Search

We present InstantStyleGaussian, an innovative 3D style transfer method based on the 3D Gaussian Splatting (3DGS) scene representation. By inputting a target style image, it quickly generates new 3D GS scenes. Our approach operates on pre-reconstructed GS scenes, combining diffusion models with an improved iterative dataset update strategy. It utilizes diffusion models to generate target style images, adds these new images to the training dataset, and uses this dataset to iteratively update and optimize the GS scenes. Extensive experimental results demonstrate that our method ensures high-quality stylized scenes while offering significant advantages in style transfer speed and consistency.

Thermal imaging has a variety of applications, from agricultural monitoring to building inspection to imaging under poor visibility, such as in low light, fog, and rain. However, reconstructing thermal scenes in 3D presents several challenges due to the comparatively lower resolution and limited features present in long-wave infrared (LWIR) images. To overcome these challenges, we propose a unified framework for scene reconstruction from a set of LWIR and RGB images, using a multispectral radiance field to represent a scene viewed by both visible and infrared cameras, thus leveraging information across both spectra. We calibrate the RGB and infrared cameras with respect to each other, as a preprocessing step using a simple calibration target. We demonstrate our method on real-world sets of RGB and LWIR photographs captured from a handheld thermal camera, showing the effectiveness of our method at scene representation across the visible and infrared spectra. We show that our method is capable of thermal super-resolution, as well as visually removing obstacles to reveal objects that are occluded in either the RGB or thermal channels. Please see https://yvette256.github.io/thermalnerf for video results as well as our code and dataset release., Comment: Presented at ICCP 2024; project page at https://yvette256.github.io/thermalnerf

Efficient inference in Large Language Models (LLMs) is impeded by the growing memory demands of key-value (KV) caching, especially for longer sequences. Traditional KV cache eviction strategies, which prioritize less critical KV-pairs based on attention scores, often degrade generation quality, leading to issues such as context loss or hallucinations. To address this, we introduce Dynamic Discriminative Operations (D2O), a novel method that utilizes two-level discriminative strategies to optimize KV cache size without fine-tuning, while preserving essential context. Initially, by observing varying densities of attention weights between shallow and deep layers, we use this insight to determine which layers should avoid excessive eviction to minimize information loss. Subsequently, for the eviction strategy in each layer, D2O innovatively incorporates a compensation mechanism that maintains a similarity threshold to re-discriminate the importance of previously discarded tokens, determining whether they should be recalled and merged with similar tokens. Our approach not only achieves significant memory savings and enhances inference throughput by more than 3 times but also maintains high-quality long-text generation. Extensive experiments across various benchmarks and LLM architectures have demonstrated that D2O significantly enhances performance with a constrained KV cache budget.

This paper proposes an empirical model of dynamic discrete choice to allow for non-separable time preferences, generalizing the well-known Rust (1987) model. Under weak conditions, we show the existence of value functions and hence well-defined optimal choices. We construct a contraction mapping of the value function and propose an estimation method similar to Rust's nested fixed point algorithm. Finally, we apply the framework to the bus engine replacement data. We improve the fit of the data with our general model and reject the null hypothesis that Harold Zuercher has separable time preferences. Misspecifying an agent's preference as time-separable when it is not leads to biased inferences about structure parameters (such as the agent's risk attitudes) and misleading policy recommendations.

In spite of significant progress in the research of fast radio bursts (FRBs) in recent decade, their origin is still under extensive debate. Investigation on the population of FRBs can provide new insight into this interesting problem. In this paper, based on the first CHIME/FRB catalog, we construct a Bayesian framework to analyze the FRB population, with the selection effect of the CHIME telescope being properly taken into account. The energy function is modeled as the power-law with an exponential cutoff. Four redshift distribution models are considered, i.e., the star formation history (SFH) model, and three time-delayed models (Gaussian delay, log-normal delay, and power-law delay). The free parameters are simultaneously constrained using Bayesian inference method, and the Bayesian information criterion (BIC) is used in model comparison. According to BIC, the log-normal delay model fits the data best. The power-law delay model and Gaussian delay model can also give reasonable fits, although they are not as good as the log-normal delay model. However, the SFH model is strongly disfavored compared with the three time-delayed models. The energy function is tightly constrained and is almost independent of the redshift models, with the best-fitting power-law index $\alpha\approx 1.8$, and cut-off energy $\log(E_c/{\rm erg})\approx 42$. The FRB population shows on average $3\sim 5$ billion years time delay with respect to the SFH. Therefore, the hypothesis that the FRB population traces the SFH is conclusively ruled out., Comment: 12 pages, 2 figures, 2 tables. Accepted by Astrophysical Journal

