual details of long videos. Our idea is based on leveraging cross-modal query and inter-frame dependencies to adaptively reduce temporal and spatial redundancy in videos. Specifically, we leverage DINOv2 features to remove redundant frames that exhibit high similarity. Then we utilize text-guided cross-modal query for selective frame feature reduction. Further, we perform spatial token reduction across frames based on their temporal dependencies. Our adaptive compression strategy effectively processes a large number of frames with little visual information loss within given context length. Our LongVU consistently surpass existing methods across a variety of video understanding benchmarks, especially on hour-long video understanding tasks such as VideoMME and MLVU. Given a light-weight LLM, our LongVU also scales effectively into a smaller size with state-of-the-art video understanding performance.

Understanding and reasoning over long videos pose significant challenges for large video language models (LVLMs) due to the difficulty in processing intensive video tokens beyond context window and retaining long-term sequential information. Retrieval-Augmented Generation (RAG) has demonstrated effectiveness in processing long context for Large Language Models (LLMs); however, applying RAG to long video faces challenges such as disrupted temporal dependencies and inclusion of irrelevant information that can hinder accurate reasoning. To address these limitations, we propose Vgent, a novel \textbf{graph-based retrieval-reasoning-augmented generation framework} to enhance LVLMs for long video understanding. Our approach introduces two key innovations: (i) It represents videos by structured graphs with semantic relationships across video clips preserved to improve retrieval effectiveness. (ii) It introduces an intermediate reasoning step to mitigate the reasoning limitation of LVLMs, which leverages structured verification to reduce retrieval noise and facilitate the explicit aggregation of relevant information across clips, resulting in more accurate and context-aware responses. We comprehensively evaluate our framework with various open-source LVLMs on three long-video understanding benchmarks. Our approach yielded an overall performance improvement of $3.0\%\sim 5.4\%$ over base models on MLVU, and outperformed state-of-the-art video RAG methods by $8.6\%$. Our code is publicly available at https://xiaoqian-shen.github.io/Vgent.

Self-supervised learning for computer vision has progressed tremendously and improved many downstream vision tasks, such as image classification, semantic segmentation, and object detection. Among these, generative self-supervised vision learning approaches, such as MAE and BEiT, show promising performance. However, their global reconstruction mechanism is computationally demanding, especially for high-resolution images. The computational cost increases extensively when scaled to a large-scale dataset. To address this issue, we propose local masked reconstruction (LoMaR), a simple yet effective approach that reconstructs image patches from small neighboring regions. The strategy can be easily integrated into any generative self-supervised learning techniques and improves the trade-off between efficiency and accuracy compared to reconstruction over the entire image. LoMaR is$2.5\times faster$than MAE and 5.0x faster than BEiT on$384\times 384$ImageNet pretraining and surpasses them by 0.2% and 0.8% in accuracy, respectively. It is$2.1\times faster$than MAE on iNaturalist pretraining and gains 0.2% in accuracy. On MS COCO, LoMaR outperforms MAE by 0.5$AP^{box}$on object detection and 0.5$AP^{mask}$on instance segmentation. It also outperforms$MAE$by 0.2% on semantic segmentation. Our code and pretrained models are available at: https://github.com/junchen14/LoMaR.

The recent GPT-4 has demonstrated extraordinary multi-modal abilities, such as directly generating websites from handwritten text and identifying humorous elements within images. These features are rarely observed in previous vision-language models. However, the technical details behind GPT-4 continue to remain undisclosed. We believe that the enhanced multi-modal generation capabilities of GPT-4 stem from the utilization of sophisticated large language models (LLM). To examine this phenomenon, we present MiniGPT-4, which aligns a frozen visual encoder with a frozen advanced LLM, Vicuna, using one projection layer. Our work, for the first time, uncovers that properly aligning the visual features with an advanced large language model can possess numerous advanced multi-modal abilities demonstrated by GPT-4, such as detailed image description generation and website creation from hand-drawn drafts. Furthermore, we also observe other emerging capabilities in MiniGPT-4, including writing stories and poems inspired by given images, teaching users how to cook based on food photos, and so on. In our experiment, we found that the model trained on short image caption pairs could produce unnatural language outputs (e.g., repetition and fragmentation). To address this problem, we curate a detailed image description dataset in the second stage to finetune the model, which consequently improves the model's generation reliability and overall usability.

