CSE 2022 Spring Departmental Demo Day

We’re thrilled to invite you to the eighth bi-annual Comp. Sci. & Eng. Fall Demo Day. Student groups from several CSE capstone classes will be presenting the culmination of 3-months of effort, hard work, (metaphorical) blood, sweat (well… caffeine really), and tears (see above).

  Where:

Davis Hall; 1st and 2nd Floor Atrium

  When:

Friday May 13, 2022

Schedule

2:00 PM - Event opens
2:05 PM - Opening remarks
2:10 PM - Judging starts
4:00 PM - Judges will get together to review results and pick winners
4:30 PM - Winners announced and videos played for the winning teams
5:00 PM - Event closes

Winners

1st Place
Parkinson's Gait Assistance
2nd Place
Electroskip Force Sensor Sensitivity
3rd Place
Open Beats
Innovation prize
GEP

Acknowledgements

Sponsors

Judges

Presented Projects

This year’s participating classes and projects include:

Master's Capstone Project (CSE 611)

Distributed systems simulator
  • Jayanth Gollapinni - sgollapi@buffalo.edu
  • Mahideep Tumati - ntumati@buffalo.edu
  • Sai Teja Veluru - sveluru2@buffalo.edu
  • Tejaswini Sana - sairamat@buffalo.edu
  • Vinay V.M. Urlam - vurlam@buffalo.edu
Distributed systems simulator is a web application that lets users simulate distributed systems protocols interactively. The primary intent of the application is to help students who take up distributed system classes, understand and visualize its protocols. Distributed systems are hard to understand, the protocols deal with unreliable clocks, process pauses, partial failures involving many parameters. Our interactive simulator will help distributed systems enthusiasts visualize and understand the concepts more intuitively by letting the users play with node failure, and other factors involved with the specific protocol, which will help the users deepen their understanding.

Distributed systems simulator includes Discrete event simulator API which has features of creating the computer nodes and network connecting these nodes, send message between these nodes, drop the message, turn the nodes off and on, simulation timers etc. Currently the application consists a Heartbeat protocol and a PUB-SUB protocol. This API will be helpful in creating any new protocol simulation. This application is developed in HTML, CSS and JS. The application runs in browser that supports javascript version ES6.
Food Court
  • 1. Radhakrishna Mohan Eadara (readara@buffalo.edu)
  • 2. Shiva Kumar Mukkapati (smukkapa@buffalo.edu)
  • 3. Niharika Talluri (ntalluri@buffalo.edu)
  • 4. Ritwik Doppalapudi (sdoppala@buffalo.edu)
  • 5. Varshith Reddy Danda (vdanda@buffalo.edu)
The goal of this project is to develop a restaurant dashboard, add analytics in the admin and restaurant consoles and enhance the existing mobile application that encourages people to share what they are eating and discover what others are eating to ultimately help solve the problem of not knowing what they want to eat in a fun, interactive way (group foodie).

1. Group foodie (Group of users can easily decide on what to eat and where to eat in their vicinity)
2. Analytics in the admin console with graphical representation (Daily, Weekly, Monthly)
3. Managing restaurants (Search, add and update menu items, update restaurant information, analytics)
Helpt App
  • Akshaya Mohan - akshayam@buffalo.edu
  • Akshay Sahai - asahai2@buffalo.edu
  • Aravind Balakrishnan - balakri2@buffalo.edu
  • Kiranmai Namburu - kiranmai@buffalo.edu
  • Prasad Shirvandkar - prasadda@buffalo.edu
Helpt is an On-Demand Services company looking to create a two-way delivery system for freelancers and clients to meet together.

Helpt was created to help as many people as possible with not only a great product but also by building a great community. A single platform to connect both freelancers and clients.

