Crosslight Software (Canada)

Dual-cavity Vertical-Cavity Surface-Emitting Lasers (VCSELs) provide an effective platform for controlling longitudinal mode behavior. This paper presents a comprehensive numerical investigation of dual-cavity VCSELs with both tuned and detuned cavity configurations, separated by a finite intermediate Distributed Bragg Reflector (DBR). Cold-cavity optical modes, supermode field distributions, and modal parameters are analyzed using a full-vectorial Finite-Difference Frequency-Domain (FDFD) solver implemented in Crosslight’s PICS3D. For tuned cavities, the optical coupling coefficient is extracted directly from the supermode wavelength splitting and correlated with the threshold behavior. For detuned cavities, intentional cavity-length mismatch leads to strong supermode localization and asymmetric gain overlap, resulting in preferential lasing of a single supermode. These results demonstrate that cavity detuning combined with an optimized intermediate DBR thickness provides a robust design approach to control VCSEL’s operation parameters.

Abstract This work presents the Affine Heat Method for computing logarithmic maps. These maps are local surface parameterizations defined by the direction and distance along shortest geodesic paths from a given source point, and arise in many geometric tasks from local texture mapping to geodesic distance‐based optimization. Our main insight is to define a connection Laplacian with a homogeneous coordinate accounting for the translation between tangent coordinate frames; the action of short‐time heat flow under this Laplacian gives both the direction and distance from the source, along shortest geodesics. The resulting numerical method is straightforward to implement, fast, and improves accuracy compared to past approaches. We present two variants of the method, one of which enables pre‐computation for fast repeated solves, while the other resolves the map even near the cut locus in high detail. As with prior heat methods, our approach can be applied in any dimension and to any spatial discretization, including polygonal meshes and point clouds, which we demonstrate along with applications of the method.

TVM is proposed to efficiently deploy deep learning (DL) networks to diverse hardware devices. Tensorization in TVM enables designers to manually set up mappings between tensor computation patterns in DL networks and specific hardware tensor instructions, so as to fully exploit the potential of DL accelerators. However, with the development of modem DL accelerators, a much richer set of hardware instructions are being proposed. This leads to a one-to-many mapping between a tensor computation pattern and hardware instructions, making it a non-trivial task to choose a proper mapping. To this end, we propose Cost Aware TVM (CAT) tensorization. The key insight of CAT is that both tensor computation patterns and instructions share similar tensor features. Based on this insight, CAT first extracts the tensor features of tensor computation patterns and locates eligible hardware instructions accordingly, and then selects the least time consuming hardware instruction via a cost model. Our experimental results across 8 deep learning models show that CAT tensorization can automatically map tensor computation patterns to proper hardware instructions, and our cost model based selection policy achieves an average speed up of 6.4x compared with randomly selecting an eligible hardware instruction for a tensor computation pattern.

In this paper, the FDFD is used to analyze Vertical Cavity Surface Emitting Laser (VCSEL). Based on the structure of the VCSEL, two main models can be used; the 2.5D and the 3D Finite Difference Frequency Domain (FDFD). The full vectorial solver is well suited for the fundamental as well as the higher-order modes and includes different field polarization. The method was used to solve both the reference VCSEL and the Surface-Relief VCSEL.

In recent years, artificial intelligence has seen increasing interest and popularity in the financial services and the other areas like supply chain management. Since the late 1970s, artificial intelligence has been developed to improve human decision-making processes and productivity in businesses, due to the ability to understand patterns and businesses phenomena, to search and analyze information and automating tasks repeatedly by humans. According to Tungsten Network, that valuable time and money are wasted on trivial supply chain related-tasks that are conducted operationally by humans. So, a business could automate some tasks to reduce wasting time with the robotic process, and machine learning algorithms are being integrated into analytics and CRM platforms to uncover information on how to better serve customers. In addition to using artificial intelligence in the supply chain, its presence in financial technology can be noted, including advanced machine learning software and expert systems that are capable of learning and performing intelligent analysis, as well as the automation of some business processes in financial. Meanwhile, given the focus on supply chain finance, which includes solutions for suppliers, manufacturers, vendors, and customers that improve financial processes in the supply chain, the industry is moving toward a revolution and a massive transformation in this approach. This paper gives an overview of artificial intelligence and the application areas of this technology and the current use in financial technology and supply chain management. This paper will also explore the benefits using Artificial Intelligence technologies in the Financial Technology and Supply Chain Management businesses.

Direct tunneling and trap-assisted tunneling models are modelled and implemented into APSYS™, a device simulation tool. By comparing with experimental data, the trap-assisted tunneling model was found to better describe leakage currents than the direct tunneling model. The trap-assisted tunneling model was shown to catch the temperature dependence of the gate leakage current from 300K to 500K.

Proceduralism has become increasingly important to both film and game pipelines. By utilizing procedural approaches, production teams in both film and games can significantly reduce timelines and enhance flexibility. In recent years, due to the increasing demand on gaming pipelines to provide ever increasing amounts of content, and the need to retain revisability of content late into the production cycle, procedural approaches have been gaining traction in gaming pipelines as well. Houdini is a software package known for its procedural approach. However, Houdini's core technology was rather inaccessible because its workflows and interfaces weren't familiar to many artists.Houdini Engine was a project to make Houdini's core technology easier to integrate with other software. This paper details its evolution, where we focus on the many technical decisions and challenges that we had to face along the way, detailing some of the reasoning and tradeoffs behind our software design choices.

Non-common atom interface effects for InAs/GaSb type II superlattices and their implementation in a finite difference k.p model are discussed in this paper. We demonstrate that the additional term added to the Hamiltonian as a result of this perturbation is not compatible with the commonly used block-diagonalization scheme and show some of the limitations associated with using only the transition bandgap to fit the strength of the interface interaction.