Monitoring animal behavior can facilitate conservation efforts by providing key insights into wildlife health, population status, and ecosystem function. Automatic recognition of animals and their behaviors is critical for capitalizing on the large unlabeled datasets generated by modern video devices and for accelerating monitoring efforts at scale. However, the development of automated recognition systems is currently hindered by a lack of appropriately labeled datasets. Existing video datasets 1) do not classify animals according to established biological taxonomies; 2) are too small to facilitate large-scale behavioral studies and are often limited to a single species; and 3) do not feature temporally localized annotations and therefore do not facilitate localization of targeted behaviors within longer video sequences. Thus, we propose MammalNet, a new large-scale animal behavior dataset with taxonomy-guided annotations of mammals and their common behaviors. MammalNet contains over 18K videos totaling 539 hours, which is ~10 times larger than the largest existing animal behavior dataset [36]. It covers 17 orders, 69 families, and 173 mammal categories for animal categorization and captures 12 high-level animal behaviors that received focus in previous animal behavior studies. We establish three benchmarks on MammalNet: standard animal and behavior recognition, compositional low-shot animal and behavior recognition, and behavior detection. Our dataset and code have been made available at: https://mammalnet.github.io.

Semantic segmentation is a crucial task in computer vision that involves segmenting images into semantically meaningful regions at the pixel level. However, existing approaches often rely on expensive human annotations as supervision for model training, limiting their scalability to large, unlabeled datasets. To address this challenge, we present ZeroSeg, a novel method that leverages the existing pretrained vision-language (VL) model (e.g. CLIP vision encoder [39]) to train open-vocabulary zero-shot semantic segmentation models. Although acquired extensive knowledge of visual concepts, it is non-trivial to exploit knowledge from these VL models to the task of semantic segmentation, as they are usually trained at an image level. ZeroSeg overcomes this by distilling the visual concepts learned by VL models into a set of segment tokens, each summarizing a localized region of the target image. We evaluate ZeroSeg on multiple popular segmentation benchmarks, including PASCAL VOC 2012, PASCAL Context, and COCO, in a zero-shot manner Our approach achieves state-of-the-art performance when compared to other zero-shot segmentation methods under the same training data, while also performing competitively compared to strongly supervised methods. Finally, we also demonstrated the effectiveness of ZeroSeg on open-vocabulary segmentation, through both human studies and qualitative visualizations. The code is publicly available at https://github.com/facebookresearch/ZeroSeg

The visual relationship recognition (VRR) task aims at understanding the pairwise visual relationships between interacting objects in an image. These relationships typically have a long-tail distribution due to their compositional nature. This problem gets more severe when the vocabulary becomes large, rendering this task very challenging. This paper shows that modeling an effective message-passing flow through an attention mechanism can be critical to tackling the compositionality and long-tail challenges in VRR. The method, called RelTransformer, represents each image as a fully-connected scene graph and restructures the whole scene into the relation-triplet and global-scene contexts. It directly passes the message from each element in the relation-triplet and global-scene contexts to the target relation via self-attention. We also design a learnable memory to augment the long-tail relation representation learning. Through extensive experiments, we find that our model generalizes well on many VRR benchmarks. Our model outperforms the best-performing models on two large-scale long-tail VRR benchmarks, VG8K-LT (+2.0% overall acc) and GQA-LT (+26.0% overall acc), both having a highly skewed distribution towards the tail. It also achieves strong results on the VG200 relation detection task. Our code is available at https://github.com/Vision-CAIR/ReITransformer.

