Abstract
The Exa.TrkX project presents a graph neural network (GNN) technique for low-level reconstruction of neutrino interactions in a Liquid Argon Time Projection Chamber (LArTPC). GNNs are still ...a relatively novel technique, and have shown great promise for similar reconstruction tasks in the Large Hadron Collider (LHC). Graphs describing particle interactions are formed by treating each detector hit as a node, with edges describing the relationships between hits. We utilise a multi-head attention message passing network which performs graph convolutions in order to label each node with a particle type.
We present an updated variant of our GNN architecture, with several improvements. After testing the model on more realistic simulation with regions of unresponsive wires, the target was modified from edge classification to node classification in order to increase robustness. Removing edges as a classification target opens up a broader possibility space for edge-forming techniques; we explore the model’s performance across a variety of approaches, such as Delaunay triangulation, kNN, and radius-based methods. We also extend this model to the 3D context, sharing information between detector views. By using reconstructed 3D spacepoints to map detector hits from each wire plane, the model naively constructs 2D representations that are independent yet fully consistent.
Online trading has become an important component of the portfolio of investors as it provides an attractive and profitable investment option in spite of the risk involved. Many individuals embraced ...diversified financial portfolios and started investing in Stocks, Cryptocurrencies, Bonds, Metals, and Commodities through online portals. It needs constant attention to the price of the financial commodities to obtain profits. However, the majority of the investors are landing in losses because of the lack of continual monitoring of the prices to ascertain the changes in the prices of financial commodities all the time. In this work, to assist users in order to avoid this tiresome task of continual monitoring, a prototype of a low-cost portable device for change indication in the price of multiple financial commodities of various trading paradigms based on NodeMCU is proposed. The device includes a Data retrieval and processing (DRP) unit in the form of nodeMCU, control unit, and notification unit. The DRP unit initially obtains the list of the financial commodities and control parameters including sampling rate, threshold, and time delay from the Google Firebase real-time database which are collected from user through a mobile application. Further, it fetches the real-time price data of the financial commodities to compute the percentage change in the given time duration and obtain the prioritized list i.e., the list of financial commodities as per the descending order of absolute percentage change and instructs the notification unit to generate alarm through an audio buzzer in case at least if one of the financial commodity exhibits an absolute percentage change greater than a predefined threshold. The notification unit also contains Red and Green LEDs corresponding to each of the financial commodities to indicate negative and positive changes that exceed the threshold. Further, an LCD displays the details of the first financial quantity in the prioritized list. The control unit comprises two designated switches to change the financial quantity whose details are to be displayed on the LCD and a silent mode switch to disable the audio notifications for the sake of convenience of the user. A prototype of the proposed device is realized with 16 cryptocurrencies, a sampling rate of one sample per minute, and a time delay of one hour for the purpose of testing the device. A maximum of 0.20 % of the mean error rate is observed in the price of Dogecoin, and 0% of the mean error rate is observed in the percentage change of price corresponding to all cryptocurrencies. Further, the number of notifications using RED and GREEN LEDs are computed for all cryptocurrencies which are found to be exactly in match with the original data with various threshold values of 1%,2%,5%, and 10%. The total number of notifications including both Red and Green LEDs is found to be higher in HBAR. The proposed device assists investors in making timely decisions for profitable trading and provides valuable insights about the momentum of various markets to readjust the investments and balance the portfolios accordingly.
