arXiv subjects

Yu Liu

Publications and source records attributed to Yu Liu.

At least 595 records · Page 33Linked to original sources

arXiv2024

Comparison and Analysis of Cognitive Load under 2D/3D Visual Stimuli

With the increasing prevalence of 3D videos, investigating the differences of viewing experiences between 2D and 3D videos has become an important issue. In this study, we explored the cognitive load induced by 2D and 3D video stimuli under various cognitive tasks utilizing electroencephalogram (EEG) data. We also introduced the Cognitive Load Index (CLI), a metric which combines θ and α oscillations to evaluate the cognitive differences. Four video stimuli, each associated with typical cognitive tasks were adopted in our experiments. Subjects were exposed to both 2D and 3D video stimuli, and the corresponding EEG data were recorded. Then, we analyzed the power within the 0.5-45 Hz frequency of EEG data, and CLI was utilized to evaluate the brain activity of different subjects. According to our experiments and analysis, videos that involve simple observational tasks (P <0.05) consistently induced a higher cognitive load in subjects when they were viewing 3D videos. However, for videos that involve calculation tasks (P >0.05), the differences in cognitive load induced by 2D and 3D video were not obvious. Thus, we concluded that 3D videos could generally induce a higher cognitive load, but the extent of the differences also depended on the contents of the video stimuli and the viewing purpose.

q-bio.NC↗

arXiv2024

Learning Disentangled Identifiers for Action-Customized Text-to-Image Generation

This study focuses on a novel task in text-to-image (T2I) generation, namely action customization. The objective of this task is to learn the co-existing action from limited data and generalize it to unseen humans or even animals. Experimental results show that existing subject-driven customization methods fail to learn the representative characteristics of actions and struggle in decoupling actions from context features, including appearance. To overcome the preference for low-level features and the entanglement of high-level features, we propose an inversion-based method Action-Disentangled Identifier (ADI) to learn action-specific identifiers from the exemplar images. ADI first expands the semantic conditioning space by introducing layer-wise identifier tokens, thereby increasing the representational richness while distributing the inversion across different features. Then, to block the inversion of action-agnostic features, ADI extracts the gradient invariance from the constructed sample triples and masks the updates of irrelevant channels. To comprehensively evaluate the task, we present an ActionBench that includes a variety of actions, each accompanied by meticulously selected samples. Both quantitative and qualitative results show that our ADI outperforms existing baselines in action-customized T2I generation. Our project page is at https://adi-t2i.github.io/ADI.

cs.CV↗

arXiv2024

Novel Class Discovery for Ultra-Fine-Grained Visual Categorization

Ultra-fine-grained visual categorization (Ultra-FGVC) aims at distinguishing highly similar sub-categories within fine-grained objects, such as different soybean cultivars. Compared to traditional fine-grained visual categorization, Ultra-FGVC encounters more hurdles due to the small inter-class and large intra-class variation. Given these challenges, relying on human annotation for Ultra-FGVC is impractical. To this end, our work introduces a novel task termed Ultra-Fine-Grained Novel Class Discovery (UFG-NCD), which leverages partially annotated data to identify new categories of unlabeled images for Ultra-FGVC. To tackle this problem, we devise a Region-Aligned Proxy Learning (RAPL) framework, which comprises a Channel-wise Region Alignment (CRA) module and a Semi-Supervised Proxy Learning (SemiPL) strategy. The CRA module is designed to extract and utilize discriminative features from local regions, facilitating knowledge transfer from labeled to unlabeled classes. Furthermore, SemiPL strengthens representation learning and knowledge transfer with proxy-guided supervised learning and proxy-guided contrastive learning. Such techniques leverage class distribution information in the embedding space, improving the mining of subtle differences between labeled and unlabeled ultra-fine-grained classes. Extensive experiments demonstrate that RAPL significantly outperforms baselines across various datasets, indicating its effectiveness in handling the challenges of UFG-NCD. Code is available at https://github.com/SSDUT-Caiyq/UFG-NCD.

