Showing 1–20 of 20 results for author: Steiner, D

PathAlign: A vision-language model for whole slide images in histopathology

Authors: Faruk Ahmed, Andrew Sellergren, Lin Yang, Shawn Xu, Boris Babenko, Abbi Ward, Niels Olson, Arash Mohtashamian, Yossi Matias, Greg S. Corrado, Quang Duong, Dale R. Webster, Shravya Shetty, Daniel Golden, Yun Liu, David F. Steiner, Ellery Wulczyn

Abstract: Microscopic interpretation of histopathology images underlies many important diagnostic and treatment decisions. While advances in vision-language modeling raise new opportunities for analysis of such images, the gigapixel-scale size of whole slide images (WSIs) introduces unique challenges. Additionally, pathology reports simultaneously highlight key findings from small regions while also aggrega… ▽ More Microscopic interpretation of histopathology images underlies many important diagnostic and treatment decisions. While advances in vision-language modeling raise new opportunities for analysis of such images, the gigapixel-scale size of whole slide images (WSIs) introduces unique challenges. Additionally, pathology reports simultaneously highlight key findings from small regions while also aggregating interpretation across multiple slides, often making it difficult to create robust image-text pairs. As such, pathology reports remain a largely untapped source of supervision in computational pathology, with most efforts relying on region-of-interest annotations or self-supervision at the patch-level. In this work, we develop a vision-language model based on the BLIP-2 framework using WSIs paired with curated text from pathology reports. This enables applications utilizing a shared image-text embedding space, such as text or image retrieval for finding cases of interest, as well as integration of the WSI encoder with a frozen large language model (LLM) for WSI-based generative text capabilities such as report generation or AI-in-the-loop interactions. We utilize a de-identified dataset of over 350,000 WSIs and diagnostic text pairs, spanning a wide range of diagnoses, procedure types, and tissue types. We present pathologist evaluation of text generation and text retrieval using WSI embeddings, as well as results for WSI classification and workflow prioritization (slide-level triaging). Model-generated text for WSIs was rated by pathologists as accurate, without clinically significant error or omission, for 78% of WSIs on average. This work demonstrates exciting potential capabilities for language-aligned WSI embeddings. △ Less

Submitted 27 June, 2024; originally announced June 2024.

Comments: 9 main pages and 19 pages of supplemental material; 3 main tables, 3 main figures and 11 supplemental tables, 7 supplemental figures

Time-dependent long-term hydrodynamic simulations of the inner protoplanetary disk III: The influence of photoevaporation

Authors: M. Cecil, L. Gehrig, D. Steiner

Abstract: The final stages of a protoplanetary disk are essential for our understanding of the formation and evolution of planets. Photoevaporation is an important mechanism that contributes to the dispersal of an accretion disk and has significant consequences for the disk's lifetime. However, the combined effects of photoevaporation and star-disk interaction have not been investigated in previous studies.… ▽ More The final stages of a protoplanetary disk are essential for our understanding of the formation and evolution of planets. Photoevaporation is an important mechanism that contributes to the dispersal of an accretion disk and has significant consequences for the disk's lifetime. However, the combined effects of photoevaporation and star-disk interaction have not been investigated in previous studies. We combined an implicit disk evolution model with a photoevaporative mass-loss profile. By including the innermost disk regions down to 0.01 AU, we could calculate the star-disk interaction, the stellar spin evolution, and the transition from an accreting disk to the propeller regime self-consistently. Starting from an early Class II star-disk system, we calculated the long-term evolution of the system until the disk becomes almost completely dissolved. Photoevaporation has a significant effect on disk structure and evolution. The radial extent of the dead zone decreases, and the number of episodic accretion events (outbursts) is reduced by high stellar X-ray luminosities. Reasonable accretion rates in combination with photoevaporative gaps are possible for a dead zone that is still massive enough to develop episodic accretion events. Furthermore, the stellar spin evolution during the Class II evolution is less affected by the star-disk interaction in the case of high X-ray luminosities. Our results suggest that the formation of planets, especially habitable planets, in the dead zone is strongly impaired in the case of strong X-ray luminosities. Additionally, the importance of the star-disk interaction during the Class II phase with respect to the stellar spin evolution is reduced. △ Less

Submitted 9 May, 2024; originally announced May 2024.

