Showing 1–48 of 48 results for author: Hu, V

Diffusion Models and Representation Learning: A Survey

Authors: Michael Fuest, **chuan Ma, Ming Gui, Johannes S. Fischer, Vincent Tao Hu, Bjorn Ommer

Abstract: Diffusion Models are popular generative modeling methods in various vision tasks, attracting significant attention. They can be considered a unique instance of self-supervised learning methods due to their independence from label annotation. This survey explores the interplay between diffusion models and representation learning. It provides an overview of diffusion models' essential aspects, inclu… ▽ More Diffusion Models are popular generative modeling methods in various vision tasks, attracting significant attention. They can be considered a unique instance of self-supervised learning methods due to their independence from label annotation. This survey explores the interplay between diffusion models and representation learning. It provides an overview of diffusion models' essential aspects, including mathematical foundations, popular denoising network architectures, and guidance methods. Various approaches related to diffusion models and representation learning are detailed. These include frameworks that leverage representations learned from pre-trained diffusion models for subsequent recognition tasks and methods that utilize advancements in representation and self-supervised learning to enhance diffusion models. This survey aims to offer a comprehensive overview of the taxonomy between diffusion models and representation learning, identifying key areas of existing concerns and potential exploration. Github link: https://github.com/dongzhuoyao/Diffusion-Representation-Learning-Survey-Taxonomy △ Less

Submitted 30 June, 2024; originally announced July 2024.

Comments: Github Repo: https://github.com/dongzhuoyao/Diffusion-Representation-Learning-Survey-Taxonomy

Insights on the Formation Conditions of Uranus and Neptune from their Deep Elemental Compositions

Authors: Olivier Mousis, Antoine Schneeberger, Thibault Cavalié, Kathleen E. Mandt, Artyom Aguichine, Jonathan I. Lunine, Tom Benest Couzinou, Vincent Hue, Raphaël Moreno

Abstract: This study, placed in the context of the preparation for the Uranus Orbiter Probe mission, aims to predict the bulk volatile compositions of Uranus and Neptune. Using a protoplanetary disk model, it examines the evolution of trace species through vapor and solid transport as dust and pebbles. Due to the high carbon abundance found in their envelopes, the two planets are postulated to have formed a… ▽ More This study, placed in the context of the preparation for the Uranus Orbiter Probe mission, aims to predict the bulk volatile compositions of Uranus and Neptune. Using a protoplanetary disk model, it examines the evolution of trace species through vapor and solid transport as dust and pebbles. Due to the high carbon abundance found in their envelopes, the two planets are postulated to have formed at the carbon monoxide iceline within the protosolar nebula. The time evolution of the abundances of the major volatile species at the location of the CO iceline is then calculated to derive the abundance ratios of the corresponding key elements, including the heavy noble gases, in the feeding zones of Uranus and Neptune. Supersolar metallicity in their envelopes likely results from accreting solids in these zones. Two types of solids are considered: pure condensates (Case 1) and a mixture of pure condensates and clathrates (Case 2). The model, calibrated to observed carbon enrichments, predicts deep compositions. In Case 1, argon is deeply depleted, while nitrogen, oxygen, krypton, phosphorus, sulfur, and xenon are significantly enriched relative to their protosolar abundances in the two planets. Case 2 predicts significant enrichments for all species, including argon, relative to their protosolar abundances. Consequently, Case 1 predicts near-zero Ar/Kr or Ar/Xe ratios, while Case 2 suggests these ratios are 0.1 and 0.5-1 times their protosolar ratios. Both cases predict a bulk sulfur-to-nitrogen ratio consistent with atmospheric measurements. △ Less

Submitted 17 June, 2024; originally announced June 2024.

Comments: Accepted for publication in the Planetary Science Journal

Temperature and composition disturbances in the southern auroral region of Jupiter revealed by JWST/MIRI

Authors: Pablo Rodríguez-Ovalle, Thierry Fouchet, Sandrine Guerlet, Thibault Cavalié, Vincent Hue, Manuel López-Puertas, Emmanuel Lellouch, James A. Sinclair, Imke de Pater, Leigh N. Fletcher, Michael H. Wong, Jake Harkett, Glenn S. Orton, Ricardo Hueso, Agustín Sánchez-Lavega, Tom S. Stallard, Dominique Bockelee-Morvan, Oliver King, Michael T. Roman, Henrik Melin

Abstract: Jupiters south polar region was observed by JWST Mid Infrared Instrument in December 2022. We used the Medium Resolution Spectrometer mode to provide new information about Jupiters South Polar stratosphere. The southern auroral region was visible and influenced the atmosphere in several ways. 1: In the interior of the southern auroral oval, we retrieved peak temperatures at two distinct pressure l… ▽ More Jupiters south polar region was observed by JWST Mid Infrared Instrument in December 2022. We used the Medium Resolution Spectrometer mode to provide new information about Jupiters South Polar stratosphere. The southern auroral region was visible and influenced the atmosphere in several ways. 1: In the interior of the southern auroral oval, we retrieved peak temperatures at two distinct pressure levels near 0.01 and 1 mbar, with warmer temperatures with respect to non auroral regions of 12 pm 2 K and 37 pm 4 K respectively. A cold polar vortex is centered at 65S at 10 mbar. 2: We found that the homopause is elevated to 590+25-118 km above the 1-bar pressure level inside the auroral oval compared to 460+60-50 km at neighboring latitudes and with an upper altitude of 350 km in regions not affected by auroral precipitation. 3: The retrieved abundance of C2H2 shows an increase within the auroral oval, and it exhibits high abundances throughout the polar region. The retrieved abundance of C2H6 increases towards the pole, without being localized in the auroral oval, in contrast with previous analysis. We determined that the warming at 0.01 mbar and the elevated homopause might be caused by the flux of charged particles depositing their energy in the South Polar Region. The 1 mbar hotspot may arise from adiabatic heating resulting from auroral driven downwelling. The cold region at 10 mbar may be caused by radiative cooling by stratospheric aerosols. The differences in spatial distribution seem to indicate that the hydrocarbons analyzed are affected differently by auroral precipitation. △ Less

Submitted 12 June, 2024; originally announced June 2024.

Recipes for forming a carbon-rich giant planet

Authors: Olivier Mousis, Thibault Cavalié, Jonathan I. Lunine, Kathleen E. Mandt, Ricardo Hueso, Artyom Aguichine, Antoine Schneeberger, Tom Benest Couzinou, David H. Atkinson, Vincent Hue, Mark Hofstadter, Udomlerd Srisuchinwong

Abstract: The exploration of carbon-to-oxygen ratios has yielded intriguing insights into the composition of close-in giant exoplanets, giving rise to a distinct classification: carbon-rich planets, characterized by a carbon-to-oxygen ratio $\ge$ 1 in their atmospheres, as opposed to giant planets exhibiting carbon-to-oxygen ratios close to the protosolar value. In contrast, despite numerous space missions… ▽ More The exploration of carbon-to-oxygen ratios has yielded intriguing insights into the composition of close-in giant exoplanets, giving rise to a distinct classification: carbon-rich planets, characterized by a carbon-to-oxygen ratio $\ge$ 1 in their atmospheres, as opposed to giant planets exhibiting carbon-to-oxygen ratios close to the protosolar value. In contrast, despite numerous space missions dispatched to the outer solar system and the proximity of Jupiter, Saturn, Uranus, and Neptune, our understanding of the carbon-to-oxygen ratio in these giants remains notably deficient. Determining this ratio is crucial as it serves as a marker linking a planet's volatile composition directly to its formation region within the disk. This article provides an overview of the current understanding of the carbon-to-oxygen ratio in the four gas giants of our solar system and explores why there is yet no definitive dismissal of the possibility that Jupiter, Saturn, Uranus, or Neptune could be considered carbon-rich planets. Additionally, we delve into the three primary formation scenarios proposed in existing literature to account for a bulk carbon-to-oxygen ratio $\ge$ 1 in a giant planet. A significant challenge lies in accurately inferring the bulk carbon-to-oxygen ratio of our solar system's gas giants. Retrieval methods involve integrating in situ measurements from entry probes equipped with mass spectrometers and remote sensing observations conducted at microwave wavelengths by orbiters. However, these methods fall short of fully discerning the deep carbon-to-oxygen abundance in the gas giants due to their limited probing depth, typically within the 10-100 bar range. △ Less

Submitted 30 May, 2024; originally announced May 2024.

Comments: Space Science Reviews, in press

Properties of electrons accelerated by the Ganymede-magnetosphere interaction: survey of Juno high-latitude observations

Authors: J. Rabia, V. Hue, N. Andre, Q. Nenon, J. R. Szalay, F. Allegrini, A. H. Sulaiman, C. K. Louis, T. K. Greathouse, Y. Sarkango, D. Santos-Costa, M. Blanc, E. Penou, P. Louarn, R. W. Ebert, G. R. Gladstone, A. Mura, J. E. P. Connerney, S. J. Bolton

Abstract: The encounter between the Jovian co-rotating plasma and Ganymede gives rise to electromagnetic waves that propagate along the magnetic field lines and accelerate particles by resonant or non-resonant wave-particle interaction. They ultimately precipitate into Jupiter's atmosphere and trigger auroral emissions. In this study, we use Juno/JADE, Juno/UVS data, and magnetic field line tracing to chara… ▽ More The encounter between the Jovian co-rotating plasma and Ganymede gives rise to electromagnetic waves that propagate along the magnetic field lines and accelerate particles by resonant or non-resonant wave-particle interaction. They ultimately precipitate into Jupiter's atmosphere and trigger auroral emissions. In this study, we use Juno/JADE, Juno/UVS data, and magnetic field line tracing to characterize the properties of electrons accelerated by the Ganymede-magnetosphere interaction in the far-field region. We show that the precipitating energy flux exhibits an exponential decay as a function of downtail distance from the moon, with an e-folding value of 29°, consistent with previous UV observations from the Hubble Space Telescope (HST). We characterize the electron energy distributions and show that two distributions exist. Electrons creating the Main Alfvén Wing (MAW) spot and the auroral tail always have broadband distribution and a mean characteristic energy of 2.2 keV while in the region connected to the Transhemispheric Electron Beam (TEB) spot the electrons are distributed non-monotonically, with a higher characteristic energy above 10 keV. Based on the observation of bidirectional electron beams, we suggest that Juno was located within the acceleration region during the 11 observations reported. We thus estimate that the acceleration region is extended, at least, between an altitude of 0.5 and 1.3 Jupiter radius above the 1-bar surface. Finally, we estimate the size of the interaction region in the Ganymede orbital plane using far-field measurements. These observations provide important insights for the study of particle acceleration processes involved in moon-magnetosphere interactions. △ Less

Submitted 3 May, 2024; originally announced May 2024.

