We consider the problem of transferring a temporal action segmentation system initially designed for exocentric (fixed) cameras to an egocentric scenario, where wearable cameras capture video data. The conventional supervised approach requires the collection and labeling of a new set of egocentric videos to adapt the model, which is costly and time-consuming. Instead, we propose a novel methodology which performs the adaptation leveraging existing labeled exocentric videos and a new set of unlabeled, synchronized exocentric-egocentric video pairs, for which temporal action segmentation annotations do not need to be collected. We implement the proposed methodology with an approach based on knowledge distillation, which we investigate both at the feature and model level. To evaluate our approach, we introduce a new benchmark based on the Assembly101 dataset. Results demonstrate the feasibility and effectiveness of the proposed method against classic unsupervised domain adaptation and temporal sequence alignment approaches. Remarkably, without bells and whistles, our best model performs on par with supervised approaches trained on labeled egocentric data, without ever seeing a single egocentric label, achieving a +15.99% (28.59% vs 12.60%) improvement in the edit score on the Assembly101 dataset compared to a baseline model trained solely on exocentric data.
相關內容
Long patch validation time is a limiting factor for automated program repair (APR). Though the duality between patch validation and mutation testing is recognized, so far there exists no study of systematically adapting mutation testing techniques to general-purpose patch validation. To address this gap, we investigate existing mutation testing techniques and identify five classes of acceleration techniques that are suitable for general-purpose patch validation. Among them, mutant schemata and mutant deduplication have not been adapted to general-purpose patch validation due to the arbitrary changes that third-party APR approaches may introduce. This presents two problems for adaption: 1) the difficulty of implementing the static equivalence analysis required by the state-of-the-art mutant deduplication approach; 2) the difficulty of capturing the changes of patches to the system state at runtime. To overcome these problems, we propose two novel approaches: 1) execution scheduling, which detects the equivalence between patches online, avoiding the static equivalence analysis and its imprecision; 2) interception-based instrumentation, which intercepts the changes of patches to the system state, avoiding a full interpreter and its overhead. Based on the contributions above, we implement ExpressAPR, a general-purpose patch validator for Java that integrates all recognized classes of techniques suitable for patch validation. Our large-scale evaluation with four APR approaches shows that ExpressAPR accelerates patch validation by 137.1x over plainvalidation or 8.8x over the state-of-the-art approach, making patch validation no longer the time bottleneck of APR. Patch validation time for a single bug can be reduced to within a few minutes on mainstream CPUs.
Advancements in generative models have sparked significant interest in generating images while adhering to specific structural guidelines. Scene graph to image generation is one such task of generating images which are consistent with the given scene graph. However, the complexity of visual scenes poses a challenge in accurately aligning objects based on specified relations within the scene graph. Existing methods approach this task by first predicting a scene layout and generating images from these layouts using adversarial training. In this work, we introduce a novel approach to generate images from scene graphs which eliminates the need of predicting intermediate layouts. We leverage pre-trained text-to-image diffusion models and CLIP guidance to translate graph knowledge into images. Towards this, we first pre-train our graph encoder to align graph features with CLIP features of corresponding images using a GAN based training. Further, we fuse the graph features with CLIP embedding of object labels present in the given scene graph to create a graph consistent CLIP guided conditioning signal. In the conditioning input, object embeddings provide coarse structure of the image and graph features provide structural alignment based on relationships among objects. Finally, we fine tune a pre-trained diffusion model with the graph consistent conditioning signal with reconstruction and CLIP alignment loss. Elaborate experiments reveal that our method outperforms existing methods on standard benchmarks of COCO-stuff and Visual Genome dataset.