Comments: 15 pages, 10 figures, to be published in Astronomy & Astrophysics

Advancing Multimodal Medical Capabilities of Gemini

Authors: Lin Yang, Shawn Xu, Andrew Sellergren, Timo Kohlberger, Yuchen Zhou, Ira Ktena, Atilla Kiraly, Faruk Ahmed, Farhad Hormozdiari, Tiam Jaroensri, Eric Wang, Ellery Wulczyn, Fayaz Jamil, Theo Guidroz, Chuck Lau, Siyuan Qiao, Yun Liu, Akshay Goel, Kendall Park, Arnav Agharwal, Nick George, Yang Wang, Ryutaro Tanno, David G. T. Barrett, Wei-Hung Weng , et al. (22 additional authors not shown)

Abstract: Many clinical tasks require an understanding of specialized data, such as medical images and genomics, which is not typically found in general-purpose large multimodal models. Building upon Gemini's multimodal models, we develop several models within the new Med-Gemini family that inherit core capabilities of Gemini and are optimized for medical use via fine-tuning with 2D and 3D radiology, histop… ▽ More Many clinical tasks require an understanding of specialized data, such as medical images and genomics, which is not typically found in general-purpose large multimodal models. Building upon Gemini's multimodal models, we develop several models within the new Med-Gemini family that inherit core capabilities of Gemini and are optimized for medical use via fine-tuning with 2D and 3D radiology, histopathology, ophthalmology, dermatology and genomic data. Med-Gemini-2D sets a new standard for AI-based chest X-ray (CXR) report generation based on expert evaluation, exceeding previous best results across two separate datasets by an absolute margin of 1% and 12%, where 57% and 96% of AI reports on normal cases, and 43% and 65% on abnormal cases, are evaluated as "equivalent or better" than the original radiologists' reports. We demonstrate the first ever large multimodal model-based report generation for 3D computed tomography (CT) volumes using Med-Gemini-3D, with 53% of AI reports considered clinically acceptable, although additional research is needed to meet expert radiologist reporting quality. Beyond report generation, Med-Gemini-2D surpasses the previous best performance in CXR visual question answering (VQA) and performs well in CXR classification and radiology VQA, exceeding SoTA or baselines on 17 of 20 tasks. In histopathology, ophthalmology, and dermatology image classification, Med-Gemini-2D surpasses baselines across 18 out of 20 tasks and approaches task-specific model performance. Beyond imaging, Med-Gemini-Polygenic outperforms the standard linear polygenic risk score-based approach for disease risk prediction and generalizes to genetically correlated diseases for which it has never been trained. Although further development and evaluation are necessary in the safety-critical medical domain, our results highlight the potential of Med-Gemini across a wide range of medical tasks. △ Less

Submitted 6 May, 2024; originally announced May 2024.

Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

Authors: Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent, Zhufeng Pan, Shibo Wang, Soroosh Mariooryad, Yifan Ding, Xinyang Geng, Fred Alcober, Roy Frostig, Mark Omernick, Lexi Walker, Cosmin Paduraru, Christina Sorokin, Andrea Tacchetti, Colin Gaffney, Samira Daruki, Olcan Sercinoglu, Zach Gleicher, Juliette Love , et al. (1092 additional authors not shown)

Abstract: In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February… ▽ More In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content. △ Less

Submitted 14 June, 2024; v1 submitted 8 March, 2024; originally announced March 2024.

Gemini: A Family of Highly Capable Multimodal Models

Authors: Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M. Dai, Anja Hauth, Katie Millican, David Silver, Melvin Johnson, Ioannis Antonoglou, Julian Schrittwieser, Amelia Glaese, Jilin Chen, Emily Pitler, Timothy Lillicrap, Angeliki Lazaridou, Orhan Firat, James Molloy, Michael Isard, Paul R. Barham, Tom Hennigan, Benjamin Lee , et al. (1325 additional authors not shown)

Abstract: This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultr… ▽ More This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of the Gemini family in cross-modal reasoning and language understanding will enable a wide variety of use cases. We discuss our approach toward post-training and deploying Gemini models responsibly to users through services including Gemini, Gemini Advanced, Google AI Studio, and Cloud Vertex AI. △ Less

Submitted 17 June, 2024; v1 submitted 18 December, 2023; originally announced December 2023.

Domain-specific optimization and diverse evaluation of self-supervised models for histopathology

Authors: Jeremy Lai, Faruk Ahmed, Supriya Vijay, Tiam Jaroensri, Jessica Loo, Saurabh Vyawahare, Saloni Agarwal, Fayaz Jamil, Yossi Matias, Greg S. Corrado, Dale R. Webster, Jonathan Krause, Yun Liu, Po-Hsuan Cameron Chen, Ellery Wulczyn, David F. Steiner