Comments: Accepted for publication in JGR: Space Physics

Effect of magnetospheric conditions on the morphology of Jupiter's UV main auroral emission, as observed by Juno-UVS

Authors: L. A. Head, D. Grodent, B. Bonfond, A. Moirano, B. Benmahi, G. Sicorello, J-C Gérard, M. F. Vogt, V. Hue, T. Greathouse, G. R. Gladstone, Z. Yao

Abstract: Auroral emissions are a reflection of magnetospheric processes, and, at Jupiter, it is not entirely certain how the morphology of the UV main emission (ME) varies with magnetospheric compression or the strength of the central current sheet. This work leverages the observations from Juno-UVS to link ME variability with magnetospheric states. Novel arc-detection techniques are used to determine new… ▽ More Auroral emissions are a reflection of magnetospheric processes, and, at Jupiter, it is not entirely certain how the morphology of the UV main emission (ME) varies with magnetospheric compression or the strength of the central current sheet. This work leverages the observations from Juno-UVS to link ME variability with magnetospheric states. Novel arc-detection techniques are used to determine new reference ovals for the ME from perijoves 1 through 54, in both hemispheres, and analyse how the size and shape of the ME vary compared to this reference oval. The morphology and brightness of the ME vary in local time: the dawn-side ME is typically expanded and the dusk-side ME typically contracted compared to the reference oval, and the dusk-side ME being typically twice as bright as the dawn-side ME. Both the northern and southern ME, and the day-side and night-side ME, expand and contract from their reference ovals synchronously, which indicates that the variable size of the ME is caused by a process occurring throughout the jovian magnetosphere. The poleward latitudinal shift of the auroral footprint of Ganymede correlates with the poleward motion of the ME, whereas a similar relation is not present for the footprint of Io. Additionally, the expansion of the ME correlates well with an increase in magnetodisc current. These two results suggest that a changing current-sheet magnetic field is partially responsible for the variable size of the ME. Finally, magnetospheric compression is linked to a global ME contraction and brightening, though this brightening occurs predominantly in the day-side ME. This observation, and the observation that the dusk-side ME is typically brighter than the dawn-side ME, stands in contrast to the modelled and observed behaviour of field-aligned currents and thus weakens the theoretical link between field-aligned currents and the generation of the auroral ME. △ Less

Submitted 8 April, 2024; v1 submitted 5 April, 2024; originally announced April 2024.

Continuous, Subject-Specific Attribute Control in T2I Models by Identifying Semantic Directions

Authors: Stefan Andreas Baumann, Felix Krause, Michael Neumayr, Nick Stracke, Vincent Tao Hu, Björn Ommer

Abstract: In recent years, advances in text-to-image (T2I) diffusion models have substantially elevated the quality of their generated images. However, achieving fine-grained control over attributes remains a challenge due to the limitations of natural language prompts (such as no continuous set of intermediate descriptions existing between ``person'' and ``old person''). Even though many methods were intro… ▽ More In recent years, advances in text-to-image (T2I) diffusion models have substantially elevated the quality of their generated images. However, achieving fine-grained control over attributes remains a challenge due to the limitations of natural language prompts (such as no continuous set of intermediate descriptions existing between ``person'' and ``old person''). Even though many methods were introduced that augment the model or generation process to enable such control, methods that do not require a fixed reference image are limited to either enabling global fine-grained attribute expression control or coarse attribute expression control localized to specific subjects, not both simultaneously. We show that there exist directions in the commonly used token-level CLIP text embeddings that enable fine-grained subject-specific control of high-level attributes in text-to-image models. Based on this observation, we introduce one efficient optimization-free and one robust optimization-based method to identify these directions for specific attributes from contrastive text prompts. We demonstrate that these directions can be used to augment the prompt text input with fine-grained control over attributes of specific subjects in a compositional manner (control over multiple attributes of a single subject) without having to adapt the diffusion model. Project page: https://compvis.github.io/attribute-control. Code is available at https://github.com/CompVis/attribute-control. △ Less

Submitted 25 March, 2024; originally announced March 2024.

Comments: Project page: https://compvis.github.io/attribute-control

ZigMa: A DiT-style Zigzag Mamba Diffusion Model

Authors: Vincent Tao Hu, Stefan Andreas Baumann, Ming Gui, Olga Grebenkova, **chuan Ma, Johannes Fischer, Björn Ommer

Abstract: The diffusion model has long been plagued by scalability and quadratic complexity issues, especially within transformer-based structures. In this study, we aim to leverage the long sequence modeling capability of a State-Space Model called Mamba to extend its applicability to visual data generation. Firstly, we identify a critical oversight in most current Mamba-based vision methods, namely the la… ▽ More The diffusion model has long been plagued by scalability and quadratic complexity issues, especially within transformer-based structures. In this study, we aim to leverage the long sequence modeling capability of a State-Space Model called Mamba to extend its applicability to visual data generation. Firstly, we identify a critical oversight in most current Mamba-based vision methods, namely the lack of consideration for spatial continuity in the scan scheme of Mamba. Secondly, building upon this insight, we introduce a simple, plug-and-play, zero-parameter method named Zigzag Mamba, which outperforms Mamba-based baselines and demonstrates improved speed and memory utilization compared to transformer-based baselines. Lastly, we integrate Zigzag Mamba with the Stochastic Interpolant framework to investigate the scalability of the model on large-resolution visual datasets, such as FacesHQ $1024\times 1024$ and UCF101, MultiModal-CelebA-HQ, and MS COCO $256\times 256$ . Code will be released at https://taohu.me/zigma/ △ Less

Submitted 1 April, 2024; v1 submitted 20 March, 2024; originally announced March 2024.

Comments: Project Page: https://taohu.me/zigma/

DepthFM: Fast Monocular Depth Estimation with Flow Matching

Authors: Ming Gui, Johannes S. Fischer, Ulrich Prestel, **chuan Ma, Dmytro Kotovenko, Olga Grebenkova, Stefan Andreas Baumann, Vincent Tao Hu, Björn Ommer

Abstract: Monocular depth estimation is crucial for numerous downstream vision tasks and applications. Current discriminative approaches to this problem are limited due to blurry artifacts, while state-of-the-art generative methods suffer from slow sampling due to their SDE nature. Rather than starting from noise, we seek a direct map** from input image to depth map. We observe that this can be effectivel… ▽ More Monocular depth estimation is crucial for numerous downstream vision tasks and applications. Current discriminative approaches to this problem are limited due to blurry artifacts, while state-of-the-art generative methods suffer from slow sampling due to their SDE nature. Rather than starting from noise, we seek a direct map** from input image to depth map. We observe that this can be effectively framed using flow matching, since its straight trajectories through solution space offer efficiency and high quality. Our study demonstrates that a pre-trained image diffusion model can serve as an adequate prior for a flow matching depth model, allowing efficient training on only synthetic data to generalize to real images. We find that an auxiliary surface normals loss further improves the depth estimates. Due to the generative nature of our approach, our model reliably predicts the confidence of its depth estimates. On standard benchmarks of complex natural scenes, our lightweight approach exhibits state-of-the-art performance at favorable low computational cost despite only being trained on little synthetic data. △ Less

Submitted 20 March, 2024; originally announced March 2024.

TMU at TREC Clinical Trials Track 2023

Authors: Aritra Kumar Lahiri, Emrul Hasan, Qinmin Vivian Hu, Cherie Ding

Abstract: This paper describes Toronto Metropolitan University's participation in the TREC Clinical Trials Track for 2023. As part of the tasks, we utilize advanced natural language processing techniques and neural language models in our experiments to retrieve the most relevant clinical trials. We illustrate the overall methodology, experimental settings, and results of our implementation for the run submi… ▽ More This paper describes Toronto Metropolitan University's participation in the TREC Clinical Trials Track for 2023. As part of the tasks, we utilize advanced natural language processing techniques and neural language models in our experiments to retrieve the most relevant clinical trials. We illustrate the overall methodology, experimental settings, and results of our implementation for the run submission as part of Team - V-TorontoMU. △ Less

Submitted 11 March, 2024; originally announced March 2024.

Training Class-Imbalanced Diffusion Model Via Overlap Optimization

Authors: Divin Yan, Lu Qi, Vincent Tao Hu, Ming-Hsuan Yang, Meng Tang

Abstract: Diffusion models have made significant advances recently in high-quality image synthesis and related tasks. However, diffusion models trained on real-world datasets, which often follow long-tailed distributions, yield inferior fidelity for tail classes. Deep generative models, including diffusion models, are biased towards classes with abundant training images. To address the observed appearance o… ▽ More Diffusion models have made significant advances recently in high-quality image synthesis and related tasks. However, diffusion models trained on real-world datasets, which often follow long-tailed distributions, yield inferior fidelity for tail classes. Deep generative models, including diffusion models, are biased towards classes with abundant training images. To address the observed appearance overlap between synthesized images of rare classes and tail classes, we propose a method based on contrastive learning to minimize the overlap between distributions of synthetic images for different classes. We show variants of our probabilistic contrastive learning method can be applied to any class conditional diffusion model. We show significant improvement in image synthesis using our loss for multiple datasets with long-tailed distribution. Extensive experimental results demonstrate that the proposed method can effectively handle imbalanced data for diffusion-based generation and classification models. Our code and datasets will be publicly available at https://github.com/yanliang3612/DiffROP. △ Less

Submitted 16 February, 2024; originally announced February 2024.

Comments: Technique Report

Influence of the Jovian current sheet models on the map** of the UV auroral footprints of Io, Europa, and Ganymede

Authors: Jonas Rabia, Quentin Nénon, Nicolas André, Vincent Hue, Daniel Santos-Costa, Aneesah Kamran, Michel Blanc

Abstract: The in-situ characterization of moon-magnetosphere interactions at Jupiter and the map** of moon auroral footpaths require accurate global models of the magnetospheric magnetic field. In this study, we compare the ability of two widely-used current sheet models, Khurana-2005 (KK2005) and Connerney-2020 (CON2020) combined with the most recent measurements acquired at low, medium, and high latitud… ▽ More The in-situ characterization of moon-magnetosphere interactions at Jupiter and the map** of moon auroral footpaths require accurate global models of the magnetospheric magnetic field. In this study, we compare the ability of two widely-used current sheet models, Khurana-2005 (KK2005) and Connerney-2020 (CON2020) combined with the most recent measurements acquired at low, medium, and high latitudes. With the adjustments of the KK2005 model to JRM33, we show that in the outer and middle magnetosphere (R>15RJ), JRM33+KK2005 is found to be the best model to reproduce the magnetic field observations of Galileo and Juno as it accounts for local time effects. JRM33+CON2020 gives the most accurate representation of the inner magnetosphere. This finding is drawn from comparisons with Juno in-situ magnetic field measurements and confirmed by contrasting the timing of the crossings of the Io, Europa, and Ganymede flux tubes identified in the Juno particles data with the two model estimates. JRM33+CON2020 also maps more accurately the UV auroral footpath of Io, Europa, and Ganymede observed by Juno than JRM33+KK2005. The JRM33+KK2005 model predicts a local time asymmetry in position of the moons' footprints, which is however not detected in Juno's UV measurements.This could indicate that local time effects on the magnetic field are marginal at the orbital locations of Io, Europa, and Ganymede. Finally, the accuracy of the models and their predictions as a function of hemisphere, local time, and longitude is explored. △ Less

Submitted 18 January, 2024; originally announced January 2024.

