Search

Molecular clouds (MCs) are the birthplaces of new stars in galaxies. A key component of MCs are photodissociation regions (PDRs), where far-ultraviolet radiation plays a crucial role in determining the gas's physical and chemical state. Traditional PDR models assume chemical steady state (CSS), where the rates of H$_2$ formation and photodissociation are balanced. However, real MCs are dynamic and can be out of CSS. In this study, we demonstrate that combining H$_2$ emission lines observed in the far-ultraviolet or infrared with column density observations can be used to derive the rates of H$_2$ formation and photodissociation. We derive analytical formulae that relate these rates to observable quantities, which we validate using synthetic H$_2$ line emission maps derived from the SILCC-Zoom hydrodynamical simulation. Our method estimates integrated H$_2$ formation and dissociation rates to within 29\% accuracy. Our simulations cover a wide dynamic range in H$_2$ formation and photodissociation rates, showing significant deviations from CSS, with 74\% of the MC's mass deviating from CSS by a factor greater than 2. Our analytical formulae can effectively distinguish between regions in and out of CSS. When applied to actual H$_2$ line observations, our method can assess the chemical state of MCs, providing insights into their evolutionary stages and lifetimes., Comment: Submitted to ApJ. **Comments Welcome!** This paper is the first ("Paper-I") in 2-paper series. Paper II, entitled "The Molecular Cloud Lifecycle II: Formation and Destruction of Molecular Clouds Diagnosed via H 2 Fluorescent Emission Emission" by Burkhart et al has been recently accepted for publication in ApJ, and may be found here: arXiv:2402.01587

We introduce LocoTrack, a highly accurate and efficient model designed for the task of tracking any point (TAP) across video sequences. Previous approaches in this task often rely on local 2D correlation maps to establish correspondences from a point in the query image to a local region in the target image, which often struggle with homogeneous regions or repetitive features, leading to matching ambiguities. LocoTrack overcomes this challenge with a novel approach that utilizes all-pair correspondences across regions, i.e., local 4D correlation, to establish precise correspondences, with bidirectional correspondence and matching smoothness significantly enhancing robustness against ambiguities. We also incorporate a lightweight correlation encoder to enhance computational efficiency, and a compact Transformer architecture to integrate long-term temporal information. LocoTrack achieves unmatched accuracy on all TAP-Vid benchmarks and operates at a speed almost 6 times faster than the current state-of-the-art., Comment: ECCV 2024. Project page: https://ku-cvlab.github.io/locotrack Code: https://github.com/KU-CVLAB/locotrack

The vocabulary size in temporal action localization (TAL) is constrained by the scarcity of large-scale annotated datasets. To address this, recent works incorporate powerful pre-trained vision-language models (VLMs), such as CLIP, to perform open-vocabulary TAL (OV-TAL). However, unlike VLMs trained on extensive image/video-text pairs, existing OV-TAL methods still rely on small, fully labeled TAL datasets for training an action localizer. In this paper, we explore the scalability of self-training with unlabeled YouTube videos for OV-TAL. Our self-training approach consists of two stages. First, a class-agnostic action localizer is trained on a human-labeled TAL dataset and used to generate pseudo-labels for unlabeled videos. Second, the large-scale pseudo-labeled dataset is combined with the human-labeled dataset to train the localizer. Extensive experiments demonstrate that leveraging web-scale videos in self-training significantly enhances the generalizability of an action localizer. Additionally, we highlighted issues with existing OV-TAL evaluation schemes and proposed a new evaluation protocol. Code is released at https://github.com/HYUNJS/STOV-TAL

We introduce HyperCLOVA X, a family of large language models (LLMs) tailored to the Korean language and culture, along with competitive capabilities in English, math, and coding. HyperCLOVA X was trained on a balanced mix of Korean, English, and code data, followed by instruction-tuning with high-quality human-annotated datasets while abiding by strict safety guidelines reflecting our commitment to responsible AI. The model is evaluated across various benchmarks, including comprehensive reasoning, knowledge, commonsense, factuality, coding, math, chatting, instruction-following, and harmlessness, in both Korean and English. HyperCLOVA X exhibits strong reasoning capabilities in Korean backed by a deep understanding of the language and cultural nuances. Further analysis of the inherent bilingual nature and its extension to multilingualism highlights the model's cross-lingual proficiency and strong generalization ability to untargeted languages, including machine translation between several language pairs and cross-lingual inference tasks. We believe that HyperCLOVA X can provide helpful guidance for regions or countries in developing their sovereign LLMs., Comment: 44 pages; updated authors list and fixed author names

