IRS-Assisted OTFS: Beamforming Design and Signal Detection

Intelligent reflecting surface (IRS) technology has become a crucial enabler for creating cost effective, innovative, and adaptable wireless communication environments. This study investigates an IRS-assisted orthogonal time frequency space (OTFS) modulation that facilitates communication between users and the base station (BS). The users attainable downlink rate can be boosted by collaboratively improving the reflection coefficient (RC) matrix at the IRS and beamforming matrix at the BS. Then, in the IRS-aided OTFS network, the problem of cooperative precoding at BS and IRS to improve the network throughput is framed. The precoding design problem is non-convex and highly complicated; an alternate optimization (AO) approach is proposed to solve this. Specifically, an approach based on strongest tap maximization (STM) and fractional programming is proposed. It solves RC matrix (at IRS) and beamforming matrix (at BS) alternatively. Moreover, an efficient signal detector for IRS-aided OTFS communication systems using the alternating direction method of multipliers (ADMM) is proposed. Finally, to estimate the cascaded MIMO channel, using a parallel factor tensor model that separates the IRS-User and BS-IRS MIMO channels, respectively is suggested. Simulation results show that the proposed method significantly enhances the system capacity and bit error rate (BER) performance compared to conventional OTFS.

VISU at WASSA 2023 Shared Task: Detecting Emotions in Reaction to News Stories Leveraging BERT and Stacked Embeddings

Our system, VISU, participated in the WASSA 2023 Shared Task (3) of Emotion Classification from essays written in reaction to news articles. Emotion detection from complex dialogues is challenging and often requires context/domain understanding. Therefore in this research, we have focused on developing deep learning (DL) models using the combination of word embedding representations with tailored prepossessing strategies to capture the nuances of emotions expressed. Our experiments used static and contextual embeddings (individual and stacked) with Bidirectional Long short-term memory (BiLSTM) and Transformer based models. We occupied rank tenth in the emotion detection task by scoring a Macro F1-Score of 0.2717, validating the efficacy of our implemented approaches for small and imbalanced datasets with mixed categories of target emotions.