Comments: 22 pages, 9 figures, 1 table. Accepted for publication in Journal of Geophysical Research: Space physics

Latent Space Editing in Transformer-Based Flow Matching

Authors: Vincent Tao Hu, David W Zhang, Pascal Mettes, Meng Tang, Deli Zhao, Cees G. M. Snoek

Abstract: This paper strives for image editing via generative models. Flow Matching is an emerging generative modeling technique that offers the advantage of simple and efficient training. Simultaneously, a new transformer-based U-ViT has recently been proposed to replace the commonly used UNet for better scalability and performance in generative modeling. Hence, Flow Matching with a transformer backbone of… ▽ More This paper strives for image editing via generative models. Flow Matching is an emerging generative modeling technique that offers the advantage of simple and efficient training. Simultaneously, a new transformer-based U-ViT has recently been proposed to replace the commonly used UNet for better scalability and performance in generative modeling. Hence, Flow Matching with a transformer backbone offers the potential for scalable and high-quality generative modeling, but their latent structure and editing ability are as of yet unknown. Hence, we adopt this setting and explore how to edit images through latent space manipulation. We introduce an editing space, which we call $u$-space, that can be manipulated in a controllable, accumulative, and composable manner. Additionally, we propose a tailored sampling solution to enable sampling with the more efficient adaptive step-size ODE solvers. Lastly, we put forth a straightforward yet powerful method for achieving fine-grained and nuanced editing using text prompts. Our framework is simple and efficient, all while being highly effective at editing images while preserving the essence of the original content. Our code will be publicly available at https://taohu.me/lfm/ △ Less

Submitted 17 December, 2023; originally announced December 2023.

Comments: AAAI 2024 with Appendix

Motion Flow Matching for Human Motion Synthesis and Editing

Authors: Vincent Tao Hu, Wenzhe Yin, **chuan Ma, Yunlu Chen, Basura Fernando, Yuki M Asano, Efstratios Gavves, Pascal Mettes, Bjorn Ommer, Cees G. M. Snoek

Abstract: Human motion synthesis is a fundamental task in computer animation. Recent methods based on diffusion models or GPT structure demonstrate commendable performance but exhibit drawbacks in terms of slow sampling speeds and error accumulation. In this paper, we propose \emph{Motion Flow Matching}, a novel generative model designed for human motion generation featuring efficient sampling and effective… ▽ More Human motion synthesis is a fundamental task in computer animation. Recent methods based on diffusion models or GPT structure demonstrate commendable performance but exhibit drawbacks in terms of slow sampling speeds and error accumulation. In this paper, we propose \emph{Motion Flow Matching}, a novel generative model designed for human motion generation featuring efficient sampling and effectiveness in motion editing applications. Our method reduces the sampling complexity from thousand steps in previous diffusion models to just ten steps, while achieving comparable performance in text-to-motion and action-to-motion generation benchmarks. Noticeably, our approach establishes a new state-of-the-art Fréchet Inception Distance on the KIT-ML dataset. What is more, we tailor a straightforward motion editing paradigm named \emph{sampling trajectory rewriting} leveraging the ODE-style generative models and apply it to various editing scenarios including motion prediction, motion in-between prediction, motion interpolation, and upper-body editing. Our code will be released. △ Less

Submitted 14 December, 2023; originally announced December 2023.

Comments: WIP

Guided Diffusion from Self-Supervised Diffusion Features

Authors: Vincent Tao Hu, Yunlu Chen, Mathilde Caron, Yuki M. Asano, Cees G. M. Snoek, Bjorn Ommer

Abstract: Guidance serves as a key concept in diffusion models, yet its effectiveness is often limited by the need for extra data annotation or classifier pretraining. That is why guidance was harnessed from self-supervised learning backbones, like DINO. However, recent studies have revealed that the feature representation derived from diffusion model itself is discriminative for numerous downstream tasks a… ▽ More Guidance serves as a key concept in diffusion models, yet its effectiveness is often limited by the need for extra data annotation or classifier pretraining. That is why guidance was harnessed from self-supervised learning backbones, like DINO. However, recent studies have revealed that the feature representation derived from diffusion model itself is discriminative for numerous downstream tasks as well, which prompts us to propose a framework to extract guidance from, and specifically for, diffusion models. Our research has yielded several significant contributions. Firstly, the guidance signals from diffusion models are on par with those from class-conditioned diffusion models. Secondly, feature regularization, when based on the Sinkhorn-Knopp algorithm, can further enhance feature discriminability in comparison to unconditional diffusion models. Thirdly, we have constructed an online training approach that can concurrently derive guidance from diffusion models for diffusion models. Lastly, we have extended the application of diffusion models along the constant velocity path of ODE to achieve a more favorable balance between sampling steps and fidelity. The performance of our methods has been outstanding, outperforming related baseline comparisons in large-resolution datasets, such as ImageNet256, ImageNet256-100 and LSUN-Churches. Our code will be released. △ Less

Submitted 14 December, 2023; originally announced December 2023.

Comments: Work In Progress

ScribbleGen: Generative Data Augmentation Improves Scribble-supervised Semantic Segmentation

Authors: Jacob Schnell, Jieke Wang, Lu Qi, Vincent Tao Hu, Meng Tang

Abstract: Recent advances in generative models, such as diffusion models, have made generating high-quality synthetic images widely accessible. Prior works have shown that training on synthetic images improves many perception tasks, such as image classification, object detection, and semantic segmentation. We are the first to explore generative data augmentations for scribble-supervised semantic segmentatio… ▽ More Recent advances in generative models, such as diffusion models, have made generating high-quality synthetic images widely accessible. Prior works have shown that training on synthetic images improves many perception tasks, such as image classification, object detection, and semantic segmentation. We are the first to explore generative data augmentations for scribble-supervised semantic segmentation. We propose ScribbleGen, a generative data augmentation method that leverages a ControlNet diffusion model conditioned on semantic scribbles to produce high-quality training data. However, naive implementations of generative data augmentations may inadvertently harm the performance of the downstream segmentor rather than improve it. We leverage classifier-free diffusion guidance to enforce class consistency and introduce encode ratios to trade off data diversity for data realism. Using the guidance scale and encode ratio, we can generate a spectrum of high-quality training images. We propose multiple augmentation schemes and find that these schemes significantly impact model performance, especially in the low-data regime. Our framework further reduces the gap between the performance of scribble-supervised segmentation and that of fully-supervised segmentation. We also show that our framework significantly improves segmentation performance on small datasets, even surpassing fully-supervised segmentation. The code is available at https://github.com/mengtang-lab/scribblegen. △ Less

Submitted 16 April, 2024; v1 submitted 28 November, 2023; originally announced November 2023.

ToddlerDiffusion: Flash Interpretable Controllable Diffusion Model

Authors: Eslam Mohamed Bakr, Liangbing Zhao, Vincent Tao Hu, Matthieu Cord, Patrick Perez, Mohamed Elhoseiny

Abstract: Diffusion-based generative models excel in perceptually impressive synthesis but face challenges in interpretability. This paper introduces ToddlerDiffusion, an interpretable 2D diffusion image-synthesis framework inspired by the human generation system. Unlike traditional diffusion models with opaque denoising steps, our approach decomposes the generation process into simpler, interpretable stage… ▽ More Diffusion-based generative models excel in perceptually impressive synthesis but face challenges in interpretability. This paper introduces ToddlerDiffusion, an interpretable 2D diffusion image-synthesis framework inspired by the human generation system. Unlike traditional diffusion models with opaque denoising steps, our approach decomposes the generation process into simpler, interpretable stages; generating contours, a palette, and a detailed colored image. This not only enhances overall performance but also enables robust editing and interaction capabilities. Each stage is meticulously formulated for efficiency and accuracy, surpassing Stable-Diffusion (LDM). Extensive experiments on datasets like LSUN-Churches and COCO validate our approach, consistently outperforming existing methods. ToddlerDiffusion achieves notable efficiency, matching LDM performance on LSUN-Churches while operating three times faster with a 3.76 times smaller architecture. Our source code is provided in the supplementary material and will be publicly accessible. △ Less

Submitted 24 November, 2023; originally announced November 2023.

Source of radio emissions induced by the Galilean moons Io, Europa and Ganymede: in situ measurements by Juno

Authors: C. K. Louis, P. Louarn, B. Collet, N. Clément, S. Al Saati, J. R. Szalay, V. Hue, L. Lamy, S. Kotsiaros, W. S. Kurth, C. M. Jackman, Y. Wang, M. Blanc, F. Allegrini, J. E. P. Connerney, D. Gershman

Abstract: At Jupiter, part of the auroral radio emissions are induced by the Galilean moons Io, Europa and Ganymede. Until now, except for Ganymede, they have been only remotely detected, using ground-based radio-telescopes or electric antennas aboard spacecraft. The polar trajectory of the Juno orbiter allows the spacecraft to cross the range of magnetic flux tubes which sustain the various Jupiter-satelli… ▽ More At Jupiter, part of the auroral radio emissions are induced by the Galilean moons Io, Europa and Ganymede. Until now, except for Ganymede, they have been only remotely detected, using ground-based radio-telescopes or electric antennas aboard spacecraft. The polar trajectory of the Juno orbiter allows the spacecraft to cross the range of magnetic flux tubes which sustain the various Jupiter-satellite interactions, and in turn to sample in situ the associated radio emission regions. In this study, we focus on the detection and the characterization of radio sources associated with Io, Europa and Ganymede. Using electric wave measurements or radio observations (Juno/Waves), in situ electron measurements (Juno/JADE-E), and magnetic field measurements (Juno/MAG) we demonstrate that the Cyclotron Maser Instability (CMI) driven by a loss-cone electron distribution function is responsible for the encountered radio sources. We confirmed that radio emissions are associated with Main (MAW) or Reflected Alfvén Wing (RAW), but also show that for Europa and Ganymede, induced radio emissions are associated with Transhemispheric Electron Beam (TEB). For each traversed radio source, we determine the latitudinal extension, the CMI-resonant electron energy, and the bandwidth of the emission. We show that the presence of Alfvén perturbations and downward field aligned currents are necessary for the radio emissions to be amplified. △ Less

Submitted 10 August, 2023; originally announced August 2023.

ReactIE: Enhancing Chemical Reaction Extraction with Weak Supervision

Authors: Ming Zhong, Siru Ouyang, Minhao Jiang, Vivian Hu, Yizhu Jiao, Xuan Wang, Jiawei Han

Abstract: Structured chemical reaction information plays a vital role for chemists engaged in laboratory work and advanced endeavors such as computer-aided drug design. Despite the importance of extracting structured reactions from scientific literature, data annotation for this purpose is cost-prohibitive due to the significant labor required from domain experts. Consequently, the scarcity of sufficient tr… ▽ More Structured chemical reaction information plays a vital role for chemists engaged in laboratory work and advanced endeavors such as computer-aided drug design. Despite the importance of extracting structured reactions from scientific literature, data annotation for this purpose is cost-prohibitive due to the significant labor required from domain experts. Consequently, the scarcity of sufficient training data poses an obstacle to the progress of related models in this domain. In this paper, we propose ReactIE, which combines two weakly supervised approaches for pre-training. Our method utilizes frequent patterns within the text as linguistic cues to identify specific characteristics of chemical reactions. Additionally, we adopt synthetic data from patent records as distant supervision to incorporate domain knowledge into the model. Experiments demonstrate that ReactIE achieves substantial improvements and outperforms all existing baselines. △ Less

Submitted 3 July, 2023; originally announced July 2023.