The limited availability of annotated data often hinders real-world applications of machine learning. To efficiently learn from small quantities of multimodal data, we leverage the linguistic knowledge from a large pre-trained language model (PLM) and quickly adapt it to new domains of image captioning. To effectively utilize a pretrained model, it is critical to balance the visual input and prior linguistic knowledge from pretraining. We propose VisualGPT, which employs a novel self-resurrecting encoder-decoder attention mechanism to quickly adapt the PLM with a small amount of in-domain image-text data. The proposed self-resurrecting activation unit produces sparse activations that prevent accidental overwriting of linguistic knowledge. When trained on 0.1%, 0.5% and 1% of the respective training sets, VisualGPT surpasses the best baseline by up to 10.0% CIDEr on MS COCO [43] and 17.9% CIDEr on Conceptual Captions [63]. Furthermore, VisualGPT achieves the state-of-the-art result on IU X-ray [15], a medical report generation dataset. Our code is available at https://github.com/Vision-CAIR/VisualGPT.

In this paper, we study fine-grained 3D object identification in real-world scenes described by a textual query. The task aims to discriminatively understand an instance of a particular 3D object described by natural language utterances among other instances of 3D objects of the same class appearing in a visual scene. We introduce the 3DRefTransformer net, a transformer-based neural network that identifies 3D objects described by linguistic utterances in real-world scenes. The network’s input is 3D object segmented point cloud images representing a real-world scene and a language utterance that refers to one of the scene objects. The goal is to identify the referred object. Compared to the state-of-the-art models that are mostly based on graph convolutions and LSTMs, our 3DRefTrans-former net offers two key advantages. First, it is an end-to-end transformer model that operates both on language and 3D visual objects. Second, it has a natural ability to ground textual terms in the utterance to the learning representation of 3D objects in the scene. We further incorporate object pairwise spatial relation loss and contrastive learning during model training. We show in our experiments that our model improves the performance upon the current SOTA significantly on Referit3D Nr3D and Sr3D datasets. Code and Models will be made publicly available at https://vision-cair.github.io/3dreftransformer/.

Several approaches have been proposed in recent literature to alleviate the long-tail problem, mainly in object classification tasks. In this paper, we make the first largescale study concerning the task of Long-Tail Visual Relationship Recognition (LTVRR). LTVRR aims at improving the learning of structured visual relationships that come from the long-tail (e.g., "rabbit grazing on grass"). In this setup, the subject, relation, and object classes each follow a long-tail distribution. To begin our study and make a future benchmark for the community, we introduce two LTVRR-related benchmarks, dubbed VG8K-LT and GQA-LT, built upon the widely used Visual Genome and GQA datasets. We use these benchmarks to study the performance of several state-of-the-art long-tail models on the LTVRR setup. Lastly, we propose a visiolinguistic hubless (VilHub) loss and a Mixup augmentation technique adapted to LTVRR setup, dubbed as RelMix. Both VilHub and RelMix can be easily integrated on top of existing models and despite being simple, our results show that they can remarkably improve the performance, especially on tail classes. Benchmarks, code, and models have been made available at: https://github.com/Vision-CAIR/LTVRR.

