Deep Learning in the Natural Sciences: Applications to Physics

A key question for machine learning approaches in particle physics is how to best represent and learn from collider events. As an event is intrinsically a variablelength unordered set of particles, we build upon recent machine learning efforts to learn directly from sets of features or "point clouds". Adapting and specializing the "Deep Sets" framework to particle physics, we introduce Energy Flow Networks, which respect infrared and collinear safety by construction. We also develop Particle Flow Networks, which allow for general energy dependence and the inclusion of additional particle-level information such as charge and flavor. These networks feature a per-particle internal (latent) representation, and summing over all particles yields an overall event-level latent representation. We show how this latent space decomposition unifies existing event representations based on detector images and radiation moments. To demonstrate the power and simplicity of this set-based approach, we apply these networks to the collider task of discriminating quark jets from gluon jets, finding similar or improved performance compared to existing methods. We also show how the learned event representation can be directly visualized, providing insight into the inner workings of the model. These architectures lend themselves to efficiently processing and analyzing events for a wide variety of tasks at the Large Hadron Collider. Implementations and examples of our architectures are available online in our EnergyFlow package.

show abstract

“…See Refs. [54][55][56][57][58] for more detailed reviews of machine learning in high-energy physics.…”

Section: Introductionmentioning

confidence: 99%

Energy flow networks: deep sets for particle jets

Komiske

Metodiev

Thaler

2019

J. High Energ. Phys.

266

284

View full text Add to dashboard Cite

show abstract

“…Each hidden layer uses the output of the preceding layer as its input. In data science, deep learning has emerged as a powerful technique for tackling previously intractable problems in the natural world (Sadowski and Baldi, 2018;Bourilkov, 2019). This is assisted by deep learning's enhanced capacity to find intricate patterns in extremely large datasets.…”

Section: Deep Neural Networkmentioning

confidence: 99%

Climatic and seismic data-driven deep learning model for earthquake magnitude prediction

et al. 2023

View full text Add to dashboard Cite

The effects of global warming are felt not only in the Earth’s climate but also in the geology of the planet. Modest variations in stress and pore-fluid pressure brought on by temperature variations, precipitation, air pressure, and snow coverage are hypothesized to influence seismicity on local and regional scales. Earthquakes can be anticipated by intelligently evaluating historical climatic datasets and earthquake catalogs that have been collected all over the world. This study attempts to predict the magnitude of the next probable earthquake by evaluating climate data along with eight mathematically calculated seismic parameters. Global temperature has been selected as the only climatic variable for this research, as it substantially affects the planet’s ecosystem and civilization. Three popular deep neural network models, namely, long short-term memory (LSTM), bidirectional long short-term memory (Bi-LSTM), and transformer models, were used to predict the magnitude of the next earthquakes in three seismic regions: Japan, Indonesia, and the Hindu-Kush Karakoram Himalayan (HKKH) region. Several well-known metrics, such as the mean absolute error (MAE), mean squared error (MSE), log-cosh loss, and mean squared logarithmic error (MSLE), have been used to analyse these models. All models eventually settle on a small value for these cost functions, demonstrating the accuracy of these models in predicting earthquake magnitudes. These approaches produce significant and encouraging results when used to predict earthquake magnitude at diverse places, opening the way for the ultimate robust prediction mechanism that has not yet been created.

show abstract

“…(De Cauwer et al, 2006), or the use of machine learning expertise to build predictive models based on such data e.g. (Sadowski & Baldi, 2018). Though there are differences in the epistemologies of the respective disciplines mentioned in these examples, these don't collide in the process of their epistemic augmentation in the pursuit of the research goal.…”

Section: Circumventionmentioning

confidence: 99%

Interdisciplinary Research as a Complicated System

Dalton

Wolff

Bekker

2022

International Journal of Qualitative Methods

View full text Add to dashboard Cite

Notwithstanding the rapidly growing body of literature on interdisciplinary research, several of the crucial dynamics in interdisciplinary research systems remain poorly understood and undertheorized. To this effect we seek to provide a ‘fundamental’ ontology of interdisciplinary research systems. We principally understand an interdisciplinary research system as a complex system consisting of researchers from different disciplines that have undergone a pseudomorphosis (i.e. a false formation) into a complicated system through the formation of a central organizing principle. The central organizing principle provides a stricter definition of the research problem and subsequently coalesces the intentionality of system agents through a unification of their disparate aims and methodologies. This pseudomorphosis is thereby associated with an exchange of individual freedom for organizational utility resulting in internal tensions which are, we argue, most prominently expressed in the interplay of epistemic incompatibilities between disciplines. We explore three frameworks for successfully navigating these incompatibilities: circumvention, which is based on avoidance of areas of disciplinary incompatibility; pragmatism, which is based on ignoring areas of compatibility; and disciplinary synthesis which involves a paradigm shift in researchers’ understanding of their disciplinary propositions resolving perceived incompatibilities. It is anticipated that this paper may be of benefit to researchers and organizers seeking to effectively structure interdisciplinary research projects, specifically in terms of framing the research problem and the modes of inquiry, and in structuring the interdisciplinary research team.

show abstract

Deep Learning in the Natural Sciences: Applications to Physics

Cited by 5 publications

References 70 publications

Energy flow networks: deep sets for particle jets

Energy flow networks: deep sets for particle jets

Climatic and seismic data-driven deep learning model for earthquake magnitude prediction

Interdisciplinary Research as a Complicated System

Contact Info

Product

Resources

About