Comments: Findings of ACL 2023, Short Paper

The Io, Europa and Ganymede auroral footprints at Jupiter in the ultraviolet: positions and equatorial lead angles

Authors: Vincent Hue, Randy Gladstone, Corentin K. Louis, Thomas K. Greathouse, Bertrand Bonfond, Jamey R. Szalay, Alessandro Moirano, Rohini S. Giles, Joshua A. Kammer, Masafumi Imai, Alessandro Mura, Maarten H. Versteeg, George Clark, Jean-Claude Gérard, Denis C. Grodent, Jonas Rabia, Ali H. Sulaiman, Scott J. Bolton, John E. P. Connerney

Abstract: Jupiter's satellite auroral footprints are a consequence of the interaction between the Jovian magnetic field with co-rotating iogenic plasma and the Galilean moons. The disturbances created near the moons propagate as Alfvén waves along the magnetic field lines. The position of the moons is therefore "Alfvénically" connected to their respective auroral footprint. The angular separation from the i… ▽ More Jupiter's satellite auroral footprints are a consequence of the interaction between the Jovian magnetic field with co-rotating iogenic plasma and the Galilean moons. The disturbances created near the moons propagate as Alfvén waves along the magnetic field lines. The position of the moons is therefore "Alfvénically" connected to their respective auroral footprint. The angular separation from the instantaneous magnetic footprint can be estimated by the so-called lead angle. That lead angle varies periodically as a function of orbital longitude, since the time for the Alfvén waves to reach the Jovian ionosphere varies accordingly. Using spectral images of the Main Alfvén Wing auroral spots collected by Juno-UVS during the first forty-three orbits, this work provides the first empirical model of the Io, Europa and Ganymede equatorial lead angles for the northern and southern hemispheres. Alfvén travel times between the three innermost Galilean moons to Jupiter's northern and southern hemispheres are estimated from the lead angle measurements. We also demonstrate the accuracy of the map** from the Juno magnetic field reference model (JRM33) at the completion of the prime mission for M-shells extending to at least 15RJ . Finally, we shows how the added knowledge of the lead angle can improve the interpretation of the moon-induced decametric emissions. △ Less

Submitted 28 April, 2023; originally announced April 2023.

Comments: 20 pages, 8 figures, Accepted for publication in Journal of Geophysical Research: Space Physics on 20 April 2023

A high spatial and spectral resolution study of Jupiter's mid-infrared auroral emissions and their response to a solar wind compression

Authors: James A Sinclair, Thomas K Greathouse, Rohini S Giles, John Lacy, Julianne Moses, Vincent Hue, Denis Grodent, Bertrand Bonfond, Chihiro Tao, Thibault Cavalié, Emma K Dahl, Glenn S Orton, Leigh N Fletcher, Patrick G J Irwin

Abstract: We present mid-infrared spectroscopy of Jupiter's mid-to-high latitudes using Gemini-North/TEXES (Texas Echelon Cross Echelle Spectrograph) on March 17-19, 2017. These observations capture Jupiter's hydrocarbon auroral emissions before, during and after the arrival of a solar wind compression on March 18th, which highlights the coupling between the polar stratosphere and external space environment… ▽ More We present mid-infrared spectroscopy of Jupiter's mid-to-high latitudes using Gemini-North/TEXES (Texas Echelon Cross Echelle Spectrograph) on March 17-19, 2017. These observations capture Jupiter's hydrocarbon auroral emissions before, during and after the arrival of a solar wind compression on March 18th, which highlights the coupling between the polar stratosphere and external space environment. In comparing observations on March 17th and 19th, we observe a brightening of the CH$_4$, C$_2$H$_2$ and C$_2$H$_4$ emissions in regions spatially coincident with the northern, duskside main auroral emission (henceforth, MAE). In inverting the spectra to derive atmospheric information, we determine that the duskside brightening results from an upper stratospheric (p < 0.1 mbar/z > 200 km) heating (e.g. $ΔT$ = 9.1 $\pm$ 2.1 K at 9 $μ$bar at 67.5$^\circ$N, 162.5$^\circ$W) with negligible heating at deeper pressures. Our interpretation is that the arrival of the solar wind enhancement drove magnetospheric dynamics through compression and/or viscous interactions on the flank. These dynamics accelerated currents and/or generated higher Poynting fluxes, which ultimately warmed the atmosphere through Joule heating and ion-neutral collisions. Poleward of the southern MAE, temperature retrievals demonstrate that auroral-related heating penetrates as deep as the 10-mbar level, in contrast to poleward of the northern MAE, where heating is only observed as deep as $\sim$3 mbar. We suggest this results from the south having higher Pedersen conductivities, and therefore stronger currents and acceleration of the neutrals, as well as the poleward heating overlap** with the apex of Jupiter's circulation thereby inhibiting efficient horizontal mixing/advection. △ Less

Submitted 17 April, 2023; originally announced April 2023.

Visual based Tomato Size Measurement System for an Indoor Farming Environment

Authors: Andy Kweon, Vishnu Hu, Jong Yoon Lim, Trevor Gee, Edmond Liu, Henry Williams, Bruce A. MacDonald, Mahla Nejati, Inkyu Sa, Ho Seok Ahn

Abstract: As technology progresses, smart automated systems will serve an increasingly important role in the agricultural industry. Current existing vision systems for yield estimation face difficulties in occlusion and scalability as they utilize a camera system that is large and expensive, which are unsuitable for orchard environments. To overcome these problems, this paper presents a size measurement met… ▽ More As technology progresses, smart automated systems will serve an increasingly important role in the agricultural industry. Current existing vision systems for yield estimation face difficulties in occlusion and scalability as they utilize a camera system that is large and expensive, which are unsuitable for orchard environments. To overcome these problems, this paper presents a size measurement method combining a machine learning model and depth images captured from three low cost RGBD cameras to detect and measure the height and width of tomatoes. The performance of the presented system is evaluated on a lab environment with real tomato fruits and fake leaves to simulate occlusion in the real farm environment. To improve accuracy by addressing fruit occlusion, our three-camera system was able to achieve a height measurement accuracy of 0.9114 and a width accuracy of 0.9443. △ Less

Submitted 12 April, 2023; originally announced April 2023.

Comments: 10 Pages, 12 Figures

Ensemble prosody prediction for expressive speech synthesis

Authors: Tian Huey Teh, Vivian Hu, Devang S Ram Mohan, Zack Hodari, Christopher G. R. Wallis, Tomás Gomez Ibarrondo, Alexandra Torresquintero, James Leoni, Mark Gales, Simon King

Abstract: Generating expressive speech with rich and varied prosody continues to be a challenge for Text-to-Speech. Most efforts have focused on sophisticated neural architectures intended to better model the data distribution. Yet, in evaluations it is generally found that no single model is preferred for all input texts. This suggests an approach that has rarely been used before for Text-to-Speech: an ens… ▽ More Generating expressive speech with rich and varied prosody continues to be a challenge for Text-to-Speech. Most efforts have focused on sophisticated neural architectures intended to better model the data distribution. Yet, in evaluations it is generally found that no single model is preferred for all input texts. This suggests an approach that has rarely been used before for Text-to-Speech: an ensemble of models. We apply ensemble learning to prosody prediction. We construct simple ensembles of prosody predictors by varying either model architecture or model parameter values. To automatically select amongst the models in the ensemble when performing Text-to-Speech, we propose a novel, and computationally trivial, variance-based criterion. We demonstrate that even a small ensemble of prosody predictors yields useful diversity, which, combined with the proposed selection criterion, outperforms any individual model from the ensemble. △ Less

Submitted 3 April, 2023; originally announced April 2023.

Comments: ICASSP 2023

Model Checking Access Control Policies: A Case Study using Google Cloud IAM

Authors: Antonios Gouglidis, Anna Kagia, Vincent C. Hu

Abstract: Authoring access control policies is challenging and prone to misconfigurations. Access control policies must be conflict-free. Hence, administrators should identify discrepancies between policy specifications and their intended function to avoid violating security principles. This paper aims to demonstrate how to formally verify access control policies. Model checking is used to verify access con… ▽ More Authoring access control policies is challenging and prone to misconfigurations. Access control policies must be conflict-free. Hence, administrators should identify discrepancies between policy specifications and their intended function to avoid violating security principles. This paper aims to demonstrate how to formally verify access control policies. Model checking is used to verify access control properties against policies supported by an access control model. The authors consider Google's Cloud Identity and Access Management (IAM) as a case study and follow NIST's guidelines to verify access control policies automatically. Automated verification using model checking can serve as a valuable tool and assist administrators in assessing the correctness of access control policies. This enables checking violations against security principles and performing security assessments of policies for compliance purposes. The authors demonstrate how to define Google's IAM underlying role-based access control (RBAC) model, specify its supported policies, and formally verify a set of properties through three examples. △ Less

Submitted 29 March, 2023; originally announced March 2023.

Catalog of Ultraviolet Bright Stars (CUBS): Strategies for UV occultation measurements, planetary illumination modeling, and sky map analyses using hybrid IUE-Kurucz spectra

Authors: M. A. Velez, K. D. Retherford, V. Hue, J. A. Kammer, T. M. Becker, G. R. Gladstone, M. W. Davis, T. K. Greathouse, P. M. Molyneux, S. M. Brooks, U. Raut, M. H. Versteeg

Abstract: Ultraviolet spectroscopy is a powerful method to study planetary surface composition through reflectance measurements and atmospheric composition through stellar/solar occultations, transits of other planetary bodies, and direct imaging of airglow and auroral emissions. The next generation of ultraviolet spectrographs (UVS) on board ESA's JUICE (Jupiter Icy Moons Explorer) and NASA's Europa Clippe… ▽ More Ultraviolet spectroscopy is a powerful method to study planetary surface composition through reflectance measurements and atmospheric composition through stellar/solar occultations, transits of other planetary bodies, and direct imaging of airglow and auroral emissions. The next generation of ultraviolet spectrographs (UVS) on board ESA's JUICE (Jupiter Icy Moons Explorer) and NASA's Europa Clipper missions will perform such measurements of Jupiter and its moons in the early 2030's. This work presents a compilation of a detailed UV stellar catalog, named CUBS, of targets with high intensity in the 50-210 nm wavelength range with applications relevant to planetary spectroscopy. These applications include: 1) Planning and simulating occultations, including calibration measurements; 2) Modeling starlight illumination of dark, nightside planetary surfaces primarily lit by the sky; and 3) Studying the origin of diffuse galactic UV light as mapped by existing datasets from Juno-UVS and others. CUBS includes information drawn from resources such as the International Ultraviolet Explorer (IUE) catalog and SIMBAD. We have constructed model spectra at 0.1 nm resolution for almost 90,000 targets using Kurucz models and, when available, IUE spectra. CUBS also includes robust checks for agreement between the Kurucz models and the IUE data. We also present a tool for which our catalog can be used to identify the best candidates for stellar occultation observations, with applications for any UV instrument. We report on our methods for producing CUBS and discuss plans for its implementation during ongoing and upcoming planetary missions. △ Less

Submitted 7 March, 2023; originally announced March 2023.