Abstract: Task-specific deep learning models in histopathology offer promising opportunities for improving diagnosis, clinical research, and precision medicine. However, development of such models is often limited by availability of high-quality data. Foundation models in histopathology that learn general representations across a wide range of tissue types, diagnoses, and magnifications offer the potential… ▽ More Task-specific deep learning models in histopathology offer promising opportunities for improving diagnosis, clinical research, and precision medicine. However, development of such models is often limited by availability of high-quality data. Foundation models in histopathology that learn general representations across a wide range of tissue types, diagnoses, and magnifications offer the potential to reduce the data, compute, and technical expertise necessary to develop task-specific deep learning models with the required level of model performance. In this work, we describe the development and evaluation of foundation models for histopathology via self-supervised learning (SSL). We first establish a diverse set of benchmark tasks involving 17 unique tissue types and 12 unique cancer types and spanning different optimal magnifications and task types. Next, we use this benchmark to explore and evaluate histopathology-specific SSL methods followed by further evaluation on held out patch-level and weakly supervised tasks. We found that standard SSL methods thoughtfully applied to histopathology images are performant across our benchmark tasks and that domain-specific methodological improvements can further increase performance. Our findings reinforce the value of using domain-specific SSL methods in pathology, and establish a set of high quality foundation models to enable further research across diverse applications. △ Less

Submitted 19 October, 2023; originally announced October 2023.

Comments: 4 main tables, 3 main figures, additional supplemental tables and figures

Towards In-context Scene Understanding

Authors: Ivana Balažević, David Steiner, Nikhil Parthasarathy, Relja Arandjelović, Olivier J. Hénaff

Abstract: In-context learning$\unicode{x2013}$the ability to configure a model's behavior with different prompts$\unicode{x2013}$has revolutionized the field of natural language processing, alleviating the need for task-specific models and paving the way for generalist models capable of assisting with any query. Computer vision, in contrast, has largely stayed in the former regime: specialized decoders and… ▽ More In-context learning$\unicode{x2013}$the ability to configure a model's behavior with different prompts$\unicode{x2013}$has revolutionized the field of natural language processing, alleviating the need for task-specific models and paving the way for generalist models capable of assisting with any query. Computer vision, in contrast, has largely stayed in the former regime: specialized decoders and finetuning protocols are generally required to perform dense tasks such as semantic segmentation and depth estimation. In this work we explore a simple mechanism for in-context learning of such scene understanding tasks: nearest neighbor retrieval from a prompt of annotated features. We propose a new pretraining protocol$\unicode{x2013}$leveraging attention within and across images$\unicode{x2013}$which yields representations particularly useful in this regime. The resulting Hummingbird model, suitably prompted, performs various scene understanding tasks without modification while approaching the performance of specialists that have been finetuned for each task. Moreover, Hummingbird can be configured to perform new tasks much more efficiently than finetuned models, raising the possibility of scene understanding in the interactive assistant regime. △ Less

Submitted 31 October, 2023; v1 submitted 2 June, 2023; originally announced June 2023.

Time-dependent, long-term hydrodynamic simulations of the inner protoplanetary disk II: The importance of stellar rotation

Authors: Lukas Gehrig, Daniel Steiner, Eduard Vorobyov, Manuel Güdel

Abstract: The spin evolution of young protostars, surrounded by an accretion disk, still poses problems for observations and theoretical models. In recent studies, the importance of the magnetic star-disk interaction for stellar spin evolution has been elaborated. The accretion disk in these studies, however, is only represented by a simplified model and important features are not considered. We combined th… ▽ More The spin evolution of young protostars, surrounded by an accretion disk, still poses problems for observations and theoretical models. In recent studies, the importance of the magnetic star-disk interaction for stellar spin evolution has been elaborated. The accretion disk in these studies, however, is only represented by a simplified model and important features are not considered. We combined the implicit hydrodynamic TAPIR disk code with a stellar spin evolution model. The influence of stellar magnetic fields on the disk dynamics, the radial position of the inner disk radius, as well as the influence of stellar rotation on the disk were calculated self-consistently. Within a defined parameter space, we can reproduce the majority of fast and slow rotating stars observed in young stellar clusters. Additionally, the back reaction of different stellar spin evolutionary tracks on the disk can be analyzed. Disks around fast rotating stars are located closer to the star. Consequently, the disk midplane temperature in the innermost disk region increases significantly compared to slow rotating stars. We can show the effects of stellar rotation on episodic accretion outbursts. The higher temperatures of disks around fast rotating stars result in more outbursts and a longer outbursting period over the disk lifetime. The combination of a long-term hydrodynamic disk and a stellar spin evolution model allows the inclusion of previously unconsidered effects such as the back-reaction of stellar rotation on the long-term disk evolution and the occurrence of accretion outbursts. However, a wider parameter range has to be studied to further investigate these effects. Additionally, a possible interaction between our model and a more realistic stellar evolution code (e.g., the MESA code) can improve our understanding of the stellar spin evolution and its effects on the pre-main sequence star. △ Less

Submitted 18 August, 2022; originally announced August 2022.

