Directors: Sisi Aarukapalli & Aarian Ahsan
Anish Nyalakonda
We will be developing a sentiment analysis model for detecting whether a movie review is positive or negative. Specifically, we will be delving into Natural Language Processing (NLP) to construct a sentiment analysis model. Our primary objective is to create a Neural Network that is capable of ascertaining whether a given movie review carries a positive or negative sentiment so that we can unravel the intricate emotions embedded within movie reviews.
Naveen Mukkatt
Humans break down complex tasks into simpler steps, so what if we took the same approach towards teaching AIs how to navigate their environment? In Scaffold, you will try to answer this question by studying various reinforcement learning algorithms and applying them to AIs as they learn to navigate their environment. The AI will be piloting a 3D character in a game engine, so if you have familiarity with game design or Unity/Unreal (hi ATEC majors), you're an ideal fit!
Nivedha Sreenivasan
To assist forensics experts in identifying what events took place at the scene of a crime, you will learn the basics of image classification and train a model using convolutional neural networks (CNNs) to classify bloodstain patterns commonly found in crime scenes into three categories: passive, transfer, and impact.
Philip Lee
We aim to rigorously evaluate the efficiency of a Large Language Model (LLM) represented by RoBERTa, in addressing a critical classification challenge: the early detection of cancer. Throughout this study, you will gain a deep understanding of the Transformer Architecture which laid the foundation for all LLMs, and provide insights into addressing critical real-world challenges, contributing to your knowledge of both machine learning and medical applications.
Rishit Viral
Decentralized applications have emerged as a transformative force in the world of blockchain technology, enabling transparent, secure, and censorship-resistant interactions on a global scale. Among the emerging blockchain platforms, Solana stands out for its high-performance capabilities and scalability, making it a promising foundation for building dApps that can handle complex tasks and accommodate growing user demands. In this project, we delve into the exciting realm of decentralized applications on the Solana blockchain, exploring their potential to revolutionize industries and reshape the way we engage with digital services. Through comprehensive analysis and hands-on development, we aim to uncover the unique features and advantages that Solana offers, while addressing challenges and providing insights into the future landscape of decentralized applications.
Rohan Dave
Large model training demands immense computing power. Federated learning presents an innovative solution by leveraging distributed devices instead of relying on a single machine. Blockchain's distinct properties form an ideal base for interactions in this setup. You will learn the basics of deep learning and blockchain to implement federated learning using Ethereum.
Rowan White
AI is often seen as less biased than humans, but it can inherit biases from human-curated data, affecting its decisions. For example, biased hiring practices can lead AI to unfairly rank job resumes based on names. AI's inner workings are often unclear, creating transparency issues. This project delves into various AI models and how XAI can unveil their operations. As a team, you'll select a model and a real-world issue to explore, then use XAI to examine and improve that model.
Saanvi Bala
Speech Emotion Detection plays an integral part in advancing Human Computer Interactions. In this project, you will learn how to analyze audio frequencies and patterns using librosa. You will also make a CNN (Convolutional Neural Network) learning model that will teach the system to recognize the various emotions. Our goal is to make a model that can accurately predict the emotion behind speech.
Victoria Vynnychock
Steganography is used to conceal information within plain sight, while Steganalysis is employed to reveal this information. You will hide messages within images, then utilize various techniques to detect and unveil the concealed content. Afterwards, you will train a machine learning model to classify images that contain hidden information.
Vaishnavi Josyula
Streaming has become a popular form of online entertainment with AI streamers recently gaining traction. In this project, you will combine multiple neural networks to create a virtual streamer who can play games, interact with the chat, and create engaging content, among other aspects. Ultimately, you will deploy the streamer to a streaming platform like Twitch.