Comments: 13 pages, 4 figures, submitted to JATIS

Enhanced C$_2$H$_2$ absorption within Jupiter's southern auroral oval from Juno UVS observations

Authors: Rohini S. Giles, Vincent Hue, Thomas K. Greathouse, G. Randall Gladstone, Joshua A. Kammer, Maarten H. Versteeg, Bertrand Bonfond, Denis G. Grodent, Jean-Claude Gérard, James A. Sinclair, Scott J. Bolton, Steven M. Levin

Abstract: Reflected sunlight observations from the Ultraviolet Spectrograph (UVS) on the Juno spacecraft were used to study the distribution of acetylene (C$_2$H$_2$) at Jupiter's south pole. We find that the shape of the C$_2$H$_2$ absorption feature varies significantly across the polar region, and this can be used to infer spatial variability in the C$_2$H$_2$ abundance. There is a localized region of en… ▽ More Reflected sunlight observations from the Ultraviolet Spectrograph (UVS) on the Juno spacecraft were used to study the distribution of acetylene (C$_2$H$_2$) at Jupiter's south pole. We find that the shape of the C$_2$H$_2$ absorption feature varies significantly across the polar region, and this can be used to infer spatial variability in the C$_2$H$_2$ abundance. There is a localized region of enhanced C$_2$H$_2$ absorption which coincides with the location of Jupiter's southern polar aurora; the C$_2$H$_2$ abundance poleward of the auroral oval is a factor of 3 higher than adjacent quiescent, non-auroral longitudes. This builds on previous infrared studies which found enhanced C$_2$H$_2$ abundances within the northern auroral oval. This suggests that Jupiter's upper-atmosphere chemistry is being strongly influenced by the influx of charged auroral particles and demonstrates the necessity of develo** ion-neutral photochemical models of Jupiter's polar regions. △ Less

Submitted 21 February, 2023; originally announced February 2023.

Comments: Accepted in JGR: Planets

Self-Guided Diffusion Models

Authors: Vincent Tao Hu, David W Zhang, Yuki M. Asano, Gertjan J. Burghouts, Cees G. M. Snoek

Abstract: Diffusion models have demonstrated remarkable progress in image generation quality, especially when guidance is used to control the generative process. However, guidance requires a large amount of image-annotation pairs for training and is thus dependent on their availability, correctness and unbiasedness. In this paper, we eliminate the need for such annotation by instead leveraging the flexibili… ▽ More Diffusion models have demonstrated remarkable progress in image generation quality, especially when guidance is used to control the generative process. However, guidance requires a large amount of image-annotation pairs for training and is thus dependent on their availability, correctness and unbiasedness. In this paper, we eliminate the need for such annotation by instead leveraging the flexibility of self-supervision signals to design a framework for self-guided diffusion models. By leveraging a feature extraction function and a self-annotation function, our method provides guidance signals at various image granularities: from the level of holistic images to object boxes and even segmentation masks. Our experiments on single-label and multi-label image datasets demonstrate that self-labeled guidance always outperforms diffusion models without guidance and may even surpass guidance based on ground-truth labels, especially on unbalanced data. When equipped with self-supervised box or mask proposals, our method further generates visually diverse yet semantically consistent images, without the need for any class, box, or segment label annotation. Self-guided diffusion is simple, flexible and expected to profit from deployment at scale. Source code will be at: https://taohu.me/sgdm/ △ Less

Submitted 27 November, 2023; v1 submitted 12 October, 2022; originally announced October 2022.

Comments: CVPR 2023

Principles of the Battery Data Genome

Authors: Logan Ward, Susan Babinec, Eric J. Dufek, David A. Howey, Venkatasubramanian Viswanathan, Muratahan Aykol, David A. C. Beck, Ben Blaiszik, Bor-Rong Chen, George Crabtree, Valerio de Angelis, Philipp Dechent, Matthieu Dubarry, Erica E. Eggleton, Donal P. Finegan, Ian Foster, Chirranjeevi Gopal, Patrick Herring, Victor W. Hu, Noah H. Paulson, Yuliya Preger, Dirk Uwe Sauer, Kandler Smith, Seth Snyder, Shashank Sripad , et al. (2 additional authors not shown)

Abstract: Electrochemical energy storage is central to modern society -- from consumer electronics to electrified transportation and the power grid. It is no longer just a convenience but a critical enabler of the transition to a resilient, low-carbon economy. The large pluralistic battery research and development community serving these needs has evolved into diverse specialties spanning materials discover… ▽ More Electrochemical energy storage is central to modern society -- from consumer electronics to electrified transportation and the power grid. It is no longer just a convenience but a critical enabler of the transition to a resilient, low-carbon economy. The large pluralistic battery research and development community serving these needs has evolved into diverse specialties spanning materials discovery, battery chemistry, design innovation, scale-up, manufacturing and deployment. Despite the maturity and the impact of battery science and technology, the data and software practices among these disparate groups are far behind the state-of-the-art in other fields (e.g. drug discovery), which have enjoyed significant increases in the rate of innovation. Incremental performance gains and lost research productivity, which are the consequences, retard innovation and societal progress. Examples span every field of battery research , from the slow and iterative nature of materials discovery, to the repeated and time-consuming performance testing of cells and the mitigation of degradation and failures. The fundamental issue is that modern data science methods require large amounts of data and the battery community lacks the requisite scalable, standardized data hubs required for immediate use of these approaches. Lack of uniform data practices is a central barrier to the scale problem. In this perspective we identify the data- and software-sharing gaps and propose the unifying principles and tools needed to build a robust community of data hubs, which provide flexible sharing formats to address diverse needs. The Battery Data Genome is offered as a data-centric initiative that will enable the transformative acceleration of battery science and technology, and will ultimately serve as a catalyst to revolutionize our approach to innovation. △ Less

Submitted 3 December, 2021; v1 submitted 14 September, 2021; originally announced September 2021.

Comments: corrected author list

Meridional variations of C$_2$H$_2$ in Jupiter's stratosphere from Juno UVS observations

Authors: Rohini S. Giles, Thomas K. Greathouse, Vincent Hue, G. Randall Gladstone, Henrik Melin, Leigh N. Fletcher, Patrick G. J. Irwin, Joshua A. Kammer, Maarten H. Versteeg, Bertrand Bonfond, Denis C. Grodent, Scott J. Bolton, Steven M. Levin

Abstract: The UVS instrument on the Juno mission records far-ultraviolet reflected sunlight from Jupiter. These spectra are sensitive to the abundances of chemical species in the upper atmosphere and to the distribution of the stratospheric haze layer. We combine observations from the first 30 perijoves of the mission in order to study the meridional distribution of acetylene (C$_2$H$_2$) in Jupiter's strat… ▽ More The UVS instrument on the Juno mission records far-ultraviolet reflected sunlight from Jupiter. These spectra are sensitive to the abundances of chemical species in the upper atmosphere and to the distribution of the stratospheric haze layer. We combine observations from the first 30 perijoves of the mission in order to study the meridional distribution of acetylene (C$_2$H$_2$) in Jupiter's stratosphere. We find that the abundance of C$_2$H$_2$ decreases towards the poles by a factor of 2-4, in agreement with previous analyses of mid-infrared spectra. This result is expected from insolation rates: near the equator, the UV solar flux is higher, allowing more C$_2$H$_2$ to be generated from the UV photolysis of CH$_4$. The decrease in abundance towards the poles suggests that horizontal mixing rates are not rapid enough to homogenize the latitudinal distribution. △ Less

Submitted 23 July, 2021; originally announced July 2021.

Comments: Accepted in JGR Planets. 19 pages, 3 figures

Map** the zonal winds of Jupiter's stratospheric equatorial oscillation

Authors: B. Benmahi, T. Cavalié, T. K. Greathouse, V. Hue, R. Giles, S. Guerlet, A. Spiga, R. Cosentino

Abstract: Since the 1950s, quasi-periodic oscillations have been studied in the terrestrial equatorial stratosphere. Other planets of the solar system present (or are expected to present) such oscillations, like the Jupiter Equatorial Oscillation(JEO) and the Saturn Semi-Annual Oscillation (SSAO). In Jupiter's stratosphere, the equatorial oscillation of its relative temperature structure about the equator,… ▽ More Since the 1950s, quasi-periodic oscillations have been studied in the terrestrial equatorial stratosphere. Other planets of the solar system present (or are expected to present) such oscillations, like the Jupiter Equatorial Oscillation(JEO) and the Saturn Semi-Annual Oscillation (SSAO). In Jupiter's stratosphere, the equatorial oscillation of its relative temperature structure about the equator, is characterized by a quasi-period of 4.4 years. The stratospheric wind field in Jupiter's equatorial zone has never been directly observed. In this paper, we aim at map** the absolute wind speeds in Jupiter's equatorial stratosphere to quantify vertical and horizontal wind and temperature shear. Assuming geostrophic equilibrium, we apply the thermal wind balance using nearly simultaneous stratospheric temperature measurements between 0.1 and 30 mbar performed with Gemini/TEXES and direct zonal wind measurements derived at 1 mbar from ALMA observations, all carried out between March 14th and 22nd, 2017. We are thus able to calculate self-consistently the zonal wind field in Jupiter's stratosphere where the JEO occurs. We obtain stratospheric map of the zonal wind speeds as a function of latitude and pressure about Jupiter's equator for the first time. The winds are vertically layered with successive eastward and westward jets. We find a 200 m/s westward jet at 4 mbar at the equator, with a typical longitudinal variability on the order of ~50 m/s. By extending our wind calculations to the upper troposphere, we find a wind structure qualitatively close to the wind observed using cloud-tracking techniques. Nearly simultaneous temperature and wind measurements, both in the stratosphere, are a powerful tool for future investigations of the JEO (and other planetary equatorial oscillations) and its temporal evolution. △ Less

Submitted 16 July, 2021; v1 submitted 13 July, 2021; originally announced July 2021.

Comments: the article was accepted on July 12, 2021

Journal ref: A&A 652, A125 (2021)

Self-supervised Video Representation Learning with Cross-Stream Prototypical Contrasting

Authors: Martine Toering, Ioannis Gatopoulos, Maarten Stol, Vincent Tao Hu

Abstract: Instance-level contrastive learning techniques, which rely on data augmentation and a contrastive loss function, have found great success in the domain of visual representation learning. They are not suitable for exploiting the rich dynamical structure of video however, as operations are done on many augmented instances. In this paper we propose "Video Cross-Stream Prototypical Contrasting", a nov… ▽ More Instance-level contrastive learning techniques, which rely on data augmentation and a contrastive loss function, have found great success in the domain of visual representation learning. They are not suitable for exploiting the rich dynamical structure of video however, as operations are done on many augmented instances. In this paper we propose "Video Cross-Stream Prototypical Contrasting", a novel method which predicts consistent prototype assignments from both RGB and optical flow views, operating on sets of samples. Specifically, we alternate the optimization process; while optimizing one of the streams, all views are mapped to one set of stream prototype vectors. Each of the assignments is predicted with all views except the one matching the prediction, pushing representations closer to their assigned prototypes. As a result, more efficient video embeddings with ingrained motion information are learned, without the explicit need for optical flow computation during inference. We obtain state-of-the-art results on nearest-neighbour video retrieval and action recognition, outperforming previous best by +3.2% on UCF101 using the S3D backbone (90.5% Top-1 acc), and by +7.2% on UCF101 and +15.1% on HMDB51 using the R(2+1)D backbone. △ Less

Submitted 20 October, 2021; v1 submitted 18 June, 2021; originally announced June 2021.