Recent advances in generative models trained on large-scale datasets have made it possible to synthesize high-quality samples across various domains. Moreover, the emergence of strong inversion networks enables not only a reconstruction of real-world images but also the modification of attributes through various editing methods. However, in certain domains related to privacy issues, e.g., human faces, advanced generative models along with strong inversion methods can lead to potential misuses. In this paper, we propose an essential yet under-explored task called generative identity unlearning, which steers the model not to generate an image of a specific identity. In the generative identity unlearning, we target the following objectives: (i) preventing the generation of images with a certain identity, and (ii) preserving the overall quality of the generative model. To satisfy these goals, we propose a novel framework, Generative Unlearning for Any Identity (GUIDE), which prevents the reconstruction of a specific identity by unlearning the generator with only a single image. GUIDE consists of two parts: (i) finding a target point for optimization that un-identifies the source latent code and (ii) novel loss functions that facilitate the unlearning procedure while less affecting the learned distribution. Our extensive experiments demonstrate that our proposed method achieves state-of-the-art performance in the generative machine unlearning task. The code is available at https://github.com/KHU-AGI/GUIDE., Comment: 15 pages, 17 figures, 10 tables, CVPR 2024 Poster

Human matting is a foundation task in image and video processing, where human foreground pixels are extracted from the input. Prior works either improve the accuracy by additional guidance or improve the temporal consistency of a single instance across frames. We propose a new framework MaGGIe, Masked Guided Gradual Human Instance Matting, which predicts alpha mattes progressively for each human instances while maintaining the computational cost, precision, and consistency. Our method leverages modern architectures, including transformer attention and sparse convolution, to output all instance mattes simultaneously without exploding memory and latency. Although keeping constant inference costs in the multiple-instance scenario, our framework achieves robust and versatile performance on our proposed synthesized benchmarks. With the higher quality image and video matting benchmarks, the novel multi-instance synthesis approach from publicly available sources is introduced to increase the generalization of models in real-world scenarios., Comment: CVPR 2024. Project link: https://maggie-matt.github.io

Data augmentations are widely used in training medical image deep learning models to increase the diversity and size of sparse datasets. However, commonly used augmentation techniques can result in loss of clinically relevant information from medical images, leading to incorrect predictions at inference time. We propose the Interactive Medical Image Learning (IMIL) framework, a novel approach for improving the training of medical image analysis algorithms that enables clinician-guided intermediate training data augmentations on misprediction outliers, focusing the algorithm on relevant visual information. To prevent the model from using irrelevant features during training, IMIL will 'blackout' clinician-designated irrelevant regions and replace the original images with the augmented samples. This ensures that for originally mispredicted samples, the algorithm subsequently attends only to relevant regions and correctly correlates them with the respective diagnosis. We validate the efficacy of IMIL using radiology residents and compare its performance to state-of-the-art data augmentations. A 4.2% improvement in accuracy over ResNet-50 was observed when using IMIL on only 4% of the training set. Our study demonstrates the utility of clinician-guided interactive training to achieve meaningful data augmentations for medical image analysis algorithms., Comment: CVPR 2024 Workshop on Domain adaptation, Explainability and Fairness in AI for Medical Image Analysis (DEF-AI-MIA)

While there has been significant progress in customizing text-to-image generation models, generating images that combine multiple personalized concepts remains challenging. In this work, we introduce Concept Weaver, a method for composing customized text-to-image diffusion models at inference time. Specifically, the method breaks the process into two steps: creating a template image aligned with the semantics of input prompts, and then personalizing the template using a concept fusion strategy. The fusion strategy incorporates the appearance of the target concepts into the template image while retaining its structural details. The results indicate that our method can generate multiple custom concepts with higher identity fidelity compared to alternative approaches. Furthermore, the method is shown to seamlessly handle more than two concepts and closely follow the semantic meaning of the input prompt without blending appearances across different subjects., Comment: CVPR 2024

Large language models (LLMs) are susceptible to a variety of risks, from non-faithful output to biased and toxic generations. Due to several limiting factors surrounding LLMs (training cost, API access, data availability, etc.), it may not always be feasible to impose direct safety constraints on a deployed model. Therefore, an efficient and reliable alternative is required. To this end, we present our ongoing efforts to create and deploy a library of detectors: compact and easy-to-build classification models that provide labels for various harms. In addition to the detectors themselves, we discuss a wide range of uses for these detector models - from acting as guardrails to enabling effective AI governance. We also deep dive into inherent challenges in their development and discuss future work aimed at making the detectors more reliable and broadening their scope.