In randomized controlled trials (RCT) with time-to-event outcomes, intercurrent events occur as semi-competing/competing events, and they could affect the hazard of outcomes or render outcomes ill-defined. Although five strategies have been proposed in ICH E9 (R1) addendum to address intercurrent events in RCT, they did not readily extend to the context of time-to-event data for studying causal effects with rigorously stated implications. In this study, we show how to define, estimate, and infer the time-dependent cumulative incidence of outcome events in such contexts for obtaining causal interpretations. Specifically, we derive the mathematical forms of the scientific objective (i.e., causal estimands) under the five strategies and clarify the required data structure to identify these causal estimands. Furthermore, we summarize estimation and inference methods for these causal estimands by adopting methodologies in survival analysis, including analytic formulas for asymptotic analysis and hypothesis testing. We illustrate our methods with the LEADER Trial on investigating the effect of liraglutide on cardiovascular outcomes. Studies of multiple endpoints and combining strategies to address multiple intercurrent events can help practitioners understand treatment effects more comprehensively.
The demand for accurate object detection in aerial imagery has surged with the widespread use of drones and satellite technology. Traditional object detection models, trained on datasets biased towards large objects, struggle to perform optimally in aerial scenarios where small, densely clustered objects are prevalent. To address this challenge, we present an innovative approach that combines super-resolution and an adapted lightweight YOLOv5 architecture. We employ a range of datasets, including VisDrone-2023, SeaDroneSee, VEDAI, and NWPU VHR-10, to evaluate our model's performance. Our Super Resolved YOLOv5 architecture features Transformer encoder blocks, allowing the model to capture global context and context information, leading to improved detection results, especially in high-density, occluded conditions. This lightweight model not only delivers improved accuracy but also ensures efficient resource utilization, making it well-suited for real-time applications. Our experimental results demonstrate the model's superior performance in detecting small and densely clustered objects, underlining the significance of dataset choice and architectural adaptation for this specific task. In particular, the method achieves 52.5% mAP on VisDrone, exceeding top prior works. This approach promises to significantly advance object detection in aerial imagery, contributing to more accurate and reliable results in a variety of real-world applications.
Leveraging sensing modalities across diverse spatial and temporal resolutions can improve performance of robotic manipulation tasks. Multi-spatial resolution sensing provides hierarchical information captured at different spatial scales and enables both coarse and precise motions. Simultaneously multi-temporal resolution sensing enables the agent to exhibit high reactivity and real-time control. In this work, we propose a framework, MResT (Multi-Resolution Transformer), for learning generalizable language-conditioned multi-task policies that utilize sensing at different spatial and temporal resolutions using networks of varying capacities to effectively perform real time control of precise and reactive tasks. We leverage off-the-shelf pretrained vision-language models to operate on low-frequency global features along with small non-pretrained models to adapt to high frequency local feedback. Through extensive experiments in 3 domains (coarse, precise and dynamic manipulation tasks), we show that our approach significantly improves (2X on average) over recent multi-task baselines. Further, our approach generalizes well to visual and geometric variations in target objects and to varying interaction forces.
Advancements in generative models have sparked significant interest in generating images while adhering to specific structural guidelines. Scene graph to image generation is one such task of generating images which are consistent with the given scene graph. However, the complexity of visual scenes poses a challenge in accurately aligning objects based on specified relations within the scene graph. Existing methods approach this task by first predicting a scene layout and generating images from these layouts using adversarial training. In this work, we introduce a novel approach to generate images from scene graphs which eliminates the need of predicting intermediate layouts. We leverage pre-trained text-to-image diffusion models and CLIP guidance to translate graph knowledge into images. Towards this, we first pre-train our graph encoder to align graph features with CLIP features of corresponding images using a GAN based training. Further, we fuse the graph features with CLIP embedding of object labels present in the given scene graph to create a graph consistent CLIP guided conditioning signal. In the conditioning input, object embeddings provide coarse structure of the image and graph features provide structural alignment based on relationships among objects. Finally, we fine tune a pre-trained diffusion model with the graph consistent conditioning signal with reconstruction and CLIP alignment loss. Elaborate experiments reveal that our method outperforms existing methods on standard benchmarks of COCO-stuff and Visual Genome dataset.