Comments: 12 pages, 7 figures, accepted by Astronomy & Astrophysics

Journal ref: A&A 667, A46 (2022)

Graph Neural Networks for Image Classification and Reinforcement Learning using Graph representations

Authors: Naman Goyal, David Steiner

Abstract: In this paper, we will evaluate the performance of graph neural networks in two distinct domains: computer vision and reinforcement learning. In the computer vision section, we seek to learn whether a novel non-redundant representation for images as graphs can improve performance over trivial pixel to node map** on a graph-level prediction graph, specifically image classification. For the reinfo… ▽ More In this paper, we will evaluate the performance of graph neural networks in two distinct domains: computer vision and reinforcement learning. In the computer vision section, we seek to learn whether a novel non-redundant representation for images as graphs can improve performance over trivial pixel to node map** on a graph-level prediction graph, specifically image classification. For the reinforcement learning section, we seek to learn if explicitly modeling solving a Rubik's cube as a graph problem can improve performance over a standard model-free technique with no inductive bias. △ Less

Submitted 8 March, 2022; v1 submitted 7 March, 2022; originally announced March 2022.

Comments: The work was done as a project for Neural Networks and Deep Learning course, Fall 2021 offering by Prof. Richard Zemel at Columbia University

Time-dependent, long-term hydrodynamic simulations of the inner protoplanetary disk I: The importance of stellar magnetic torques

Authors: D. Steiner, L. Gehrig, B. W. Ratschiner, F. Ragossnig, E. I. Vorobyov, M. Güdel, E. A. Dorfi

Abstract: We conduct simulations of the inner regions of protoplanetary disks (PPDs) to investigate the effects of protostellar magnetic fields on their long-term evolution. We use an inner boundary model that incorporates the influence of a stellar magnetic field. The position of the inner disk is dependent on the mass accretion rate as well as the magnetic field strength. We use this model to study the re… ▽ More We conduct simulations of the inner regions of protoplanetary disks (PPDs) to investigate the effects of protostellar magnetic fields on their long-term evolution. We use an inner boundary model that incorporates the influence of a stellar magnetic field. The position of the inner disk is dependent on the mass accretion rate as well as the magnetic field strength. We use this model to study the response of a magnetically truncated inner disk to an episodic accretion event. Additionally, we vary the protostellar magnetic field strength and investigate the consequences of the magnetic field on the long-term behavior of PPDs. We use the fully implicit 1+1D TAPIR code which solves the axisymmetric hydrodynamic equations self-consistently. Our model allows us to investigate disk dynamics close to the star and to conduct long-term evolution simulations simultaneously and includes the radial radiation transport in the stationary diffusion limit. We include stellar magnetic torques, the influence of a pressure gradient, and a variable inner disk radius in the TAPIR code to describe the innermost disk region in a more self-consistent manner and can show that this approach alters the disk dynamics considerably compared to a simplified diffusive evolution equation, especially during outbursts. The influences of a prescribed stellar magnetic field, local pressure gradients, and a variable inner disk radius result in a more consistent description of the gas dynamics in the innermost regions of PPDs. Combining magnetic torques acting on the innermost disk regions with the long-term evolution of PPDs yields previously unseen results, whereby the whole disk structure is affected over its entire lifetime. Additionally, we want to emphasize that a combination of our 1+1D model with more sophisticated multi-dimensional codes could improve the understanding of PPDs even further. △ Less

Submitted 18 September, 2021; originally announced September 2021.

Journal ref: A&A 655, A110 (2021)

Predicting Prostate Cancer-Specific Mortality with A.I.-based Gleason Grading

Authors: Ellery Wulczyn, Kunal Nagpal, Matthew Symonds, Melissa Moran, Markus Plass, Robert Reihs, Farah Nader, Fraser Tan, Yuannan Cai, Trissia Brown, Isabelle Flament-Auvigne, Mahul B. Amin, Martin C. Stumpe, Heimo Muller, Peter Regitnig, Andreas Holzinger, Greg S. Corrado, Lily H. Peng, Po-Hsuan Cameron Chen, David F. Steiner, Kurt Zatloukal, Yun Liu, Craig H. Mermel