Land cover information is indispensable for advancing the United Nations' sustainable development goals, and land cover mapping under a more detailed category system would significantly contribute to economic livelihood tracking and environmental degradation measurement. However, the substantial difficulty in acquiring fine-grained training data makes the implementation of this task particularly challenging. Here, we propose to combine fully labeled source domain and weakly labeled target domain for weakly supervised domain adaptation (WSDA). This is beneficial as the utilization of sparse and coarse weak labels can considerably alleviate the labor required for precise and detailed land cover annotation. Specifically, we introduce the Prototype-based pseudo-label Rectification and Expansion (PRE) approach, which leverages the prototypes (i.e., the class-wise feature centroids) as the bridge to connect sparse labels and global feature distributions. According to the feature distances to the prototypes, the confidence of pseudo-labels predicted in the unlabeled regions of the target domain is assessed. This confidence is then utilized to guide the dynamic expansion and rectification of pseudo-labels. Based on PRE, we carry out high categorical resolution land cover mapping for 10 cities in different regions around the world, severally using PlanetScope, Gaofen-1, and Sentinel-2 satellite images. In the study areas, we achieve cross-sensor, cross-category, and cross-continent WSDA, with the overall accuracy exceeding 80%. The promising results indicate that PRE is capable of reducing the dependency of land cover classification on high-quality annotations, thereby improving label efficiency. We expect our work to enable global fine-grained land cover mapping, which in turn promote Earth observation to provide more precise and thorough information for environmental monitoring.

With advanced AI/ML, there has been growing research on explainable AI (XAI) and studies on how humans interact with AI and XAI for effective human-AI collaborative decision-making. However, we still have a lack of understanding of how AI systems and XAI should be first presented to users without technical backgrounds. In this paper, we present the findings of semi-structured interviews with health professionals (n=12) and students (n=4) majoring in medicine and health to study how to improve onboarding with AI and XAI. For the interviews, we built upon human-AI interaction guidelines to create onboarding materials of an AI system for stroke rehabilitation assessment and AI explanations and introduce them to the participants. Our findings reveal that beyond presenting traditional performance metrics on AI, participants desired benchmark information, the practical benefits of AI, and interaction trials to better contextualize AI performance, and refine the objectives and performance of AI. Based on these findings, we highlight directions for improving onboarding with AI and XAI and human-AI collaborative decision-making.

Recently, the strong latent Diffusion Probabilistic Model (DPM) has been applied to high-quality Text-to-Image (T2I) generation (e.g., Stable Diffusion), by injecting the encoded target text prompt into the gradually denoised diffusion image generator. Despite the success of DPM in practice, the mechanism behind it remains to be explored. To fill this blank, we begin by examining the intermediate statuses during the gradual denoising generation process in DPM. The empirical observations indicate, the shape of image is reconstructed after the first few denoising steps, and then the image is filled with details (e.g., texture). The phenomenon is because the low-frequency signal (shape relevant) of the noisy image is not corrupted until the final stage in the forward process (initial stage of generation) of adding noise in DPM. Inspired by the observations, we proceed to explore the influence of each token in the text prompt during the two stages. After a series of experiments of T2I generations conditioned on a set of text prompts. We conclude that in the earlier generation stage, the image is mostly decided by the special token [\texttt{EOS}] in the text prompt, and the information in the text prompt is already conveyed in this stage. After that, the diffusion model completes the details of generated images by information from themselves. Finally, we propose to apply this observation to accelerate the process of T2I generation by properly removing text guidance, which finally accelerates the sampling up to 25\%+.