With Helpt, your help is at your fingertips. Just open the app and enter what service you need help with, and a nearby helper will help you complete the job reliably.
Look the Part
  • varshith@buffalo.edu
  • mmusku@buffalo.edu
  • bandarup@buffalo.edu
  • mythreye@buffalo.edu
  • saisrava@buffalo.edu
  • sitikala@buffalo.edu
Timecapsule
  • kc64746@buffalo.edu - 50418476
  • sakethba@buffalo.edu - 50414412
  • sahithba@buffalo.edu - 50414758
  • lalithsa@buffalo.edu - 50414463
  • hramdas@buffalo.edu - 50376171
The goal of this project is to build a mobile application and a web console called Time Capsule. TimeCapsule is an app that helps you to talk to your future self and also with future generations - family/friends so that you are more connected to your loved ones, anywhere, anytime, even after you are gone. This allows your legacy to live on for many generations to come.

The project is to Create, preserve memories and experiences to cherish by looking at those after some long time and also to remind beautiful moments to the future generations. In the application, The users must sign up. Free tier includes storage upto 50gb. Premium service includes different plans and users can sign up for premium service based on their needs. Users can book a session with a time capsule by selecting the “Contact us” section of the mobile application. Admin can access the web application and upload videos of a particular user in their premium section.

Essentially, The time capsule app assists in saving important events that the user wants to preserve and can share among other users within the app. Users can set up (create an account) their profile in the time capsule app. Upon creating a new account.
The user can perform the following actions with the application.
- The user can save photos, videos, voice recordings, descriptions.
- The user data is saved in a timely fashion which means the user selects any year and he can go through the saved events anytime and anywhere.
- The user can retrieve the saved data from a cloud storage.
- The user can share the data with other users.
- It has inbuilt video player , Audio player and Image viewer for the users to access their contents.
Dancestry
  • Srikar Panuganti <srikarpa@buffalo.edu>
  • Tarun Kodavati <vkodavat@buffalo.edu>
  • Jeevan Sagar Batana <jbatana@buffalo.edu>
  • Tejasri Karuturi <tejasrik@buffalo.edu>
  • Sumanth Reddy Adupala <sadupala@buffalo.edu>
  • Bharath Gangishetti <bgangish@buffalo.edu>
Dancestry is an interactive, app-based lineal network illustrating connections between dance artists, their teachers and choreographers they have danced in the work of.

We will be illustrating lineal network in form of a interactive network graph.
STOOTY
  • Adithya Venkatesh-venkate9@buffalo.edu
  • Ankit Shetty- ankitdiw@buffalo.edu
  • Neha Bisht - nehabish@buffalo.edu
  • Shubhangi Balasubramanian- balasub7@buffalo.edu
  • Sushikth Shetty- sushikth@buffalo.edu
  • Meet Thosani - meetpiyu@buffalo.edu
STOOTY, an online service in the Entertainment fintech and Creative production domain, solves the problem of creators and studio operators by providing them a platform to perform activities like book spaces and services simply and instantly, and studio owners can easily finance, market, book, and manage their production services using this platform.

We have added the following features in the application for a smooth user experience wherein the creators and artists will be able to share their social media platforms to display their content and vice versa, they can use @ or # to search for a user or a listing respectively, we have also given admin rights to manage multiple users, a listing can also be promoted which shows up as a sponsored ad banner to the users, users can comment and give reviews on studio spaces and services, the booking details can also be integrated with google calendar of the user.

Tech Stack:
Languages:Front-end :HTML/CSS3, BOOTSTRAP5, JAVASCRIPT, JQUERY
Back-endPHP
Database:MySql
Cloud Service:Namecheap
Version Control:git
ACV Auctions
  • Rithvik - Rithvikas@buffalo.edu
  • Bhushan Parwani - Bhushanp@buffalo.edu
  • Ed Christian - Christi6@buffalo.edu (will not be in attendance on presentation day due to conflict)
An ML object detection application built with PyTorch and YOLO5 to identify instrument cluster lights in vehicles, streamlining an otherwise manual, error prone, and time-consuming process.