Abstract: Gleason grading of prostate cancer is an important prognostic factor but suffers from poor reproducibility, particularly among non-subspecialist pathologists. Although artificial intelligence (A.I.) tools have demonstrated Gleason grading on-par with expert pathologists, it remains an open question whether A.I. grading translates to better prognostication. In this study, we developed a system to p… ▽ More Gleason grading of prostate cancer is an important prognostic factor but suffers from poor reproducibility, particularly among non-subspecialist pathologists. Although artificial intelligence (A.I.) tools have demonstrated Gleason grading on-par with expert pathologists, it remains an open question whether A.I. grading translates to better prognostication. In this study, we developed a system to predict prostate-cancer specific mortality via A.I.-based Gleason grading and subsequently evaluated its ability to risk-stratify patients on an independent retrospective cohort of 2,807 prostatectomy cases from a single European center with 5-25 years of follow-up (median: 13, interquartile range 9-17). The A.I.'s risk scores produced a C-index of 0.84 (95%CI 0.80-0.87) for prostate cancer-specific mortality. Upon discretizing these risk scores into risk groups analogous to pathologist Grade Groups (GG), the A.I. had a C-index of 0.82 (95%CI 0.78-0.85). On the subset of cases with a GG in the original pathology report (n=1,517), the A.I.'s C-indices were 0.87 and 0.85 for continuous and discrete grading, respectively, compared to 0.79 (95%CI 0.71-0.86) for GG obtained from the reports. These represent improvements of 0.08 (95%CI 0.01-0.15) and 0.07 (95%CI 0.00-0.14) respectively. Our results suggest that A.I.-based Gleason grading can lead to effective risk-stratification and warrants further evaluation for improving disease management. △ Less

Submitted 24 November, 2020; originally announced December 2020.

Journal ref: Nature Communications Medicine (2021)

Interpretable Survival Prediction for Colorectal Cancer using Deep Learning

Authors: Ellery Wulczyn, David F. Steiner, Melissa Moran, Markus Plass, Robert Reihs, Fraser Tan, Isabelle Flament-Auvigne, Trissia Brown, Peter Regitnig, Po-Hsuan Cameron Chen, Narayan Hegde, Apaar Sadhwani, Robert MacDonald, Benny Ayalew, Greg S. Corrado, Lily H. Peng, Daniel Tse, Heimo Müller, Zhaoyang Xu, Yun Liu, Martin C. Stumpe, Kurt Zatloukal, Craig H. Mermel

Abstract: Deriving interpretable prognostic features from deep-learning-based prognostic histopathology models remains a challenge. In this study, we developed a deep learning system (DLS) for predicting disease specific survival for stage II and III colorectal cancer using 3,652 cases (27,300 slides). When evaluated on two validation datasets containing 1,239 cases (9,340 slides) and 738 cases (7,140 slide… ▽ More Deriving interpretable prognostic features from deep-learning-based prognostic histopathology models remains a challenge. In this study, we developed a deep learning system (DLS) for predicting disease specific survival for stage II and III colorectal cancer using 3,652 cases (27,300 slides). When evaluated on two validation datasets containing 1,239 cases (9,340 slides) and 738 cases (7,140 slides) respectively, the DLS achieved a 5-year disease-specific survival AUC of 0.70 (95%CI 0.66-0.73) and 0.69 (95%CI 0.64-0.72), and added significant predictive value to a set of 9 clinicopathologic features. To interpret the DLS, we explored the ability of different human-interpretable features to explain the variance in DLS scores. We observed that clinicopathologic features such as T-category, N-category, and grade explained a small fraction of the variance in DLS scores (R2=18% in both validation sets). Next, we generated human-interpretable histologic features by clustering embeddings from a deep-learning based image-similarity model and showed that they explain the majority of the variance (R2 of 73% to 80%). Furthermore, the clustering-derived feature most strongly associated with high DLS scores was also highly prognostic in isolation. With a distinct visual appearance (poorly differentiated tumor cell clusters adjacent to adipose tissue), this feature was identified by annotators with 87.0-95.5% accuracy. Our approach can be used to explain predictions from a prognostic deep learning model and uncover potentially-novel prognostic features that can be reliably identified by people for future validation studies. △ Less

Submitted 17 November, 2020; originally announced November 2020.

Journal ref: Nature Partner Journal Digital Medicine (2021)

1+1D implicit disk computations

Authors: Florian Ragossnig, Ernst A. Dorfi, Bernhard Ratschiner, Lukas Gehrig, Daniel Steiner, Alexander Stökl, Colin P. Johnstone

Abstract: We present an implicit numerical method to solve the time-dependent equations of radiation hydrodynamics (RHD) in axial symmetry assuming hydrostatic equilibrium perpendicular to the equatorial plane (1+1D) of a gaseous disk. The equations are formulated in conservative form on an adaptive grid and the corresponding fluxes are calculated by a spacial second order advection scheme. Self-gravity of… ▽ More We present an implicit numerical method to solve the time-dependent equations of radiation hydrodynamics (RHD) in axial symmetry assuming hydrostatic equilibrium perpendicular to the equatorial plane (1+1D) of a gaseous disk. The equations are formulated in conservative form on an adaptive grid and the corresponding fluxes are calculated by a spacial second order advection scheme. Self-gravity of the disk is included by solving the Possion equation. We test the resulting numerical method through comparison with a simplified analytical solution as well as through the long term viscous evolution of protoplanetary disk when due to viscosity matter is transported towards the central host star and the disk depletes. The importance of the inner boundary conditions on the structural behaviour of disks is demonstrated with several examples. △ Less

Submitted 23 June, 2020; originally announced June 2020.