Comments: WACV2022

Ctrl-P: Temporal Control of Prosodic Variation for Speech Synthesis

Authors: Devang S Ram Mohan, Vivian Hu, Tian Huey Teh, Alexandra Torresquintero, Christopher G. R. Wallis, Marlene Staib, Lorenzo Foglianti, Jiameng Gao, Simon King

Abstract: Text does not fully specify the spoken form, so text-to-speech models must be able to learn from speech data that vary in ways not explained by the corresponding text. One way to reduce the amount of unexplained variation in training data is to provide acoustic information as an additional learning signal. When generating speech, modifying this acoustic information enables multiple distinct rendit… ▽ More Text does not fully specify the spoken form, so text-to-speech models must be able to learn from speech data that vary in ways not explained by the corresponding text. One way to reduce the amount of unexplained variation in training data is to provide acoustic information as an additional learning signal. When generating speech, modifying this acoustic information enables multiple distinct renditions of a text to be produced. Since much of the unexplained variation is in the prosody, we propose a model that generates speech explicitly conditioned on the three primary acoustic correlates of prosody: $F_{0}$, energy and duration. The model is flexible about how the values of these features are specified: they can be externally provided, or predicted from text, or predicted then subsequently modified. Compared to a model that employs a variational auto-encoder to learn unsupervised latent features, our model provides more interpretable, temporally-precise, and disentangled control. When automatically predicting the acoustic features from text, it generates speech that is more natural than that from a Tacotron 2 model with reference encoder. Subsequent human-in-the-loop modification of the predicted acoustic features can significantly further increase naturalness. △ Less

Submitted 15 June, 2021; originally announced June 2021.

Comments: To be published in Interspeech 2021. 5 pages, 4 figures

ADEPT: A Dataset for Evaluating Prosody Transfer

Authors: Alexandra Torresquintero, Tian Huey Teh, Christopher G. R. Wallis, Marlene Staib, Devang S Ram Mohan, Vivian Hu, Lorenzo Foglianti, Jiameng Gao, Simon King

Abstract: Text-to-speech is now able to achieve near-human naturalness and research focus has shifted to increasing expressivity. One popular method is to transfer the prosody from a reference speech sample. There have been considerable advances in using prosody transfer to generate more expressive speech, but the field lacks a clear definition of what successful prosody transfer means and a method for meas… ▽ More Text-to-speech is now able to achieve near-human naturalness and research focus has shifted to increasing expressivity. One popular method is to transfer the prosody from a reference speech sample. There have been considerable advances in using prosody transfer to generate more expressive speech, but the field lacks a clear definition of what successful prosody transfer means and a method for measuring it. We introduce a dataset of prosodically-varied reference natural speech samples for evaluating prosody transfer. The samples include global variations reflecting emotion and interpersonal attitude, and local variations reflecting topical emphasis, propositional attitude, syntactic phrasing and marked tonicity. The corpus only includes prosodic variations that listeners are able to distinguish with reasonable accuracy, and we report these figures as a benchmark against which text-to-speech prosody transfer can be compared. We conclude the paper with a demonstration of our proposed evaluation methodology, using the corpus to evaluate two text-to-speech models that perform prosody transfer. △ Less

Submitted 21 July, 2021; v1 submitted 15 June, 2021; originally announced June 2021.

Comments: 5 pages, 1 figure, accepted to Interspeech 2021

First direct measurement of auroral and equatorial jets in the stratosphere of Jupiter

Authors: T. Cavalié, B. Benmahi, V. Hue, R. Moreno, E. Lellouch, T. Fouchet, P. Hartogh, L. Rezac, T. K. Greathouse, G. R. Gladstone, J. A. Sinclair, M. Dobrijevic, F. Billebaud, C. Jarchow

Abstract: Context. The tropospheric wind pattern in Jupiter consists of alternating prograde and retrograde zonal jets with typical velocities of up to 100 m/s around the equator. At much higher altitudes, in the ionosphere, strong auroral jets have been discovered with velocities of 1-2 km/s. There is no such direct measurement in the stratosphere of the planet. Aims. In this paper, we bridge the altitude… ▽ More Context. The tropospheric wind pattern in Jupiter consists of alternating prograde and retrograde zonal jets with typical velocities of up to 100 m/s around the equator. At much higher altitudes, in the ionosphere, strong auroral jets have been discovered with velocities of 1-2 km/s. There is no such direct measurement in the stratosphere of the planet. Aims. In this paper, we bridge the altitude gap between these measurements by directly measuring the wind speeds in Jupiter's stratosphere. Methods. We use the Atacama Large Millimeter/submillimeter Array's very high spectral and angular resolution imaging of the stratosphere of Jupiter to retrieve the wind speeds as a function of latitude by fitting the Doppler shifts induced by the winds on the spectral lines. Results. We detect for the first time equatorial zonal jets that reside at 1 mbar, i.e. above the altitudes where Jupiter's Quasi-Quadrennial Oscillation occurs. Most noticeably, we find 300-400 m/s non-zonal winds at 0.1 mbar over the polar regions underneath the main auroral ovals. They are in counter-rotation and lie several hundreds of kilometers below the ionospheric auroral winds. We suspect them to be the lower tail of the ionospheric auroral winds. Conclusions. We detect directly and for the first time strong winds in Jupiter's stratosphere. They are zonal at low-to-mid latitudes and non-zonal at polar latitudes. The wind system found at polar latitudes may help increase the effciency of chemical complexification by confining the photochemical products in a region of large energetic electron precipitation. △ Less

Submitted 22 March, 2021; originally announced March 2021.

Comments: 7 pages and 3 figures (+ 5 pages and 5 figures for the appendix). Published in A&A 647, L8

Journal ref: Astronomy and Astrophysics, Volume 647, L8 (2021)

Detection of a bolide in Jupiter's atmosphere with Juno UVS

Authors: Rohini S. Giles, Thomas K. Greathouse, Joshua A. Kammer, G. Randall Gladstone, Bertrand Bonfond, Vincent Hue, Denis C. Grodent, Jean-Claude Gérard, Maarten H. Versteeg, Scott J. Bolton, John E. P. Connerney, Steven M. Levin

Abstract: The UVS instrument on the Juno mission recorded transient bright emission from a point source in Jupiter's atmosphere. The spectrum shows that the emission is consistent with a 9600-K blackbody located 225 km above the 1-bar level and the duration of the emission was between 17 ms and 150 s. These characteristics are consistent with a bolide in Jupiter's atmosphere. Based on the energy emitted, we… ▽ More The UVS instrument on the Juno mission recorded transient bright emission from a point source in Jupiter's atmosphere. The spectrum shows that the emission is consistent with a 9600-K blackbody located 225 km above the 1-bar level and the duration of the emission was between 17 ms and 150 s. These characteristics are consistent with a bolide in Jupiter's atmosphere. Based on the energy emitted, we estimate that the impactor had a mass of 250-5000 kg, which corresponds to a diameter of 1-4 m. By considering all observations made with Juno UVS over the first 27 perijoves of the mission, we estimate an impact flux rate of 24,000 per year for impactors with masses greater than 250-5000 kg. △ Less

Submitted 8 February, 2021; originally announced February 2021.

Comments: Accepted in GRL. 21 pages, 3 figures

Saturn's Seasonal Atmosphere at Northern Summer Solstice

Authors: L. N. Fletcher, L. Sromovsky, V. Hue, J. I. Moses, S. Guerlet, R. A. West, T. Koskinen

Abstract: The incredible longevity of Cassini's orbital mission at Saturn has provided the most comprehensive exploration of a seasonal giant planet to date. This review explores Saturn's changing global temperatures, composition, and aerosol properties between northern spring and summer solstice (2015-2017), extending our previous review of Cassini's remote sensing investigations (2004-14, Fletcher et al.,… ▽ More The incredible longevity of Cassini's orbital mission at Saturn has provided the most comprehensive exploration of a seasonal giant planet to date. This review explores Saturn's changing global temperatures, composition, and aerosol properties between northern spring and summer solstice (2015-2017), extending our previous review of Cassini's remote sensing investigations (2004-14, Fletcher et al., 2018) to the grand finale. The result is an unprecedented record of Saturn's climate that spans almost half a Saturnian year, which can be used to test the seasonal predictions of radiative climate models, neutral and ion photochemistry models, and atmospheric circulation models. Hemispheric asymmetries in tropospheric and stratospheric temperatures were observed to reverse from northern winter to northern summer; spatial distributions of hydrocarbons and para-hydrogen shifted in response to atmospheric dynamics (e.g., seasonally-reversing Hadley cells, polar stratospheric vortex formation, equatorial stratospheric oscillations, and inter-hemispheric transport); and upper tropospheric and stratospheric aerosols exhibited changes in optical thickness that modulated Saturn's visible colours (from blue hues to a golden appearance in the north near solstice), reflectivity, and near-infrared emission. Numerical simulations of radiative balance and photochemistry do a good job in reproducing the observed seasonal change and phase lags, but discrepancies between models and observations still persist, indicating a crucial role for atmospheric dynamics and the need to couple chemical and radiative schemes to the next generation of circulation models. With Cassini's demise, an extended study of Saturn's seasons, from northern summer to autumn, will require the capabilities of ground- and space-based observatories, as we eagerly await the next orbital explorer at Saturn. △ Less

Submitted 6 May, 2023; v1 submitted 16 December, 2020; originally announced December 2020.

Comments: 24 pages, 17 figures, accepted for publication by Cambridge University Press as part of a multi-volume work edited by Kevin Baines, Michael Flasar, Norbert Krupp, and Thomas Stallard, entitled "Cassini at Saturn: The Grand Finale." The copy of the Chapter, as displayed on this website, is a draft, pre-publication copy only

1D photochemical model of the ionosphere and the stratosphere of Neptune

Authors: M. Dobrijevic, J. C. Loison, V. Hue, T. Cavalié, K. M. Hickson

Abstract: Neptune remains a mysterious world that deserves further exploration and is a high-priority objective for a future planetary mission in order to better understand the formation and evolution of ice giant planets. We have developed a coupled ion-neutral 1D photochemical model of Neptune's atmosphere to study the origin and evolution of the hydrocarbons and the oxygen species. The up-to-date chemica… ▽ More Neptune remains a mysterious world that deserves further exploration and is a high-priority objective for a future planetary mission in order to better understand the formation and evolution of ice giant planets. We have developed a coupled ion-neutral 1D photochemical model of Neptune's atmosphere to study the origin and evolution of the hydrocarbons and the oxygen species. The up-to-date chemical scheme is derived from one used for Titan's atmosphere, which led to good agreements with the Cassini-CIRS observations for oxygen species and the main hydrocarbons. The main results we obtain are the following: The ion-neutral chemistry coupling produces aromatics (and benzene in particular) in the atmosphere of Neptune with relatively high abundances. Our model results are in good agreement with observations (taking model uncertainties into account). Two ionospheric peaks are present in the atmosphere located above the pressure level of 10$^{-5}$ mbar and around 10$^{-3}$ mbar. The influx of oxygen species in the upper atmosphere of Neptune has an effect on the concentration of many ions. We show that in situ exploration of Neptune's atmosphere would provide very interesting constraints for photochemical models concerning in particular the origin of oxygen species and the contribution of ion chemistry. A precise description of upper atmospheric chemistry is crucial for a better understanding of the internal composition and the formation processes of this planet. △ Less

Submitted 16 November, 2020; originally announced November 2020.