A common limitation of autonomous tissue manipulation in robotic minimally invasive surgery (MIS) is the absence of force sensing and control at the tool level. Recently, our team has developed haptics-enabled forceps that can simultaneously measure the grasping and pulling forces during tissue manipulation. Based on this design, here we further present a method to automate tissue traction with controlled grasping and pulling forces. Specifically, the grasping stage relies on a controlled grasping force, while the pulling stage is under the guidance of a controlled pulling force. Notably, during the pulling process, the simultaneous control of both grasping and pulling forces is also enabled for more precise tissue traction, achieved through force decoupling. The force controller is built upon a static model of tissue manipulation, considering the interaction between the haptics-enabled forceps and soft tissue. The efficacy of this force control approach is validated through a series of experiments comparing targeted, estimated, and actual reference forces. To verify the feasibility of the proposed method in surgical applications, various tissue resections are conducted on ex vivo tissues employing a dual-arm robotic setup. Finally, we discuss the benefits of multi-force control in tissue traction, evidenced through comparative analyses of various ex vivo tissue resections. The results affirm the feasibility of implementing automatic tissue traction using micro-sized forceps with multi-force control, suggesting its potential to promote autonomous MIS. A video demonstrating the experiments can be found at //youtu.be/8fe8o8IFrjE.
We study the problem of completing various visual document understanding (VDU) tasks, e.g., question answering and information extraction, on real-world documents through human-written instructions. To this end, we propose InstructDoc, the first large-scale collection of 30 publicly available VDU datasets, each with diverse instructions in a unified format, which covers a wide range of 12 tasks and includes open document types/formats. Furthermore, to enhance the generalization performance on VDU tasks, we design a new instruction-based document reading and understanding model, InstructDr, that connects document images, image encoders, and large language models (LLMs) through a trainable bridging module. Experiments demonstrate that InstructDr can effectively adapt to new VDU datasets, tasks, and domains via given instructions and outperforms existing multimodal LLMs and ChatGPT without specific training.
We propose a robust and reliable evaluation metric for generative models by introducing topological and statistical treatments for rigorous support estimation. Existing metrics, such as Inception Score (IS), Frechet Inception Distance (FID), and the variants of Precision and Recall (P&R), heavily rely on supports that are estimated from sample features. However, the reliability of their estimation has not been seriously discussed (and overlooked) even though the quality of the evaluation entirely depends on it. In this paper, we propose Topological Precision and Recall (TopP&R, pronounced 'topper'), which provides a systematic approach to estimating supports, retaining only topologically and statistically important features with a certain level of confidence. This not only makes TopP&R strong for noisy features, but also provides statistical consistency. Our theoretical and experimental results show that TopP&R is robust to outliers and non-independent and identically distributed (Non-IID) perturbations, while accurately capturing the true trend of change in samples. To the best of our knowledge, this is the first evaluation metric focused on the robust estimation of the support and provides its statistical consistency under noise.
The ubiquity of large-scale Pre-Trained Models (PTMs) is on the rise, sparking interest in model hubs, and dedicated platforms for hosting PTMs. Despite this trend, a comprehensive exploration of the challenges that users encounter and how the community leverages PTMs remains lacking. To address this gap, we conducted an extensive mixed-methods empirical study by focusing on discussion forums and the model hub of HuggingFace, the largest public model hub. Based on our qualitative analysis, we present a taxonomy of the challenges and benefits associated with PTM reuse within this community. We then conduct a quantitative study to track model-type trends and model documentation evolution over time. Our findings highlight prevalent challenges such as limited guidance for beginner users, struggles with model output comprehensibility in training or inference, and a lack of model understanding. We also identified interesting trends among models where some models maintain high upload rates despite a decline in topics related to them. Additionally, we found that despite the introduction of model documentation tools, its quantity has not increased over time, leading to difficulties in model comprehension and selection among users. Our study sheds light on new challenges in reusing PTMs that were not reported before and we provide recommendations for various stakeholders involved in PTM reuse.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191