Parameter-efficient fine-tuning (PEFT) has emerged as a popular approach for adapting pre-trained Vision Transformer (ViT) models to downstream applications. While current PEFT methods achieve parameter efficiency, they overlook GPU memory and time efficiency during both fine-tuning and inference, due to the repeated computation of redundant tokens in the ViT architecture. This falls short of practical requirements for downstream task adaptation. In this paper, we propose \textbf{Sparse-Tuning}, a novel tuning paradigm that substantially enhances both fine-tuning and inference efficiency for pre-trained ViT models. Sparse-Tuning efficiently fine-tunes the pre-trained ViT by sparsely preserving the informative tokens and merging redundant ones, enabling the ViT to focus on the foreground while reducing computational costs on background regions in the images. To accurately distinguish informative tokens from uninformative ones, we introduce a tailored Dense Adapter, which establishes dense connections across different encoder layers in the ViT, thereby enhancing the representational capacity and quality of token sparsification. Empirical results on VTAB-1K, three complete image datasets, and two complete video datasets demonstrate that Sparse-Tuning reduces the GFLOPs to \textbf{62\%-70\%} of the original ViT-B while achieving state-of-the-art performance. Source code is available at \url{https://github.com/liuting20/Sparse-Tuning}.

In this study, we investigate the electronic structure of a spin-triplet superconductor candidate CeRh$_2$As$_2$ using high-resolution angle-resolved photoemission spectroscopy and density functional theory calculations. Notably, Fermi surface nesting hints at connections to magnetic excitation or quadrupole density wave phenomena, elucidating the superconducting mechanisms. Measured band structures reveal primarily localized 4f electrons, with minor itinerant contributions. Additionally, a transition from localized to itinerant behavior and significant c-f hybridization anisotropy underscore the role of f-electrons in shaping electronic properties. These findings deepen our understanding of CeRh$_2$As$_2$'s unconventional superconductivity and magnetism. Further exploration promises advances in superconductivity research., Comment: 6 pages, 3 figures

The cosmic distance duality relation (CDDR), expressed as DL(z) = (1 + z)2DA(z), plays an important role in modern cosmology. In this paper, we propose a new method of testing CDDR using strongly lensed gravitational wave (SLGW) signals. Under the geometric optics approximation, we calculate the gravitational lens effects of two lens models, the point mass and singular isothermal sphere. We use functions of {\eta}1(z) = 1 + {\eta}0z and {\eta}2(z) = 1 + {\eta}0z=(1 + z) to parameterize the deviation of CDDR. By reparameterizing the SLGW waveform with CDDR and the distance-redshift relation, we include the deviation parameters {\eta}0 of CDDR as waveform parameters. We evaluate the ability of this method by calculating the parameter estimation of simulated SLGW signals from massive binary black holes. We apply the Fisher information matrix and Markov Chain Monte Carlo methods to calculate parameter estimation. We find that with only one SLGW signal, the measurement precision of {\eta}0 can reach a considerable level of 0.5-1.3% for {\eta}1(z) and 1.1-2.6% for {\eta}2(z), depending on the lens model and parameters., Comment: 15 pages, 7 figures

Large-scale pre-trained vision models (PVMs) have shown great potential for adaptability across various downstream vision tasks. However, with state-of-the-art PVMs growing to billions or even trillions of parameters, the standard full fine-tuning paradigm is becoming unsustainable due to high computational and storage demands. In response, researchers are exploring parameter-efficient fine-tuning (PEFT), which seeks to exceed the performance of full fine-tuning with minimal parameter modifications. This survey provides a comprehensive overview and future directions for visual PEFT, offering a systematic review of the latest advancements. First, we provide a formal definition of PEFT and discuss model pre-training methods. We then categorize existing methods into three categories: addition-based, partial-based, and unified-based. Finally, we introduce the commonly used datasets and applications and suggest potential future research challenges. A comprehensive collection of resources is available at https://github.com/synbol/Awesome-Parameter-Efficient-Transfer-Learning., Comment: 9 pages, 3 figures, 2 tables