Comments: for details check https://authors.elsevier.com/tracking/article/details.do?aid=107437&jid=COMPHY&surname=Ragossnig

Mass limits for stationary protoplanetary accretion disks

Authors: Florian Ragossnig, Lukas Gehrig, Ernst A. Dorfi, Daniel Steiner, Alexander Stökl

Abstract: The collapse of interstellar gaseous clouds towards a protostar leads to the formation of accretion disks around the central star. Such disks can be dynamically stable if they settle in an axisymmetric state. In this letter, we investigate the long-term stability of astrophysical viscous disks around various protostars. We apply an implicit numerical code which solves the equations of radiation hy… ▽ More The collapse of interstellar gaseous clouds towards a protostar leads to the formation of accretion disks around the central star. Such disks can be dynamically stable if they settle in an axisymmetric state. In this letter, we investigate the long-term stability of astrophysical viscous disks around various protostars. We apply an implicit numerical code which solves the equations of radiation hydrodynamics and treats turbulence-induced viscosity according to the $α$-viscosity model. We show how the viscosity is related to the disk mass. A stability criterion to determine the maximum disk mass can be formulated. We analyse such instabilities for a variety of radial points with different orbital distances from the host star and discuss the feedback on the disk in the event of an unstable protoplanetary disk. Additionally, we examine the critical disk-mass for disks with variable outer boundaries and compare them to observations of protostellar disks in the Upper Scorpius OB Association and near the Lupus complex. We derive an easily applicable method to obtain an estimate for maximum disk masses when the outer disk radius ins known. △ Less

Submitted 4 September, 2020; v1 submitted 10 January, 2020; originally announced January 2020.

Comments: Not yet published

Deep learning-based survival prediction for multiple cancer types using histopathology images

Authors: Ellery Wulczyn, David F. Steiner, Zhaoyang Xu, Apaar Sadhwani, Hongwu Wang, Isabelle Flament, Craig H. Mermel, Po-Hsuan Cameron Chen, Yun Liu, Martin C. Stumpe

Abstract: Prognostic information at diagnosis has important implications for cancer treatment and monitoring. Although cancer staging, histopathological assessment, molecular features, and clinical variables can provide useful prognostic insights, improving risk stratification remains an active research area. We developed a deep learning system (DLS) to predict disease specific survival across 10 cancer typ… ▽ More Prognostic information at diagnosis has important implications for cancer treatment and monitoring. Although cancer staging, histopathological assessment, molecular features, and clinical variables can provide useful prognostic insights, improving risk stratification remains an active research area. We developed a deep learning system (DLS) to predict disease specific survival across 10 cancer types from The Cancer Genome Atlas (TCGA). We used a weakly-supervised approach without pixel-level annotations, and tested three different survival loss functions. The DLS was developed using 9,086 slides from 3,664 cases and evaluated using 3,009 slides from 1,216 cases. In multivariable Cox regression analysis of the combined cohort including all 10 cancers, the DLS was significantly associated with disease specific survival (hazard ratio of 1.58, 95% CI 1.28-1.70, p<0.0001) after adjusting for cancer type, stage, age, and sex. In a per-cancer adjusted subanalysis, the DLS remained a significant predictor of survival in 5 of 10 cancer types. Compared to a baseline model including stage, age, and sex, the c-index of the model demonstrated an absolute 3.7% improvement (95% CI 1.0-6.5) in the combined cohort. Additionally, our models stratified patients within individual cancer stages, particularly stage II (p=0.025) and stage III (p<0.001). By develo** and evaluating prognostic models across multiple cancer types, this work represents one of the most comprehensive studies exploring the direct prediction of clinical outcomes using deep learning and histopathology images. Our analysis demonstrates the potential for this approach to provide prognostic information in multiple cancer types, and even within specific pathologic stages. However, given the relatively small number of clinical events, we observed wide confidence intervals, suggesting that future work will benefit from larger datasets. △ Less

Submitted 16 December, 2019; originally announced December 2019.