Demo involves a live dashboard and imaging to demonstrate real-time the object detection capabilities of the application.
LifeTreeWNY
  • dgarude@buffalo.edu
  • anishana@buffalo.edu
  • ldodda@buffalo.edu
  • anushnag@buffalo.edu
  • shaikazi@buffalo.edu
The project is a web application for the management of services provided by the business LifeTreeWNY in the Western New York region. The webapp houses the entire lifecycle right from getting customer orders to managing invoices, to ensuring the the job is done with automatic follow up emails and a maps view to visualize the locations of jobs to be performed.

The project integrates Quickbooks, an accounting software along with dropbox to provide employees access to customer files and uses google maps to provide the location data of the customer's requests.
Patcom Medical
  • Mounisha Kallepalli - mounisha@buffalo.edu
  • Vishnu Priya Bollam - vbollam@buffalo.edu
  • Kaushik Reddy Kadapakonda - kkadapak@buffalo.edu
  • Sai Teja Cherukuri – cheruku3@buffalo.edu
  • Rahul Reddy Karumuri – rkarumur@buffalo.edu
  • Dileep Narayana Garimella – dileepna@buffalo.edu
Findings of Endoscopic evaluation of swallowing

The project is a web-based application that allows authorized medical users to do the FEES study, create reports and store the patient’s information. The user can log on to the application, record the videos through the endoscopic cameras. While recording the videos, a manual process of screenshotting the video can be done. These are later added to the reports respective to the patients via basic button click. Furthermore, screenshots can be edited using basic graphic tools to highlight interesting or crucial parts of the image.
Essential Vending
  • Saqlain Naveed Ahmed (sahmed34@buffalo.edu)
  • Pranav Moreshwar Sorte (pranavmo@buffalo.edu)
  • Nupur Atul Chauhan (nupuratu@buffalo.edu)
  • Shreya Satish Shetty (sshetty5@buffalo.edu)
  • Carolyn Joseph Bose (cbose@buffalo.edu)
Essential Vending is a mobile application that aims to make the process of purchasing items from the vending machine more convenient. Currently the vending machines only accept payments through cash. However, this method is rather inefficient since more than 60% of people don’t carry cash everyday. Our goal is to design a system that is capable of enabling hassle-free purchase of items within the vending machine through several online payment options.

The application also includes features such as navigation using Google Maps to locate the vending machines, giving detailed statistics and insights about the purchases and user activity to the administrators as well as incorporating ads and promotions within the application.
Anastocare
  • Dexson (dexsonjo@buffalo.edu)
  • Anuj (anujswap@buffalo.edu)
  • Saistha (saisthak@buffalo.edu)
  • Mojitha (mojitham@buffalo.edu)
  • Mousam (mousamsa@buffalo.edu)
Limitless medical technologies

Anastocare is a web and mobile application that evaluates a patient's health and alerts the user and their healthcare professional if the patient is at risk of developing an abdominal leak in the near future. This allows both the doctor and the patient to be aware of the patient’s health condition and take appropriate measures to avert complications.
Ionita Lab Project
  • Kusal Nagabhairava ( kusalnag@buffalo.edu)
  • Praneeth Sai Kotha (pkotha@buffalo.edu)
  • Sri Rudra Tetali (srirudra@buffalo.edu)
  • Jeevana Chintapalli (jeevanac@buffalo.edu)
  • Manoj Reddy Kesireddy ( mkesired@buffalo.edu).
The Ionita project principally aims to develop a tool that can help the surgeons real-time
during an aneurysm treatment. It analyses the DSA ( Digital Subtraction Angiography) images taken before and after the aneurysm procedure and predicts the possibility of success of the treatment. Angiographic Parametric Imaging Maps (API) are generated from the fed DSA images, and specific key parameters are measured using the tool, which are required for the prediction.

