Two-Stage Learning for Uplink Channel Estimation in One-Bit Massive MIMO. (arXiv:1911.12461v1 [eess.SP])

We develop a two-stage deep learning pipeline architecture to estimate the uplink massive MIMO channel with one-bit ADCs. This deep learning pipeline is composed of two separate generative deep learning models. The first one is a supervised learning model and designed to compensate for the quantization loss. The second one is an unsupervised learning model…

Spectrum Cartography via Coupled Block-Term Tensor Decomposition. (arXiv:1911.12468v1 [eess.SP])

Spectrum cartography aims at estimating power propagation patterns over a geographical region across multiple frequency bands (i.e., a radio map)—from limited samples taken sparsely over the region. Classic cartography methods are mostly concerned with recovering the aggregate radio frequency (RF) information while ignoring the constituents of the radio map—but fine-grained emitter-level RF information is of…

Minimum Bayes Risk Training of RNN-Transducer for End-to-End Speech Recognition. (arXiv:1911.12487v1 [cs.CL])

In this work, we propose minimum Bayes risk (MBR) training of RNN-Transducer (RNN-T) for end-to-end speech recognition. Specifically, initialized with a RNN-T trained model, MBR training is conducted via minimizing the expected edit distance between the reference label sequence and on-the-fly generated N-best hypothesis. We also introduce a heuristic to incorporate an external neural network…

A Dynamical Model of Oncotripsy by Mechanical Cell Fatigue: Selective Cancer Cell Ablation by Low-Intensity Pulsed Ultrasound (LIPUS). (arXiv:1911.12407v1 [physics.bio-ph])

The method of oncotripsy, first proposed in [S. Heyden and M. Ortiz (2016). Oncotripsy: Targeting cancer cells selectively via resonant harmonic excitation. Journal of the Mechanics and Physics of Solids, 92:164-175], exploits aberrations in the material properties and morphology of cancerous cells in order to ablate them selectively by means of tuned low-intensity pulsed ultrasound…

Classification of coherent enhancements of light-harvesting processes. (arXiv:1911.12376v1 [physics.chem-ph])

In recent years, several kinds of coherence have been shown to affect the performance of light-harvesting systems, in some cases significantly improving their efficiency. Here, we classify the possible mechanisms of coherent efficiency enhancements, based on the types of coherence that can characterise a light-harvesting system and the types of processes these coherences can affect.…

Dust-acoustic rogue waves in non-thermal plasmas. (arXiv:1911.12386v1 [physics.plasm-ph])

The nonlinear propagation of dust-acoustic (DA) waves (DAWs) and associated DA rogue waves (DARWs), which are governed by the nonlinear Schr\”{o}dinger equation, is theoretically investigated in a four component plasma medium containing inertial warm negatively charged dust grains and inertialess non-thermal distributed electrons as well as iso-thermal positrons and ions. The modulationally stable and unstable…

Simultaneous Segmentation and Relaxometry for MRI through Multitask Learning. (arXiv:1911.12389v1 [physics.med-ph])

Purpose: This study demonstrated an MR signal multitask learning method for 3D simultaneous segmentation and relaxometry of human brain tissues. Materials and Methods: A 3D inversion-prepared balanced steady-state free precession sequence was used for acquiring in vivo multi-contrast brain images. The deep neural network contained 3 residual blocks, and each block had 8 fully connected…

Field emission microscopy of carbon nanotube fibers: evaluating and interpreting spatial emission. (arXiv:1911.12393v1 [physics.app-ph])

In this work, we quantify field emission properties of cathodes made from carbon nanotube (CNT) fibers. The cathodes were arranged in different configurations to determine the effect of cathode geometry on the emission properties. Various geometries were investigated including: 1) flat cut fiber tip, 2) folded fiber, 3) looped fiber and 4) and fibers wound…

Improved depth resolution and depth-of-field in temporal integral imaging systems through non-uniform and curved time-lens array. (arXiv:1911.12397v1 [physics.optics])

Observing and studying the evolution of rare non-repetitive natural phenomena such as optical rogue waves or dynamic chemical processes in living cells is a crucial necessity for developing science and technologies relating to them. One indispensable technique for investigating these fast evolutions is temporal imaging systems. However, just as conventional spatial imaging systems are incapable…

Model-Aware Deep Architectures for One-Bit Compressive Variational Autoencoding. (arXiv:1911.12410v1 [eess.SP])

Parameterized mathematical models play a central role in understanding and design of complex information systems. However, they often cannot take into account the intricate interactions innate to such systems. On the contrary, purely data-driven approaches do not need explicit mathematical models for data generation and have a wider applicability at the cost of interpretability. In…