Journal ref: PLOS ONE (2020)

Time-dependent galactic winds

Authors: E. A. Dorfi, D. Steiner, F. Ragossnig, D. Breitschwerdt

Abstract: Cosmic rays (CRs) are transported out of the galaxy by diffusion and advection due to streaming along magnetic field lines and resonant scattering off self-excited Magneto-Hydro-Dynamic (MHD) waves. Thus momentum is transferred to the plasma via the frozen-in waves as a mediator assisting the thermal pressure in driving a galactic wind. Galactic CRs (GCRs) are accelerated by shock waves generated… ▽ More Cosmic rays (CRs) are transported out of the galaxy by diffusion and advection due to streaming along magnetic field lines and resonant scattering off self-excited Magneto-Hydro-Dynamic (MHD) waves. Thus momentum is transferred to the plasma via the frozen-in waves as a mediator assisting the thermal pressure in driving a galactic wind. Galactic CRs (GCRs) are accelerated by shock waves generated in supernova remnants (SNRs), and they propagate from the disc into the halo. Therefore CR acceleration in the halo strongly depends on the inner disc boundary conditions. We performed hydrodynamical simulations of galactic winds in flux tube geometry appropriate for disc galaxies, describing the CR diffusive-advective transport in a hydrodynamical fashion (by taking appropriate moments of the Fokker-Planck equation) along with the energy exchange with self-generated MHD waves. Our time-dependent CR hydrodynamic simulations confirm that the evolution of galactic winds with feedback depends on the structure of the galactic halo. In case of a wind-structured halo, the wind breaks down after the last super nova (SN) has exploded. The mechanism described here offers a natural and elegant solution to explain the power-law distribution of CRs between the `knee' and the `ankle'. The transition will be naturally smooth, because the Galactic CRs accelerated at SN shocks will be `post-accelerated' by shocks generated at the inner boundary and travelling through the halo. △ Less

Submitted 3 September, 2019; originally announced September 2019.

Comments: Galaxies: evolution -- ISM: jets and outflows -- Galaxies: starburst -- supernova remnants -- cosmic rays

Journal ref: A&A 630, A107 (2019)

Interaction of infalling solid bodies with primordial atmospheres of disk-embedded planets

Authors: Florian Ragossnig, Alexander Stökl, Ernst Dorfi, Colin P. Johnstone, Daniel Steiner, Manuel Güdel

Abstract: Planets that form early enough to be embedded in the circumstellar gas disk accumulate thick atmospheres of nebular gas. Models of these atmospheres need to specify the surface luminosity (i.e. energy loss rate) of the planet. This luminosity is usually associated with a continuous inflow of solid bodies, where the gravitational energy released from these bodies is the source of energy. However, i… ▽ More Planets that form early enough to be embedded in the circumstellar gas disk accumulate thick atmospheres of nebular gas. Models of these atmospheres need to specify the surface luminosity (i.e. energy loss rate) of the planet. This luminosity is usually associated with a continuous inflow of solid bodies, where the gravitational energy released from these bodies is the source of energy. However, if these bodies release energy in the atmosphere instead of at the surface, this assumption might not be justified. Our aim is to explore the interactions of infalling planetesimals with primordial atmospheres at an embedded phase of evolution. We investigate effects of atmospheric interaction on the planetesimals (mass loss) and the atmosphere (heating/cooling). We used atmospheric parameters from a snapshot of time-dependent evolution simulations for embedded atmospheres and simulated purely radial, infall events of siliceous planetesimals in a 1D, explicit code. We implemented energy transfer between friction, radiation transfer by the atmosphere and the body and thermal ablation; this gives us the possibility to examine the effects on the planetesimals and the atmosphere. We find that a significant amount of gravitational energy is indeed dissipated into the atmosphere, especially for larger planetary cores, which consequently cannot contribute to the atmospheric planetary luminosity. Furthermore, we examine that planetesimal infall events for cores, $M_\mathrm{C} > 2$M$_{\oplus}$, which actually result in a local cooling of the atmosphere; this is totally in contradiction with the classical model. △ Less

Submitted 14 September, 2018; originally announced September 2018.