cs.CV↗

arXiv2024

AnyDoor: Zero-shot Object-level Image Customization

This work presents AnyDoor, a diffusion-based image generator with the power to teleport target objects to new scenes at user-specified locations in a harmonious way. Instead of tuning parameters for each object, our model is trained only once and effortlessly generalizes to diverse object-scene combinations at the inference stage. Such a challenging zero-shot setting requires an adequate characterization of a certain object. To this end, we complement the commonly used identity feature with detail features, which are carefully designed to maintain texture details yet allow versatile local variations (e.g., lighting, orientation, posture, etc.), supporting the object in favorably blending with different surroundings. We further propose to borrow knowledge from video datasets, where we can observe various forms (i.e., along the time axis) of a single object, leading to stronger model generalizability and robustness. Extensive experiments demonstrate the superiority of our approach over existing alternatives as well as its great potential in real-world applications, such as virtual try-on and object moving. Project page is https://damo-vilab.github.io/AnyDoor-Page/.

cs.CV↗

arXiv2024

Type-printable photodetector arrays for multichannel meta-infrared imaging

Multichannel meta-imaging, inspired by the parallel-processing capability of neuromorphic computing, offers significant advancements in resolution enhancement and edge discrimination in imaging systems, extending even into the mid- to far-infrared spectrum. Currently typical multichannel infrared imaging systems consist of separating optical gratings or merging multi-cameras, which require complex circuit design and heavy power consumption, hindering the implementation of advanced human-eye-like imagers. Here, we present a novel approach for printable graphene plasmonic photodetector arrays driven by a ferroelectric superdomain for multichannel meta-infrared imaging with enhanced edge discrimination. The fabricated photodetectors exhibited multiple spectral responses with zero-bias operation by directly rescaling the ferroelectric superdomain instead of reconstructing the separated gratings. We also demonstrated enhanced and faster shape classification (98.1%) and edge detection (98.2%) using our multichannel infrared images compared with single-channel detectors. Our proof-of-concept photodetector arrays simplify multichannel infrared imaging systems and hold great potential for applications in efficient edge detection in human-brain-type machine vision.

physics.optics↗

arXiv2024

Behavioral analysis in immersive learning environments: A systematic literature review and research agenda

The rapid growth of immersive technologies in educational areas has increased research interest in analyzing the specific behavioral patterns of learners in immersive learning environments. Considering the fact that research on the technical affordances of immersive technologies and the pedagogical affordances of behavioral analysis remains fragmented, this study first contributes by developing a conceptual framework that amalgamates learning requirements, specification, evaluation, and iteration into an integrated model to identify learning benefits and potential hurdles of behavioral analysis in immersive learning environments. Then, a systematic review was conducted underpinning the proposed conceptual framework to retrieve valuable empirical evidence from the 40 eligible articles during the last decade. The review findings suggest that (1) there is an essential need to sufficiently prepare the salient pedagogical requirements to define the specific learning stage, envisage intended cognitive objectives, and specify an appropriate set of learning activities, when developing comprehensive plans on behavioral analysis in immersive learning environments. (2) Researchers could customize the unique immersive experimental implementation by considering factors from four dimensions: learner, pedagogy, context, and representation. (3) The behavioral patterns constructed in immersive learning environments vary by considering the influence of behavioral analysis techniques, research themes, and immersive technical features. (4) The use of behavioral analysis in immersive learning environments faces several challenges from technical, implementation, and data processing perspectives. This study also articulates critical research agenda that could drive future investigation on behavioral analysis in immersive learning environments.

cs.HC↗

arXiv2024

Are Human Rules Necessary? Generating Reusable APIs with CoT Reasoning and In-Context Learning

Inspired by the great potential of Large Language Models (LLMs) for solving complex coding tasks, in this paper, we propose a novel approach, named Code2API, to automatically perform APIzation for Stack Overflow code snippets. Code2API does not require additional model training or any manual crafting rules and can be easily deployed on personal computers without relying on other external tools. Specifically, Code2API guides the LLMs through well-designed prompts to generate well-formed APIs for given code snippets. To elicit knowledge and logical reasoning from LLMs, we used chain-of-thought (CoT) reasoning and few-shot in-context learning, which can help the LLMs fully understand the APIzation task and solve it step by step in a manner similar to a developer. Our evaluations show that Code2API achieves a remarkable accuracy in identifying method parameters (65%) and return statements (66%) equivalent to human-generated ones, surpassing the current state-of-the-art approach, APIzator, by 15.0% and 16.5% respectively. Moreover, compared with APIzator, our user study demonstrates that Code2API exhibits superior performance in generating meaningful method names, even surpassing the human-level performance, and developers are more willing to use APIs generated by our approach, highlighting the applicability of our tool in practice. Finally, we successfully extend our framework to the Python dataset, achieving a comparable performance with Java, which verifies the generalizability of our tool.