We study the expected sensitivity to the mixing between sterile and tau neutrinos directly from the tau neutrino disappearance in the high-energy fixed target experiment. Here, the beam energy is large enough to produce tau neutrinos at the target with large luminosity. During their propagation to the detector, tau neutrinos may oscillate into sterile neutrinos. By examining the energy spectrum of the observed tau neutrino events, we can probe the mixing between sterile and tau neutrinos directly. In this paper, we consider Scattering and Neutrino Detector (SND) at SHiP experiment as a showcase, which uses 400 GeV protons from SPS at CERN, and expect to observe 7,300 tau and anti-tau neutrinos from the $2\times 10^{20}$ POT for 5 years operation. Assuming the uncertainty of 10\%, we find the sensitivity $|U_{\tau 4}|^2 \sim 0.08$\, (90\% CL) for $\Delta m_{41}^2 \sim 500\ \mathrm{eV}^2$ with 10\% background to the signal. We also consider a far SND at the end of the SHiP Hidden Sector Decay Spectrometer (HSDS), in which case the sensitivity would be enhanced to $|U_{\tau 4}|^2 \sim 0.02$. Away from this mass, the sensitivity becomes lower than $|U_{\tau 4}|^2 \sim 0.15$ for $\Delta m_{41}^2 \lesssim 100\ \mathrm{eV}^2$ or $\Delta m_{41}^2\gtrsim 10^4 \mathrm{eV}^2$., Comment: 14 pages, 8 figures

Molecular hydrogen (H$_2$) formation and dissociation are key processes that drive the gas lifecycle in galaxies. Using the SImulating the LifeCycle of Molecular Clouds (SILCC) zoom-in simulation suite, we explore the utility of future observations of H$_2$ dissociation and formation for tracking the lifecycle of molecular clouds. The simulations used in this work include non-equilibrium H$_2$ formation, stellar radiation, sink particles, and turbulence. We find that, at early times in the cloud evolution, H$_2$ formation rapidly outpaces dissociation and molecular clouds build their mass from the atomic reservoir in their environment. Rapid H$_2$ formation is also associated with a higher early star formation rate. For the clouds studied here, H$_2$ is strongly out of chemical equilibrium during the early stages of cloud formation but settles into a bursty chemical steady-state about 2 Myrs after the first stars form. At the latest stage of cloud evolution, dissociation outweighs formation and the clouds enter a dispersal phase. We discuss how theories for the molecular cloud lifecycle and the star formation efficiency may be distinguished with observational measurements of H$_2$ fluorescence with a space-based high-resolution FUV spectrometer, such as the proposed Hyperion and Eos NASA Explorer missions. Such missions would enable measurements of the H$_2$ dissociation and formation rates, which we demonstrate can be connected to different phases in a molecular cloud's star-forming life, including cloud building, rapidly star-forming, H$_2$ chemical equilibrium, and cloud destruction., Comment: Submitted to ApJ, comments welcome

In stereo-matching knowledge distillation methods of the self-supervised monocular depth estimation, the stereo-matching network's knowledge is distilled into a monocular depth network through pseudo-depth maps. In these methods, the learning-based stereo-confidence network is generally utilized to identify errors in the pseudo-depth maps to prevent transferring the errors. However, the learning-based stereo-confidence networks should be trained with ground truth (GT), which is not feasible in a self-supervised setting. In this paper, we propose a method to identify and filter errors in the pseudo-depth map using multiple disparity maps by checking their consistency without the need for GT and a training process. Experimental results show that the proposed method outperforms the previous methods and works well on various configurations by filtering out erroneous areas where the stereo-matching is vulnerable, especially such as textureless regions, occlusion boundaries, and reflective surfaces., Comment: ICASSP 2024. The first two authors are equally contributed

We propose a novel stereo-confidence that can be measured externally to various stereo-matching networks, offering an alternative input modality choice of the cost volume for learning-based approaches, especially in safety-critical systems. Grounded in the foundational concepts of disparity definition and the disparity plane sweep, the proposed stereo-confidence method is built upon the idea that any shift in a stereo-image pair should be updated in a corresponding amount shift in the disparity map. Based on this idea, the proposed stereo-confidence method can be summarized in three folds. 1) Using the disparity plane sweep, multiple disparity maps can be obtained and treated as a 3-D volume (predicted disparity volume), like the cost volume is constructed. 2) One of these disparity maps serves as an anchor, allowing us to define a desirable (or ideal) disparity profile at every spatial point. 3) By comparing the desirable and predicted disparity profiles, we can quantify the level of matching ambiguity between left and right images for confidence measurement. Extensive experimental results using various stereo-matching networks and datasets demonstrate that the proposed stereo-confidence method not only shows competitive performance on its own but also consistent performance improvements when it is used as an input modality for learning-based stereo-confidence methods., Comment: AAAI 2024. The first two authors contributed equally