Comments: 40 pages, 1 table, 20 figures. Published in Icarus

Journal ref: Icarus 335 (2020), 113375

Possible Transient Luminous Events observed in Jupiter's upper atmosphere

Authors: Rohini S. Giles, Thomas K. Greathouse, Bertrand Bonfond, G. Randall Gladstone, Joshua A. Kammer, Vincent Hue, Denis C. Grodent, Jean-Claude Gérard, Maarten H. Versteeg, Michael H. Wong, Scott J. Bolton, John E. P. Connerney, Steven M. Levin

Abstract: 11 transient bright flashes were detected in Jupiter's atmosphere using the UVS instrument on the Juno spacecraft. These bright flashes are only observed in a single spin of the spacecraft and their brightness decays exponentially with time, with a duration of ~1.4 ms. The spectra are dominated by H2 Lyman band emission and based on the level of atmospheric absorption, we estimate a source altitud… ▽ More 11 transient bright flashes were detected in Jupiter's atmosphere using the UVS instrument on the Juno spacecraft. These bright flashes are only observed in a single spin of the spacecraft and their brightness decays exponentially with time, with a duration of ~1.4 ms. The spectra are dominated by H2 Lyman band emission and based on the level of atmospheric absorption, we estimate a source altitude of 260 km above the 1-bar level. Based on these characteristics, we suggest that these are observations of Transient Luminous Events (TLEs) in Jupiter's upper atmosphere. In particular, we suggest that these are elves, sprites or sprite halos, three types of TLEs that occur in the Earth's upper atmosphere in response to tropospheric lightning strikes. This is supported by visible light imaging, which shows cloud features typical of lightning source regions at the locations of several of the bright flashes. TLEs have previously only been observed on Earth, although theoretical and experimental work has predicted that they should also be present on Jupiter. △ Less

Submitted 26 October, 2020; originally announced October 2020.

Comments: Accepted in JGR: Planets. 28 pages, 8 figures

PointMixup: Augmentation for Point Clouds

Authors: Yunlu Chen, Vincent Tao Hu, Efstratios Gavves, Thomas Mensink, Pascal Mettes, Pengwan Yang, Cees G. M. Snoek

Abstract: This paper introduces data augmentation for point clouds by interpolation between examples. Data augmentation by interpolation has shown to be a simple and effective approach in the image domain. Such a mixup is however not directly transferable to point clouds, as we do not have a one-to-one correspondence between the points of two different objects. In this paper, we define data augmentation bet… ▽ More This paper introduces data augmentation for point clouds by interpolation between examples. Data augmentation by interpolation has shown to be a simple and effective approach in the image domain. Such a mixup is however not directly transferable to point clouds, as we do not have a one-to-one correspondence between the points of two different objects. In this paper, we define data augmentation between point clouds as a shortest path linear interpolation. To that end, we introduce PointMixup, an interpolation method that generates new examples through an optimal assignment of the path function between two point clouds. We prove that our PointMixup finds the shortest path between two point clouds and that the interpolation is assignment invariant and linear. With the definition of interpolation, PointMixup allows to introduce strong interpolation-based regularizers such as mixup and manifold mixup to the point cloud domain. Experimentally, we show the potential of PointMixup for point cloud classification, especially when examples are scarce, as well as increased robustness to noise and geometric transformations to points. The code for PointMixup and the experimental details are publicly available. △ Less

Submitted 14 August, 2020; originally announced August 2020.

Comments: Accepted as Spotlight presentation at European Conference on Computer Vision (ECCV), 2020

Localizing the Common Action Among a Few Videos

Authors: Pengwan Yang, Vincent Tao Hu, Pascal Mettes, Cees G. M. Snoek

Abstract: This paper strives to localize the temporal extent of an action in a long untrimmed video. Where existing work leverages many examples with their start, their ending, and/or the class of the action during training time, we propose few-shot common action localization. The start and end of an action in a long untrimmed video is determined based on just a hand-full of trimmed video examples containin… ▽ More This paper strives to localize the temporal extent of an action in a long untrimmed video. Where existing work leverages many examples with their start, their ending, and/or the class of the action during training time, we propose few-shot common action localization. The start and end of an action in a long untrimmed video is determined based on just a hand-full of trimmed video examples containing the same action, without knowing their common class label. To address this task, we introduce a new 3D convolutional network architecture able to align representations from the support videos with the relevant query video segments. The network contains: (\textit{i}) a mutual enhancement module to simultaneously complement the representation of the few trimmed support videos and the untrimmed query video; (\textit{ii}) a progressive alignment module that iteratively fuses the support videos into the query branch; and (\textit{iii}) a pairwise matching module to weigh the importance of different support videos. Evaluation of few-shot common action localization in untrimmed videos containing a single or multiple action instances demonstrates the effectiveness and general applicability of our proposal. △ Less

Submitted 25 August, 2020; v1 submitted 13 August, 2020; originally announced August 2020.

Comments: ECCV 2020

18-year long monitoring of the evolution of H2O vapor in the stratosphere of Jupiter with the Odin space telescope

Authors: B. Benmahi, T. Cavalié, M. Dobrijevic, N. Biver, K. Bermudez-Diaz, Aa. Sandqvist, E. Lellouch, R. Moreno, T. Fouchet, V. Hue, P. Hartogh, F. Billebaud, A. Lecacheux, Å. Hjalmarson, U. Frisk, M. Olberg, andThe Odin Team

Abstract: Comet Shoemaker-Levy 9 impacted Jupiter in July 1994, leaving its stratosphere with several new species, among them water vapor (H2O). With the aid of a photochemical model H2O can be used as a dynamical tracer in the jovian stratosphere. In this paper, we aim at constraining vertical eddy diffusion (Kzz) at the levels where H2O resides. We monitored the H2O disk-averaged emission at 556.936 GHz w… ▽ More Comet Shoemaker-Levy 9 impacted Jupiter in July 1994, leaving its stratosphere with several new species, among them water vapor (H2O). With the aid of a photochemical model H2O can be used as a dynamical tracer in the jovian stratosphere. In this paper, we aim at constraining vertical eddy diffusion (Kzz) at the levels where H2O resides. We monitored the H2O disk-averaged emission at 556.936 GHz with the Odin space telescope between 2002 and 2019, covering nearly two decades. We analyzed the data with a combination of 1D photochemical and radiative transfer models to constrain vertical eddy diffusion in the stratosphere of Jupiter. The Odin observations show us that the emission of H2O has an almost linear decrease of about 40% between 2002 and 2019.We can only reproduce our time series if we increase the magnitude of Kzz in the pressure range where H2O diffuses downward from 2002 to 2019, i.e. from ~0.2 mbar to ~5 mbar. However, this modified Kzz is incompatible with hydrocarbon observations. We find that, even if allowance is made for the initially large abundances of H2O and CO at the impact latitudes, the photochemical conversion of H2O to CO2 is not sufficient to explain the progressive decline of the H2O line emission, suggestive of additional loss mechanisms. The Kzz we derived from the Odin observations of H2O can only be viewed as an upper limit in the ~0.2 mbar to ~5 mbar pressure range. The incompatibility between the interpretations made from H2O and hydrocarbon observations probably results from 1D modeling limitations. Meridional variability of H2O, most probably at auroral latitudes, would need to be assessed and compared with that of hydrocarbons to quantify the role of auroral chemistry in the temporal evolution of the H2O abundance since the SL9 impacts. Modeling the temporal evolution of SL9 species with a 2D model would be the next natural step. △ Less

Submitted 10 July, 2020; originally announced July 2020.

Comments: 9 pages, 8 figures, accepted in Astronomy and Astrophysics

Journal ref: A&A 641, A140 (2020)

The deep composition of Uranus and Neptune from in situ exploration and thermochemical modeling

Authors: Thibault Cavalié, Olivia Venot, Yamila Miguel, Leigh N. Fletcher, Peter Wurz, Olivier Mousis, Roda Bounaceur, Vincent Hue, Jérémy Leconte, Michel Dobrijevic

Abstract: The distant ice giants of the Solar System, Uranus and Neptune, have only been visited by one space mission, Voyager 2. The current knowledge on their composition remains very limited despite some recent advances. A better characterization of their composition is however essential to constrain their formation and evolution, as a significant fraction of their mass is made of heavy elements, contrar… ▽ More The distant ice giants of the Solar System, Uranus and Neptune, have only been visited by one space mission, Voyager 2. The current knowledge on their composition remains very limited despite some recent advances. A better characterization of their composition is however essential to constrain their formation and evolution, as a significant fraction of their mass is made of heavy elements, contrary to the gas giants Jupiter and Saturn. An in situ probe like Galileo would provide us with invaluable direct ground-truth composition measurements. However, some of the condensibles will remain out of the grasp of a shallow probe. While additional constraints could be obtained from a complementary orbiter, thermochemistry and diffusion modeling can further help us to increase the science return of an in situ probe. △ Less

Submitted 29 April, 2020; originally announced April 2020.

Comments: Accepted for publication in Space Science Reviews

Herschel map of Saturn's stratospheric water, delivered by the plumes of Enceladus

Authors: T. Cavalié, V. Hue, P. Hartogh, R. Moreno, E. Lellouch, H. Feuchtgruber, C. Jarchow, T. Cassidy, L. N. Fletcher, F. Billebaud, M. Dobrijevic, L. Rezac, G. S. Orton, M. Rengel, T. Fouchet, S. Guerlet

Abstract: Context. The origin of water in the stratospheres of Giant Planets has been an outstanding question ever since its first detection by ISO some 20 years ago. Water can originate from interplanetary dust particles, icy rings and satellites and large comet impacts. Analysis of Herschel Space Observatory observations have proven that the bulk of Jupiter's stratospheric water was delivered by the Shoem… ▽ More Context. The origin of water in the stratospheres of Giant Planets has been an outstanding question ever since its first detection by ISO some 20 years ago. Water can originate from interplanetary dust particles, icy rings and satellites and large comet impacts. Analysis of Herschel Space Observatory observations have proven that the bulk of Jupiter's stratospheric water was delivered by the Shoemaker-Levy 9 impacts in 1994. In 2006, the Cassini mission detected water plumes at the South Pole of Enceladus, placing the moon as a serious candidate for Saturn's stratospheric water. Further evidence was found in 2011, when Herschel demonstrated the presence of a water torus at the orbital distance of Enceladus, fed by the moon's plumes. Finally, water falling from the rings onto Saturn's uppermost atmospheric layers at low latitudes was detected during the final orbits of Cassini's end-of-mission plunge into the atmosphere. Aims. In this paper, we use Herschel map** observations of water in Saturn's stratosphere to identify its source. Methods. Several empirical models are tested against the Herschel-HIFI and -PACS observations, which were collected on December 30, 2010, and January 2nd, 2011 (respectively). Results. We demonstrate that Saturn's stratospheric water is not uniformly mixed as a function of latitude, but peaking at the equator and decreasing poleward with a Gaussian distribution. We obtain our best fit with an equatorial mole fraction 1.1 ppb and a half-width at half-maximum of 25°, when accounting for a temperature increase in the two warm stratospheric vortices produced by Saturn's Great Storm of 2010-2011. Conclusions. This work demonstrates that Enceladus is the main source of Saturn's stratospheric water. △ Less

Submitted 20 August, 2019; originally announced August 2019.