Background: Cognitive biases in clinical decision-making significantly contribute to errors in diagnosis and suboptimal patient outcomes. Addressing these biases presents a formidable challenge in the medical field. Objective: This study explores the role of large language models (LLMs) in mitigating these biases through the utilization of a multi-agent framework. We simulate the clinical decision-making processes through multi-agent conversation and evaluate its efficacy in improving diagnostic accuracy. Methods: A total of 16 published and unpublished case reports where cognitive biases have resulted in misdiagnoses were identified from the literature. In the multi-agent framework, we leveraged GPT-4 to facilitate interactions among four simulated agents to replicate clinical team dynamics. Each agent has a distinct role: 1) To make the final diagnosis after considering the discussions, 2) The devil's advocate and correct confirmation and anchoring bias, 3) The tutor and facilitator of the discussion to reduce premature closure bias, and 4) To record and summarize the findings. A total of 80 simulations were evaluated for the accuracy of initial diagnosis, top differential diagnosis and final two differential diagnoses. Results: In a total of 80 responses evaluating both initial and final diagnoses, the initial diagnosis had an accuracy of 0% (0/80), but following multi-agent discussions, the accuracy for the top differential diagnosis increased to 71.3% (57/80), and for the final two differential diagnoses, to 80.0% (64/80). Conclusions: The framework demonstrated an ability to re-evaluate and correct misconceptions, even in scenarios with misleading initial investigations. The LLM-driven multi-agent conversation framework shows promise in enhancing diagnostic accuracy in diagnostically challenging medical scenarios., Comment: 21 pages, 3 figures

Large-scale pre-trained models have achieved remarkable success in various computer vision tasks. A standard approach to leverage these models is to fine-tune all model parameters for downstream tasks, which poses challenges in terms of computational and storage costs. Recently, inspired by Natural Language Processing (NLP), parameter-efficient transfer learning has been successfully applied to vision tasks. However, most existing techniques primarily focus on single-task adaptation, and despite limited research on multi-task adaptation, these methods often exhibit suboptimal training and inference efficiency. In this paper, we first propose an once-for-all Vision Multi-Task Adapter (VMT-Adapter), which strikes approximately O(1) training and inference efficiency w.r.t task number. Concretely, VMT-Adapter shares the knowledge from multiple tasks to enhance cross-task interaction while preserves task-specific knowledge via independent knowledge extraction modules. Notably, since task-specific modules require few parameters, VMT-Adapter can handle an arbitrary number of tasks with a negligible increase of trainable parameters. We also propose VMT-Adapter-Lite, which further reduces the trainable parameters by learning shared parameters between down- and up-projections. Extensive experiments on four dense scene understanding tasks demonstrate the superiority of VMT-Adapter(-Lite), achieving a 3.96%(1.34%) relative improvement compared to single-task full fine-tuning, while utilizing merely ~1% (0.36%) trainable parameters of the pre-trained model., Comment: Accepted to AAAI2024

Multi-Task Learning (MTL) is designed to train multiple correlated tasks simultaneously, thereby enhancing the performance of individual tasks. Typically, a multi-task network structure consists of a shared backbone and task-specific decoders. However, the complexity of the decoders increases with the number of tasks. To tackle this challenge, we integrate the decoder-free vision-language model CLIP, which exhibits robust zero-shot generalization capability. Recently, parameter-efficient transfer learning methods have been extensively explored with CLIP for adapting to downstream tasks, where prompt tuning showcases strong potential. Nevertheless, these methods solely fine-tune a single modality (text or visual), disrupting the modality structure of CLIP. In this paper, we first propose Multi-modal Alignment Prompt (MmAP) for CLIP, which aligns text and visual modalities during fine-tuning process. Building upon MmAP, we develop an innovative multi-task prompt learning framework. On the one hand, to maximize the complementarity of tasks with high similarity, we utilize a gradient-driven task grouping method that partitions tasks into several disjoint groups and assign a group-shared MmAP to each group. On the other hand, to preserve the unique characteristics of each task, we assign an task-specific MmAP to each task. Comprehensive experiments on two large multi-task learning datasets demonstrate that our method achieves significant performance improvements compared to full fine-tuning while only utilizing approximately 0.09% of trainable parameters., Comment: Accepted by AAAI2024

We investigate the chemical potential effects of the equation of state and the chiral transition in an Einstein-Maxwell-dilaton-scalar system, which is obtained from an improved soft-wall AdS/QCD model coupled with an Einstein-Maxwell-dilaton system. The equations of state obtained from the model are in quantitative agreement with the lattice results at both zero and nonzero chemical potentials. The sensible chiral transition behaviors can be realized in the model. The QCD phase diagram with a CEP has also been obtained from the model.

