Yash Narayan | Searcharxiv

arXiv2025

Post-Newtonian Dynamics of Radiating Charges: Canonical Formulation and Binary Inspiral Laws

We revisit an explicit electromagnetic analogue of the Post Newtonian Hamiltonian framework widely used in gravitational wave physics. Starting from the Lorentz Dirac equation, we implement the Landau-Lifshitz order reduction to cast the 1.5PN radiation reaction force in terms of a double sum in canonical variables and incorporate this into the well known 1PN Darwin Hamiltonian system. The resulting phase space is strictly conservative when dissipation is switched off, while in presence of dissipation, it exhibits monotonic energy loss during the inspiral, accompanied by orbit circularization and eccentric bursts in the evolution of the Darwin Hamiltonian. Using this phase space framework we compute the circular and eccentric inspiral laws, including $1$PN conservative corrections. Extending to charged compact binaries in Einstein-Maxwell theory, we combine the known $2$PN ADM-type conservative Hamiltonian with leading $1.5$PN dipole dissipation and $2.5$PN gravitational quadrupole flux, obtaining gauge-invariant energy-frequency relations, closed-form inspiral laws, and a dipole-quadrupole crossover scale that separates electromagnetic and gravitational flux dominated inspirals.

gr-qc↗

arXiv2021

DeepWaste: Applying Deep Learning to Waste Classification for a Sustainable Planet

Accurate waste disposal, at the point of disposal, is crucial to fighting climate change. When materials that could be recycled or composted get diverted into landfills, they cause the emission of potent greenhouse gases such as methane. Current attempts to reduce erroneous waste disposal are expensive, inaccurate, and confusing. In this work, we propose DeepWaste, an easy-to-use mobile app, that utilizes highly optimized deep learning techniques to provide users instantaneous waste classification into trash, recycling, and compost. We experiment with several convolution neural network architectures to detect and classify waste items. Our best model, a deep learning residual neural network with 50 layers, achieves an average precision of 0.881 on the test set. We demonstrate the performance and efficiency of our app on a set of real-world images.

cs.LG↗