Comments: Accepted for publication in A&A, section 10. Planets and planetary systems. 16 pages. 17 figures, 1 table

Journal ref: A&A 630, A87 (2019)

Seasonal stratospheric photochemistry on Uranus and Neptune

Authors: Julianne I. Moses, Leigh N. Fletcher, Thomas K. Greathouse, Glenn S. Orton, Vincent Hue

Abstract: A time-variable 1D photochemical model is used to study the distribution of stratospheric hydrocarbons as a function of altitude, latitude, and season on Uranus and Neptune. The results for Neptune indicate that in the absence of stratospheric circulation or other meridional transport processes, the hydrocarbon abundances exhibit strong seasonal and meridional variations in the upper stratosphere,… ▽ More A time-variable 1D photochemical model is used to study the distribution of stratospheric hydrocarbons as a function of altitude, latitude, and season on Uranus and Neptune. The results for Neptune indicate that in the absence of stratospheric circulation or other meridional transport processes, the hydrocarbon abundances exhibit strong seasonal and meridional variations in the upper stratosphere, but that these variations become increasingly damped with depth due to increasing dynamical and chemical time scales. At high altitudes, hydrocarbon mixing ratios are typically largest where the solar insolation is the greatest, leading to strong hemispheric dichotomies between the summer-to-fall hemisphere and winter-to-spring hemisphere. At mbar pressures and deeper, slower chemistry and diffusion lead to latitude variations that become more symmetric about the equator. On Uranus, the stagnant, poorly mixed stratosphere confines methane and its photochemical products to higher pressures, where chemistry and diffusion time scales remain large. Seasonal variations in hydrocarbons are therefore predicted to be more muted on Uranus, despite the planet's very large obliquity. Radiative-transfer simulations demonstrate that latitude variations in hydrocarbons on both planets are potentially observable with future JWST mid-infrared spectral imaging. Our seasonal model predictions for Neptune compare well with retrieved C2H2 and C2H6 abundances from spatially resolved ground-based observations (no such observations currently exist for Uranus), suggesting that stratospheric circulation -- which was not included in these models -- may have little influence on the large-scale meridional hydrocarbon distributions on Neptune, unlike the situation on Jupiter and Saturn. △ Less

Submitted 27 March, 2018; originally announced March 2018.

Journal ref: Icarus 307, 124-145, 2018

Photochemistry, mixing and transport in Jupiter's stratosphere constrained by Cassini

Authors: Vincent Hue, Franck Hersant, Thibault Cavalié, Michel Dobrijevic, James Sinclair

Abstract: We aim at constraining the diffusive and advective transport processes in Jupiter's stratosphere. The IR spectrum recorded by Cassini/CIRS during the Jupiter flyby contains the fingerprints of several atmospheric compounds and allows probing its atmospheric composition. C2H2 and C2H6, the main compounds produced by methane photochemistry, were retrieved as a function of latitude at certain pressur… ▽ More We aim at constraining the diffusive and advective transport processes in Jupiter's stratosphere. The IR spectrum recorded by Cassini/CIRS during the Jupiter flyby contains the fingerprints of several atmospheric compounds and allows probing its atmospheric composition. C2H2 and C2H6, the main compounds produced by methane photochemistry, were retrieved as a function of latitude at certain pressure levels. CIRS observations suggest a different meridional distribution for these two species, difficult to reconcile with their photochemical histories, which are thought to be coupled to the methane photolysis. While the C2H2 abundance decreases with latitude, C2H6 becomes more abundant at high latitudes. A new 2D (latitude-altitude) seasonal photochemical model is developed to study whether the addition of stratospheric transport processes, such as meridional diffusion and advection, can explain the latitudinal behavior of C2H2 and C2H6. C2H2 observations are fairly well reproduced without meridional diffusion. Adding meridional diffusion to the model provides an improved agreement with the C2H6 observations by flattening its meridional distribution, but degrades the fit to the C2H2 distribution. Meridional diffusion alone cannot produce the observed increase with latitude of C2H6. When adding advective transport between 30 mbar and 0.01 mbar, with upwelling winds at the equator and downwelling winds at high latitudes, we can reproduce the C2H6 abundance increase with latitude. However, the fit to the C2H2 distribution is degraded. The strength of the advective winds needed to reproduce the C2H6 abundances is very sensitive to the value of the meridional eddy diffusion coefficient. The coupled fate of C2H2 and C2H6 suggests that an additional process is missing in the model. Ion-neutral chemistry was not accounted for in this work and might be a good candidate to solve this issue. △ Less

Submitted 23 February, 2018; originally announced February 2018.

Comments: Received 13 October 2017 Revised 12 January 2018 Accepted 13 February 2018

Journal ref: Icarus 307 (2018) 106-123

Scientific rationale for Uranus and Neptune in situ explorations

Authors: O. Mousis, D. H. Atkinson, T. Cavalié, L. N. Fletcher, M. J. Amato, S. Aslam, F. Ferri, J. -B. Renard, T. Spilker, E. Venkatapathy, P. Wurz, K. Aplin, A. Coustenis, M. Deleuil, M. Dobrijevic, T. Fouchet, T. Guillot, P. Hartogh, T. Hewagama, M. D. Hofstadter, V. Hue, R. Hueso, J. -P. Lebreton, E. Lellouch, J. Moses , et al. (31 additional authors not shown)

Abstract: The ice giants Uranus and Neptune are the least understood class of planets in our solar system but the most frequently observed type of exoplanets. Presumed to have a small rocky core, a deep interior comprising ~70% heavy elements surrounded by a more dilute outer envelope of H2 and He, Uranus and Neptune are fundamentally different from the better-explored gas giants Jupiter and Saturn. Because… ▽ More The ice giants Uranus and Neptune are the least understood class of planets in our solar system but the most frequently observed type of exoplanets. Presumed to have a small rocky core, a deep interior comprising ~70% heavy elements surrounded by a more dilute outer envelope of H2 and He, Uranus and Neptune are fundamentally different from the better-explored gas giants Jupiter and Saturn. Because of the lack of dedicated exploration missions, our knowledge of the composition and atmospheric processes of these distant worlds is primarily derived from remote sensing from Earth-based observatories and space telescopes. As a result, Uranus's and Neptune's physical and atmospheric properties remain poorly constrained and their roles in the evolution of the Solar System not well understood. Exploration of an ice giant system is therefore a high-priority science objective as these systems (including the magnetosphere, satellites, rings, atmosphere, and interior) challenge our understanding of planetary formation and evolution. Here we describe the main scientific goals to be addressed by a future in situ exploration of an ice giant. An atmospheric entry probe targeting the 10-bar level, about 5 scale heights beneath the tropopause, would yield insight into two broad themes: i) the formation history of the ice giants and, in a broader extent, that of the Solar System, and ii) the processes at play in planetary atmospheres. The probe would descend under parachute to measure composition, structure, and dynamics, with data returned to Earth using a Carrier Relay Spacecraft as a relay station. In addition, possible mission concepts and partnerships are presented, and a strawman ice-giant probe payload is described. An ice-giant atmospheric probe could represent a significant ESA contribution to a future NASA ice-giant flagship mission. △ Less

Submitted 1 August, 2017; originally announced August 2017.

Comments: Submitted to Planetary and Space Science

2D photochemical modeling of Saturn's stratosphere. Part II: Feedback between composition and temperature

Authors: V. Hue, T. K. Greathouse, T. Cavalié, M. Dobrijevic, F. Hersant

Abstract: Saturn's axial tilt produces seasons in a similar way as on Earth. Both the stratospheric temperature and composition are affected by this latitudinally varying insolation along the seasons. The thermal structure is controlled and regulated by the amount of hydrocarbons in the stratosphere, which act as absorbers and coolants from the UV to the far-IR spectral range, and this structure influences… ▽ More Saturn's axial tilt produces seasons in a similar way as on Earth. Both the stratospheric temperature and composition are affected by this latitudinally varying insolation along the seasons. The thermal structure is controlled and regulated by the amount of hydrocarbons in the stratosphere, which act as absorbers and coolants from the UV to the far-IR spectral range, and this structure influences the amount of hydrocarbons. We study here the feedback between the chemical composition and the thermal structure by coupling a latitudinal and seasonal photochemical model with a radiative seasonal model. Our results show that the seasonal temperature peak in the higher stratosphere, associated with the seasonal increase of insolation, is shifted earlier than the maximum insolation peak. This shift is increased with increasing latitudes and is caused by the low amount of stratospheric coolants in the spring season. At 80$^{\circ}$ in both hemispheres, the temperature peak at 1d-2mbar is seen to occur half a season earlier than was previously predicted by radiative seasonal models that assumed spatially and temporally uniform distribution of coolants. This shift progressively decreases with increasing pressure, up to around the 0.5mbar pressure level where it vanishes. However, the thermal field has a small feedback on the abundance distributions. This feedback modifies the predicted equator-to-pole temperature gradient. The meridional gradients of temperature at the mbar pressure levels are better reproduced when this feedback is accounted for. At lower pressure levels, the thermal structure seems to depart from pure radiative seasonal equilibrium as previously suggested by Guerlet et al. (2014). Although the agreement with the absolute value of the stratospheric temperature observed by Cassini is moderate, it is a mandatory step toward a fully coupled GCM-photochemical model. △ Less

Submitted 8 December, 2015; originally announced December 2015.

Comments: 23 pages, 8 figures, accepted for publication in Icarus

2D photochemical modeling of Saturn's stratosphere. Part I: Seasonal variation of atmospheric composition without meridional transport

Authors: Vincent Hue, Thibault Cavalié, Michel Dobrijevic, Franck Hersant, Thomas K. Greathouse

Abstract: Saturn's axial tilt of 26.7° produces seasons in a similar way as on Earth. Both the stratospheric temperature and composition are affected by this latitudinally varying insolation along Saturn's orbital path. A new time dependent 2D photochemical model is presented to study the seasonal evolution of Saturn's stratospheric composition. This study focuses on the impact of the seasonally variable th… ▽ More Saturn's axial tilt of 26.7° produces seasons in a similar way as on Earth. Both the stratospheric temperature and composition are affected by this latitudinally varying insolation along Saturn's orbital path. A new time dependent 2D photochemical model is presented to study the seasonal evolution of Saturn's stratospheric composition. This study focuses on the impact of the seasonally variable thermal field on the main stratospheric C2 hydrocarbon chemistry (C2H2 and C2H6) using a realistic radiative climate model. Meridional mixing and advective processes are implemented in the model but turned off in the present study for the sake of simplicity. The results are compared to a simple study case where a latitudinally and temporally steady thermal field is assumed. Our simulations suggest that, when the seasonally variable thermal field is accounted for, the downward diffusion of the seasonally produced hydrocarbons is faster due to the seasonal compression of the atmospheric column during winter. This effect increases with increasing latitudes which experience the most important thermal changes in the course of the seasons. The seasonal variability of C2H2 and C2H6 therefore persists at higher-pressure levels with a seasonally-variable thermal field. Cassini limb-observations of C2H2 and C2H6 (Guerlet et al., 2009) are reasonably well-reproduced from the equator to 40° in both hemispheres △ Less

Submitted 9 April, 2015; originally announced April 2015.

Comments: 63 pages, 26 figures, manuscript accepted for publication in Icarus