cs.SE↗

arXiv2024

A simple model of decision-making in the application process

In decision-making, individuals often rely on intuition, which can occasionally yield suboptimal outcomes. This study examines the impact of intuitive decision-making on individuals who are confronted with limited position information in the job application process. We propose a measure, the mismatch index, that gauges allocation efficiency by comparing the final application rate to the preset admission rate. By simulation and analytical results, we counter-intuitively find that under the intuitive strategy, acquiring more information does not always lead to more efficient allocation. Additionally, a shift from despondency to a bandwagon effect occurs when the initial application rate surpasses the admission rate, which can be observed in our field experiments. Meanwhile, experimental data also unveil variations in individuals' reliance on intuition, indicating the presence of inherent adventurous and conservative inclinations. To account for these effects, we introduce an enhancement factor into our model. The improved results align well with these real data, showing that compared to mediate competitive scenarios, individuals exhibit a stronger conservative tendency in fierce or less competitive scenarios. These findings offer significant insights into resource allocation, especially in the competitive job market context.

physics.soc-ph↗

arXiv2024

Deep Reward Supervisions for Tuning Text-to-Image Diffusion Models

Optimizing a text-to-image diffusion model with a given reward function is an important but underexplored research area. In this study, we propose Deep Reward Tuning (DRTune), an algorithm that directly supervises the final output image of a text-to-image diffusion model and back-propagates through the iterative sampling process to the input noise. We find that training earlier steps in the sampling process is crucial for low-level rewards, and deep supervision can be achieved efficiently and effectively by stopping the gradient of the denoising network input. DRTune is extensively evaluated on various reward models. It consistently outperforms other algorithms, particularly for low-level control signals, where all shallow supervision methods fail. Additionally, we fine-tune Stable Diffusion XL 1.0 (SDXL 1.0) model via DRTune to optimize Human Preference Score v2.1, resulting in the Favorable Diffusion XL 1.0 (FDXL 1.0) model. FDXL 1.0 significantly enhances image quality compared to SDXL 1.0 and reaches comparable quality compared with Midjourney v5.2.

cs.CV↗

arXiv2024

Improving TAS Adaptability with a Variable Temperature Threshold

Thermal-Aware Scheduling (TAS) provides methods to manage the thermal dissipation of a computing chip during task execution. These methods aim to avoid issues such as accelerated aging of the device, premature failure and degraded chip performance. In this work, we implement a new TAS algorithm, VTF-TAS, which makes use of a variable temperature threshold to control task execution and thermal dissipation. To enable adequate execution of the tasks to reach their deadlines, this threshold is managed based on the theory of fluid scheduling. Using an evaluation methodology as described in POD-TAS, we evaluate VTF-TAS using a set of 4 benchmarks from the COMBS benchmark suite to examine its ability to minimize chip temperature throughout schedule execution. Through our evaluation, we demonstrate that this new algorithm is able to adaptively manage the temperature threshold such that the peak temperature during schedule execution is lower than POD-TAS, with no requirement for an expensive search procedure to obtain an optimal threshold for scheduling.

eess.SY↗

arXiv2024

Continuous Invariance Learning

Invariance learning methods aim to learn invariant features in the hope that they generalize under distributional shifts. Although many tasks are naturally characterized by continuous domains, current invariance learning techniques generally assume categorically indexed domains. For example, auto-scaling in cloud computing often needs a CPU utilization prediction model that generalizes across different times (e.g., time of a day and date of a year), where `time' is a continuous domain index. In this paper, we start by theoretically showing that existing invariance learning methods can fail for continuous domain problems. Specifically, the naive solution of splitting continuous domains into discrete ones ignores the underlying relationship among domains, and therefore potentially leads to suboptimal performance. To address this challenge, we then propose Continuous Invariance Learning (CIL), which extracts invariant features across continuously indexed domains. CIL is a novel adversarial procedure that measures and controls the conditional independence between the labels and continuous domain indices given the extracted features. Our theoretical analysis demonstrates the superiority of CIL over existing invariance learning methods. Empirical results on both synthetic and real-world datasets (including data collected from production systems) show that CIL consistently outperforms strong baselines among all the tasks.