MOTIVATION: Over the past years, many computational methods have been developed to incorporate information about phenotypes for disease-gene prioritization task. These methods generally compute the similarity between a patient's phenotypes and a database of gene-phenotype to find the most phenotypically similar match. The main limitation in these methods is their reliance on knowledge about phenotypes associated with particular genes, which is not complete in humans as well as in many model organisms, such as the mouse and fish. Information about functions of gene products and anatomical site of gene expression is available for more genes and can also be related to phenotypes through ontologies and machine-learning models. RESULTS: We developed a novel graph-based machine-learning method for biomedical ontologies, which is able to exploit axioms in ontologies and other graph-structured data. Using our machine-learning method, we embed genes based on their associated phenotypes, functions of the gene products and anatomical location of gene expression. We then develop a machine-learning model to predict gene-disease associations based on the associations between genes and multiple biomedical ontologies, and this model significantly improves over state-of-the-art methods. Furthermore, we extend phenotype-based gene prioritization methods significantly to all genes, which are associated with phenotypes, functions or site of expression. AVAILABILITY AND IMPLEMENTATION: Software and data are available at https://github.com/bio-ontology-research-group/DL2Vec. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Infectious diseases caused by novel viruses have become a major public health concern. Rapid identification of virus-host interactions can reveal mechanistic insights into infectious diseases and shed light on potential treatments. Current computational prediction methods for novel viruses are based mainly on protein sequences. However, it is not clear to what extent other important features, such as the symptoms caused by the viruses, could contribute to a predictor. Disease phenotypes (i.e. signs and symptoms) are readily accessible from clinical diagnosis and we hypothesize that they may act as a potential proxy and an additional source of information for the underlying molecular interactions between the pathogens and hosts. RESULTS: We developed DeepViral, a deep learning based method that predicts protein-protein interactions (PPI) between humans and viruses. Motivated by the potential utility of infectious disease phenotypes, we first embedded human proteins and viruses in a shared space using their associated phenotypes and functions, supported by formalized background knowledge from biomedical ontologies. By jointly learning from protein sequences and phenotype features, DeepViral significantly improves over existing sequence-based methods for intra- and inter-species PPI prediction. AVAILABILITY AND IMPLEMENTATION: Code and datasets for reproduction and customization are available at https://github.com/bio-ontology-research-group/DeepViral. Prediction results for 14 virus families are available at https://doi.org/10.5281/zenodo.4429824. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Understanding the relationships between biomedical terms like viruses, drugs, and symptoms is essential in the fight against diseases. Many attempts have been made to introduce the use of machine learning to the scientific process of hypothesis generation (HG), which refers to the discovery of meaningful implicit connections between biomedical terms. However, most existing methods fail to truly capture the temporal dynamics of scientific term relations and also assume unobserved connections to be irrelevant (i.e., in a positive-negative (PN) learning setting). To break these limits, we formulate this HG problem as future connectivity prediction task on a dynamic attributed graph via positive-unlabeled (PU) learning. Then, the key is to capture the temporal evolution of node pair (term pair) relations from just the positive and unlabeled data. We propose a variational inference model to estimate the positive prior, and incorporate it in the learning of node pair embeddings, which are then used for link prediction. Experiment results on real-world biomedical term relationship datasets and case study analyses on a COVID-19 dataset validate the effectiveness of the proposed model.

A Web-scale knowledge graph (KG) typically contains millions of entities and thousands of entity types. Due to the lack of a pre-defined data schema such as the ER model, entities in KGs are loosely coupled based on their relationships, which brings challenges for effective accesses of the KGs in a structured manner like SPARQL. This demonstration presents an entity-oriented exploratory search prototype system that is able to support search and explore KGs in a exploratory search manner, where local structures of KGs can be dynamically discovered and utilized for guiding users. The system applies a path-based ranking method for recommending similar entities and their relevant information as exploration pointers. The interface is designed to assist users to investigate a domain (particular type) of entities, as well as to explore the knowledge graphs in various relevant domains. The queries are dynamically formulated by tracing the users' dynamic clicking (exploration) behaviors. In this demonstration, we will show how our system visualize the underlying entity structures, as well as explain the semantic correlations among them in a unified interface, which not only assist users to learn about the properties of entities in many aspects but also guide them to further explore the information space.

The entity set expansion problem is to expand a small set of seed entities to a more complete set of similar entities. It can be applied in applications such as web search, item recommendation and query expansion. Traditionally, people solve this problem by exploiting the co-occurrence of entities within web pages, where latent semantic correlation among seed entities cannot be revealed. We propose a novel approach to solve the problem using knowledge graphs, by considering the deficiency (e.g., incompleteness) of knowledge graphs. We design an effective ranking model based on the semantic features of seeds to retrieve the candidate entities. Extensive experiments on public datasets show that the proposed solution significantly outperforms the state-of-the-art techniques.

Large-scale knowledge graphs (KGs) contain massive entities and abundant relations among the entities. Data exploration over KGs allows users to browse the attributes of entities as well as the relations among entities. It therefore provides a good way of learning the structure and coverage of KGs. In this paper, we introduce a system called SEED that is designed to support entity-oriented exploration in large-scale KGs, based on retrieving similar entities of some seed entities as well as their semantic relations that show how entities are similar to each other. A by-product of entity exploration in SEED is to facilitate discovering the deficiency of KGs, so that the detected bugs can be easily fixed by users as they explore the KGs.