NA
SportStretchUSA
  • Akhil Reddy Yeredla (ayeredla@buffalo.edu)
  • Teja Reddy Alla (tejaredd@buffalo.edu)
  • Bhargav Sai Kantipudi (bkantipu@buffalo.edu)
  • Harikanth Dasoji (harikant@buffalo.edu)
  • Venkata Prathima Bhargavi Karri (vkarri@buffalo.edu)
SportStrechUSA is an app that connects recovery specialists with athletes for the most part. It is for just about anyone who is looking to gain flexibility in order to perform better on the field, in the gym, and in everyday activities. The appointments need to be scheduled in advance and the payment is done after the service is completed. This is where our app comes into the picture i.e. it helps the recovery specialists to manage their calendar and allot the slots for the user to book. The users can select from a list of recovery specialists in the location of their choosing and can decide whom to choose based on their ratings and reviews. Also, the admin can manage the recovery specialists registered with the app and can see all the appointments made through the app.

The app has an interactive calendar feature, google maps are integrated directly into the UI, payment module offered through the square is used for the transactions in the app.

The recovery specialists have the ability to choose the radius around them i.e. how far they can travel to serve their appointments. They can also decide the cost of a particular appointment based on parking availability, elevator availability, pets, etc. There are also subscription tiers that have certain perks as the tier goes up.

Admin can keep or remove the recovery specialists from the app based on their credibility and reputation.
HOME
  • Kyle Pellechia ktpellec@buffalo.edu
  • Meghna Sriram meghnasr@buffalo.edu
  • Lakshmi Chandana Yarramreddy lyarramr@buffalo.edu
  • Gurudev Yalagala gurudevy@buffalo.edu
  • Sai Chetan Thalla saicheta@buffalo.edu
  • Atharva Avinash Deshpande deshpan5@buffalo.edu
Educational Training website to inform people of Housing and Tenant Laws as well as giving the HOME team information to best reach out to more people

Admin console allows HOME staff to view user data, video and quiz analytics, as well as adding/editing videos and quizzes for future trainings
Road Geometry
  • Steve Thomas - stevetho@buffalo.edu
  • Anup Atul Thakkar - anupatul@buffalo.edu
  • Tanuja Joshi - tanujajo@buffalo.edu
  • Kunal Pal Varma - kunalpal@buffalo.edu
  • Pushkaraj Joshi - pjoshi6@buffalo.edu
An efficient crowd sourcing engine for mapping road geometry.

An end-to-end system, that segments crowd-sourced data from mobile devices referencing open street maps road data. This can be used for various mapping applications such as road geometry estimation, road quality (pothole/s) estimation, etc. Users can then interact with the generated mapping data road-segment wise on a web app.
CamoHealth
  • Abhiramasundari Vijayakumar - vijayak2@buffalo.edu
  • Aishwarya Rath - arath@buffalo.edu
  • Daksh Khorana - dakshsac@buffalo.edu
  • Dakshil Shah - dakshilk@buffalo.edu
  • Parth Dixit - parthjig@buffalo.edu
  • Shraddha Singireddy - ssingire@buffalo.edu
Maternal Health care website would help pregnant women to understand what kind of help or which kind of specialty they need to choose. Using the same information the patient can look for and make booking at their nearest specialist available in real time. In this project, the doctors can register themselves and manage the patient's booking and help them as per their need. There is also a complaint/review section that allows patients to provide their feedback on the appointment made.

-
Vizier DB
  • Name: Keshannagari
  • Sphoorthi
  • Email: skeshann@buffalo.edu
  • Name: Mehra
  • Sakshi
  • Email: sakshime@buffalo.edu
  • Name: M
  • Anvesh
  • Email: somanaga@buffalo.edu
  • Name: Parvat
  • Aniruddha
  • Email: aparvat@buffalo.edu
  • Name: Singh
  • Paramveer
  • Email: psingh46@buffalo.edu
  • Name: Vatwani
  • Nikhil
  • Email: nvatwani@buffalo.edu
Vizier is a reproducibility-oriented notebook environment. Currently code can be edited in the notebook itself. The aim of this project is to add a feature to open code in a local editor and enable code completion for python in user's default code editor.
Vizier notebooks do not have all the features of a text/code editor such as sophisticated autocomplete, documentation, shortcuts etc. The ability to edit on the local default editor would put the user in a familiar environment. With the “Open In Editor” button the user does not have to copy code to and from the Vizier notebook. Integrating linting with mypy plugin in user's local editor. Thus with these features the developer experience with VizierDB is enhanced.