cs.LG↗

arXiv2024

Effect Size Estimation for Duration Recommendation in Online Experiments: Leveraging Hierarchical Models and Objective Utility Approaches

The selection of the assumed effect size (AES) critically determines the duration of an experiment, and hence its accuracy and efficiency. Traditionally, experimenters determine AES based on domain knowledge. However, this method becomes impractical for online experimentation services managing numerous experiments, and a more automated approach is hence of great demand. We initiate the study of data-driven AES selection in for online experimentation services by introducing two solutions. The first employs a three-layer Gaussian Mixture Model considering the heteroskedasticity across experiments, and it seeks to estimate the true expected effect size among positive experiments. The second method, grounded in utility theory, aims to determine the optimal effect size by striking a balance between the experiment's cost and the precision of decision-making. Through comparisons with baseline methods using both simulated and real data, we showcase the superior performance of the proposed approaches.

cs.LG↗

arXiv2024

InFusion: Inpainting 3D Gaussians via Learning Depth Completion from Diffusion Prior

3D Gaussians have recently emerged as an efficient representation for novel view synthesis. This work studies its editability with a particular focus on the inpainting task, which aims to supplement an incomplete set of 3D Gaussians with additional points for visually harmonious rendering. Compared to 2D inpainting, the crux of inpainting 3D Gaussians is to figure out the rendering-relevant properties of the introduced points, whose optimization largely benefits from their initial 3D positions. To this end, we propose to guide the point initialization with an image-conditioned depth completion model, which learns to directly restore the depth map based on the observed image. Such a design allows our model to fill in depth values at an aligned scale with the original depth, and also to harness strong generalizability from largescale diffusion prior. Thanks to the more accurate depth completion, our approach, dubbed InFusion, surpasses existing alternatives with sufficiently better fidelity and efficiency under various complex scenarios. We further demonstrate the effectiveness of InFusion with several practical applications, such as inpainting with user-specific texture or with novel object insertion.

cs.CV↗

arXiv2024

Ghost-dil-NetVLAD: A Lightweight Neural Network for Visual Place Recognition

Visual place recognition (VPR) is a challenging task with the unbalance between enormous computational cost and high recognition performance. Thanks to the practical feature extraction ability of the lightweight convolution neural networks (CNNs) and the train-ability of the vector of locally aggregated descriptors (VLAD) layer, we propose a lightweight weakly supervised end-to-end neural network consisting of a front-ended perception model called GhostCNN and a learnable VLAD layer as a back-end. GhostCNN is based on Ghost modules that are lightweight CNN-based architectures. They can generate redundant feature maps using linear operations instead of the traditional convolution process, making a good trade-off between computation resources and recognition accuracy. To enhance our proposed lightweight model further, we add dilated convolutions to the Ghost module to get features containing more spatial semantic information, improving accuracy. Finally, rich experiments conducted on a commonly used public benchmark and our private dataset validate that the proposed neural network reduces the FLOPs and parameters of VGG16-NetVLAD by 99.04% and 80.16%, respectively. Besides, both models achieve similar accuracy.

cs.CV↗

arXiv2024

Predicting Accurate Hot Spots in a More Than Ten-Thousand-Core GPU with a Million-Time Speedup over FEM Enabled by a Physics-based Learning Algorithm

