Abstract :

This study presents Lexi Genius, an innovative and accessible tool based on a Large Language Model (LLM) specifically designed for offline settings. When Internet access is limited, Lexi Genius offers a suite of features designed to improve reading comprehension, including AI/ML-based text summaries, focused on Science and Technology (SandT) documents and article summaries. Above all, Lexi Genius stands out for its distinctive approach to offline text enhancement and simplicity that ensures user-friendly communication. In addition, its scalability and flexibility allow developers to easily add new

Abstract :

This paper discusses the importance of examining Gait, EEG, and Speech data to diagnose neurological conditions associated with Parkinson disease (PD). The numerous methods for analyzing Parkinson's disease by separating the electrical activity of brain signals are discussed in research articles. Although EEG systems have relatively limited spatial resolution and will be contaminated with numerous aberrations, they are frequently used to extract brain signals since they are quite capable of monitoring electrical activities of the brain. Speech signal analysis is frequently used to detect Parkinson's disease because current surveys show that roughly 70% to 90% of patients with the condition exhibit dysphonic (impaired speaking ability) symptoms. The speech quality may vary as a function of the surrounding environment and motion disturbances, which could lead to an incorrect diagnosis. Despite the above-mentioned limits of gaits, EEG and Speech Signals might result in a flawed diagnosis of the diseases and may result in subtle manifestations of faults, despite the fact that researchers have suggested a variety of risk factors for appropriate analysis of PD using these signals. Therefore, to more accurately detect these neurological illnesses using EEG and voice signal analysis, a powerful diagnostic system is required. This paper discusses the various approaches and methodologies used in diagnosing neurological conditions like PD using speech and EEG signals, which gives researchers a better understanding of how to carry out future research and create an effective integrated algorithm for diagnosing the aforementioned neurological conditions using Gaits, speech and EEG signals

Keywords: parkinson’s disease, Gaits, EEG, Speech signal.

Abstract :

In this project, we propose a multi-modal deep learning methodology to predict imminent stock market crashes, aiming to enable policymakers and financial institutions to take proactive measures. Building on previous research, we strive to develop a more accurate predictive model by integrating both quantitative financial market data and qualitative sentiment from news headlines. This approach is essential due to the increasing complexity and volatility of modern global markets. Our dataset comprises two primary components: quantitative market data, including closing prices of key Vanguard sector ETFs and the VIX volatility index from August 2008 to July 2016, and qualitative news data, encompassing Reddit WorldNews headlines for each trading day within the same period. The model architecture consists of three main components: a market data encoder, a news data encoder, and a final evaluator. For the market data encoder, various configurations of dense and LSTM layers were tested, with a 64-dense layer model performing best. For the news data, headlines were first encoded using a pre-trained Universal Sentence Encoder, followed by dense and LSTM layers, with a 2-dense layer model proving optimal. The encoded market and news data were then concatenated and passed to the evaluator, which utilized dropout layers to make the final crash prediction. The model achieved an AUC of approximately 0.7 in this task. Key findings include the effectiveness of lightweight models, likely due to the limited variability and temporal scope of the data, and the challenge some models faced in capturing meaningful spatial relationships in the financial data. The main contribution of this work is demonstrating the value of incorporating qualitative news sentiment alongside quantitative market data for more accurate crash prediction.

Abstract :

The issue of mental health has become a global concern over time. Although there are mental health services available, they are expensive and difficult for the general public to get. The objective of this research project is to explore technology and robotics in the hopes of developing a cost-effective alternative for mental health services. The goal of the project is to provide both students and non-student users with optimal access to readily available, affordable mental health services. The suggested robot companion aids in addressing procrastination and anxiety. In this study, a robot companion with application components to help with various mental health concerns is introduced. The Pomodoro Technique, an anxiety-reduction strategy, and a task list system that helps with procrastination, anxiety, and stress are the three key systems that this robot encompasses. The Adalo app is used to create the robot's virtual application. Adalo is an app creator that doesn't require knowledge of programming languages. The pomodoro technique, the anxiety-reduction system, and the anti-procrastination system are the three main systems included in the app, and the robot itself has a spinning body that works similarly to a fidget spinner to help reduce stress. In addition, several tests were conducted on the robot to guarantee its efficacy and efficiency. The survey's positive findings point to a promising future for mental health.