Journal ref: A&A 618, A19 (2018)

Equalizer 2.0 - Convergence of a Parallel Rendering Framework

Authors: Stefan Eilemann, David Steiner, Renato Pajarola

Abstract: Develo** complex, real world graphics applications which leverage multiple GPUs and computers for interactive 3D rendering tasks is a complex task. It requires expertise in distributed systems and parallel rendering in addition to the application domain itself. We present a mature parallel rendering framework which provides a large set of features, algorithms and system integration for a wide ra… ▽ More Develo** complex, real world graphics applications which leverage multiple GPUs and computers for interactive 3D rendering tasks is a complex task. It requires expertise in distributed systems and parallel rendering in addition to the application domain itself. We present a mature parallel rendering framework which provides a large set of features, algorithms and system integration for a wide range of real-world research and industry applications. Using the Equalizer parallel rendering framework, we show how a wide set of generic algorithms can be integrated in the framework to help application scalability and development in many different domains, highlighting how concrete applications benefit from the diverse aspects and use cases of Equalizer. We present novel parallel rendering algorithms, powerful abstractions for large visualization setups and virtual reality, as well as new experimental results for parallel rendering and data distribution. △ Less

Submitted 22 February, 2018; originally announced February 2018.

ACM Class: I.3.2.a; I.3.7.g; I.3.8; I.3.6.a; I.6.9.f

Genetic Complexity in a Drosophila Model of Diabetes-Associated Misfolded Human Proinsulin

Authors: Soo-Young Park, Michael Z. Ludwig, Natalia A. Tamarina, Bin Z. He, Sarah H. Carl, Desiree A. Dickerson, Levi Barse, Bharath Arun, Calvin Williams, Cecelia M. Miles, Louis H. Philipson, Donald F. Steiner, Graeme I. Bell, Martin Kreitman

Abstract: Here we use Drosophila melanogaster to create a genetic model of human permanent neonatal diabetes mellitus and present experimental results describing dimensions of this complexity. The approach involves the transgenic expression of a misfolded mutant of human preproinsulin, hINSC96Y, which is a cause of the disease. When expressed in fly imaginal discs, hINSC96Y causes a reduction of adult struc… ▽ More Here we use Drosophila melanogaster to create a genetic model of human permanent neonatal diabetes mellitus and present experimental results describing dimensions of this complexity. The approach involves the transgenic expression of a misfolded mutant of human preproinsulin, hINSC96Y, which is a cause of the disease. When expressed in fly imaginal discs, hINSC96Y causes a reduction of adult structures, including the eye, wing and notum. Eye imaginal discs exhibit defects in both the structure and arrangement of ommatidia. In the wing, expression of hINSC96Y leads to ectopic expression of veins and mechano-sensory organs, indicating disruption of wild type signaling processes regulating cell fates. These readily measurable disease phenotypes are sensitive to temperature, gene dose and sex. Mutant (but not wild type) proinsulin expression in the eye imaginal disc induces IRE1-mediated Xbp1 alternative splicing, a signal for endoplasmic reticulum stress response activation, and produces global change in gene expression. Mutant hINS transgene tester strains, when crossed to stocks from the Drosophila Genetic Reference Panel produces F1 adults with a continuous range of disease phenotypes and large broad-sense heritability. Surprisingly, the severity of mutant hINS-induced disease in the eye is not correlated with that in the notum in these crosses, nor with eye reduction phenotypes caused by the expression of two dominant eye mutants acting in two different eye development pathways, Drop (Dr) or Lobe (L) when crossed into the same genetic backgrounds. The tissue specificity of genetic variability for mutant hINS-induced disease thus has its own distinct signature. The genetic dominance of disease-specific phenotypic variability makes this approach amenable to genome-wide association study (GWAS) in a simple F1 screen of natural variation. △ Less

Submitted 31 May, 2013; originally announced June 2013.

Comments: 60 pages; 6 figures; 8 supporting figures; 11 supporting tables

Network Bargaining: Using Approximate Blocking Sets to Stabilize Unstable Instances

Authors: Jochen Koenemann, Kate Larson, David Steiner

Abstract: We study a network extension to the Nash bargaining game, as introduced by Kleinberg and Tardos (STOC'08), where the set of players corresponds to vertices in a graph $G=(V,E)$ and each edge $ij\in E$ represents a possible deal between players $i$ and $j$. We reformulate the problem as a cooperative game and study the following question: Given a game with an empty core (i.e. an unstable game) is i… ▽ More We study a network extension to the Nash bargaining game, as introduced by Kleinberg and Tardos (STOC'08), where the set of players corresponds to vertices in a graph $G=(V,E)$ and each edge $ij\in E$ represents a possible deal between players $i$ and $j$. We reformulate the problem as a cooperative game and study the following question: Given a game with an empty core (i.e. an unstable game) is it possible, through minimal changes in the underlying network, to stabilize the game? We show that by removing edges in the network that belong to a blocking set we can find a stable solution in polynomial time. This motivates the problem of finding small blocking sets. While it has been previously shown that finding the smallest blocking set is NP-hard (Biro,Kern,Paulusma, TAMC'10), we show that it is possible to efficiently find approximate blocking sets in sparse graphs. △ Less

Submitted 28 July, 2012; originally announced July 2012.