The classical proper orthogonal decomposition (POD) with the Galerkin projection (GP) has been revised for chip-level thermal simulation of microprocessors with a large number of cores. An ensemble POD-GP methodology (EnPOD-GP) is introduced to significantly improve the training effectiveness and prediction accuracy by dividing a large number of heat sources into heat source blocks (HSBs) each of which may contains one or a very small number of heat sources. Although very accurate, efficient and robust to any power map, EnPOD-GP suffers from intensive training for microprocessors with an enormous number of cores. A local-domain EnPOD-GP model (LEnPOD-GP) is thus proposed to further minimize the training burden. LEnPOD-GP utilizes the concepts of local domain truncation and generic building blocks to reduce the massive training data. LEnPOD-GP has been demonstrated on thermal simulation of NVIDIA Tesla Volta GV100, a GPU with more than 13,000 cores including FP32, FP64, INT32, and Tensor Cores. Due to the domain truncation for LEnPOD-GP, the least square error (LSE) is degraded but is still as small as 1.6% over the entire space and below 1.4% in the device layer when using 4 modes per HSB. When only the maximum temperature of the entire GPU is of interest, LEnPOD-GP offers a computing speed 1.1 million times faster than the FEM with a maximum error near 1.2 degrees over the entire simulation time.

cs.CE↗

arXiv2024

GLID: Pre-training a Generalist Encoder-Decoder Vision Model

This paper proposes a GeneraLIst encoder-Decoder (GLID) pre-training method for better handling various downstream computer vision tasks. While self-supervised pre-training approaches, e.g., Masked Autoencoder, have shown success in transfer learning, task-specific sub-architectures are still required to be appended for different downstream tasks, which cannot enjoy the benefits of large-scale pre-training. GLID overcomes this challenge by allowing the pre-trained generalist encoder-decoder to be fine-tuned on various vision tasks with minimal task-specific architecture modifications. In the GLID training scheme, pre-training pretext task and other downstream tasks are modeled as "query-to-answer" problems, including the pre-training pretext task and other downstream tasks. We pre-train a task-agnostic encoder-decoder with query-mask pairs. During fine-tuning, GLID maintains the pre-trained encoder-decoder and queries, only replacing the topmost linear transformation layer with task-specific linear heads. This minimizes the pretrain-finetune architecture inconsistency and enables the pre-trained model to better adapt to downstream tasks. GLID achieves competitive performance on various vision tasks, including object detection, image segmentation, pose estimation, and depth estimation, outperforming or matching specialist models such as Mask2Former, DETR, ViTPose, and BinsFormer.

cs.CV↗

arXiv2024

Ranni: Taming Text-to-Image Diffusion for Accurate Instruction Following

Existing text-to-image (T2I) diffusion models usually struggle in interpreting complex prompts, especially those with quantity, object-attribute binding, and multi-subject descriptions. In this work, we introduce a semantic panel as the middleware in decoding texts to images, supporting the generator to better follow instructions. The panel is obtained through arranging the visual concepts parsed from the input text by the aid of large language models, and then injected into the denoising network as a detailed control signal to complement the text condition. To facilitate text-to-panel learning, we come up with a carefully designed semantic formatting protocol, accompanied by a fully-automatic data preparation pipeline. Thanks to such a design, our approach, which we call Ranni, manages to enhance a pre-trained T2I generator regarding its textual controllability. More importantly, the introduction of the generative middleware brings a more convenient form of interaction (i.e., directly adjusting the elements in the panel or using language instructions) and further allows users to finely customize their generation, based on which we develop a practical system and showcase its potential in continuous generation and chatting-based editing. Our project page is at https://ranni-t2i.github.io/Ranni.

cs.CV↗

arXiv2024

Space Group Constrained Crystal Generation

Crystals are the foundation of numerous scientific and industrial applications. While various learning-based approaches have been proposed for crystal generation, existing methods seldom consider the space group constraint which is crucial in describing the geometry of crystals and closely relevant to many desirable properties. However, considering space group constraint is challenging owing to its diverse and nontrivial forms. In this paper, we reduce the space group constraint into an equivalent formulation that is more tractable to be handcrafted into the generation process. In particular, we translate the space group constraint into two parts: the basis constraint of the invariant logarithmic space of the lattice matrix and the Wyckoff position constraint of the fractional coordinates. Upon the derived constraints, we then propose DiffCSP++, a novel diffusion model that has enhanced a previous work DiffCSP by further taking space group constraint into account. Experiments on several popular datasets verify the benefit of the involvement of the space group constraint, and show that our DiffCSP++ achieves promising performance on crystal structure prediction, ab initio crystal generation and controllable generation with customized space groups.

cs.LG↗