Semiconductor-based superconducting qubits offer a versatile platform for studying hybrid quantum devices in circuit quantum electrodynamics (cQED) architecture. Most of these cQED experiments utilize coplanar waveguides, where the incorporation of DC gate lines is straightforward. Here, we present a technique for probing gate-tunable hybrid devices using a three-dimensional (3D) microwave cavity. A recess is machined inside the cavity wall for the placement of devices and gate lines. We validate this design using a hybrid device based on an InAs-Al nanowire Josephson junction. The coupling between the device and the cavity is facilitated by a long superconducting strip, the antenna. The Josephson junction and the antenna together form a gatemon qubit. We further demonstrate the gate-tunable cavity shift and two-tone qubit spectroscopy. This technique could be used to probe various quantum devices and materials in a 3D cQED architecture that requires DC gate voltages.

Blind and low-vision (BLV) people rely on GPS-based systems for outdoor navigation. GPS's inaccuracy, however, causes them to veer off track, run into obstacles, and struggle to reach precise destinations. While prior work has made precise navigation possible indoors via hardware installations, enabling this outdoors remains a challenge. Interestingly, many outdoor environments are already instrumented with hardware such as street cameras. In this work, we explore the idea of repurposing existing street cameras for outdoor navigation. Our community-driven approach considers both technical and sociotechnical concerns through engagements with various stakeholders: BLV users, residents, business owners, and Community Board leadership. The resulting system, StreetNav, processes a camera's video feed using computer vision and gives BLV pedestrians real-time navigation assistance. Our evaluations show that StreetNav guides users more precisely than GPS, but its technical performance is sensitive to environmental occlusions and distance from the camera. We discuss future implications for deploying such systems at scale.

The complex background in the soil image collected in the field natural environment will affect the subsequent soil image recognition based on machine vision. Segmenting the soil center area from the soil image can eliminate the influence of the complex background, which is an important preprocessing work for subsequent soil image recognition. For the first time, the deep learning method was applied to soil image segmentation, and the Mask R-CNN model was selected to complete the positioning and segmentation of soil images. Construct a soil image dataset based on the collected soil images, use the EISeg annotation tool to mark the soil area as soil, and save the annotation information; train the Mask R-CNN soil image instance segmentation model. The trained model can obtain accurate segmentation results for soil images, and can show good performance on soil images collected in different environments; the trained instance segmentation model has a loss value of 0.1999 in the training set, and the mAP of the validation set segmentation (IoU=0.5) is 0.8804, and it takes only 0.06s to complete image segmentation based on GPU acceleration, which can meet the real-time segmentation and detection of soil images in the field under natural conditions. You can get our code in the Conclusions. The homepage is https://github.com/YidaMyth., Comment: 4 pages, 5 figures, Published in 2023 3rd International Conference on Consumer Electronics and Computer Engineering

We employ a sample of 135,873 RR Lyrae stars (RRLs) with precise photometric-metallicity and distance estimates from the newly calibrated $P$--$\phi_{31}$--$R_{21}$--[Fe/H] and $Gaia$ $G$-band $P$--$R_{21}$--[Fe/H] absolute magnitude-metallicity relations of Li et al., combined with available proper motions from $Gaia$ EDR3, and 6955 systemic radial velocities from $Gaia$ DR3 and other sources, in order to explore the chemistry and kinematics of the halo of the Milky Way (MW). This sample is ideally suited for characterization of the inner- and outer-halo populations of the stellar halo, free from the bias associated with spectroscopically selected probes, and for estimation of their relative contributions as a function of Galactocentric distance. The results of a Gaussian Mixture-Model analysis of these contributions are broadly consistent with other observational studies of the halo, and with expectations from recent MW simulation studies. We apply the HDBSCAN clustering method to the specific energies and cylindrical actions ($E$, J$_{r}$, J$_{\phi}$, J$_{z}$), identifying 97 Dynamically Tagged Groups (DTGs) of RRLs, and explore their associations with recognized substructures of the MW. The precise photometric-distance determinations ($\delta\, d/d < 5$\%), and the resulting high-quality determination of dynamical parameters, yield highly statistically significant (low) dispersions of [Fe/H] for the stellar members of the DTGs compared to random draws from the full sample, indicating that they share common star-formation and chemical histories, influenced by their birth environments., Comment: 16 pages, 8 figures, submitted to MNRAS