This is the only notebook which provides seamless swithching between web ui and local editor. The feature we have implemented is cross platform. This can be precursor for direct editor integration.
Open Beats
  • Harman Singh (harmansi@buffalo.edu)
  • Akshata Singh (akshatas@buffalo.edu)
  • Ruthvik Ledalla (ruthvikl@buffalo.edu)
  • Srilakshmi Velpuri (svelpuri@buffalo.edu)
  • Bhavyasai Survepalli (bhavyasa@buffalo.edu)
  • Venkata Siva Naga Sai Aravind Kollipara (vkollipa@buffalo.edu)
A web platform that lets music artists & producers:

1.) Collaborate on music remotely.

2.) Distribute music to social media with one click.

3.) Stream music in uncompromised quality.

Open Beats is a music studio in the cloud. It provides a digital audio workstation for music artists to collaborate remotely and a social media platform to connect with other artists looking to collaborate and share music.
Energy-Effecient In-Memory Acceleration for Edge Devices
  • Montana Lee: montanal; Emma Waldvogel: emmawald
When performing calculations to make inferences based on sensory/input data, one of the most costly operations is transferring data between CPU and memory. This is done in order for the CPU to complete some calculation, then propagate the resulting data back to memory. These transfers are debilitating to devices that must be energy-efficient due to their nature (drones, etc.), since energy is used for this transfer process.

To remedy this, models of memory hardware capable of in-memory processing have been developed to use within devices that must make decisions based on some algorithm or decision model. If a machine learning model is trained and its training weights are used within memory to perform the computation-transfer heavy operations, then rapid inferences can be made from input.

To improve upon the model used within the previous paper, two more models have been developed and tested. The MNIST handwritten number dataset was simply used to simulate an input and classifications problem- any situation in which a system must make inferences based off of a trained model may replicate a similar hardware and algorithm solution.


see project description
Drug Simplifier
  • 1. Shreya Jain - sjain28@buffalo.edu
  • 2. Sasank Yalavarthi - sasankya@buffalo.edu
  • 3. Aprishyta Nayak - aprishyt@buffalo.edu
  • 4. Vishal Patil - vpatil4@buffalo.edu
  • 5. Palek Naithani - paleknai@buffalo.edu
  • 6. Aniket Bhamani - abhamani@buffalo.edu
The project is called "Adawi" and there are two sides for this project: Admin and User. There can be three types of users; patient, caregiver, or both. Patients can either manage their own prescriptions or ask caregiver to do it for them. Sometimes, it is difficult to recognize a medicine just by a name. In order to avoid the confusion, we have included images, audio, and description of the medicine so that the user can easily identify it. User can either choose an existing medicine, or can add the medicine with their own images and audio. If they are not feeling good either with the UI or the medicine, they can chat with the administrator and get their problems resolved. The major role of the admin is to approve the medicines, add medicine to the database, and resolve the complaints of the user. Our project is a web-based application and the tech stack used for this project is Java, React, AWS, and MySQL.

Not Applicable
Find a Mechanic
  • Shivani Reddy Katta - skatta@buffalo.edu
  • Saideep ganireddy - saideepg@buffalo.edu
  • Nikhil Bandari - nbandari@buffalo.edu
  • Venkata Subramanya Krishna Kuruganty - vkurugan@buffalo.edu
  • Ashiwn Panditrao Jadhav - ajadhav5@buffalo.edu
  • Roshan Saundankar- rsaundan@buffalo.edu
Find a mechanic is a mobile application that allows the general public to locate trustworthy and genuine mechanics or automobile technicians in their preferred locality to take care of their car.

1)Mobile application - Used by users and mechanics (features : Calendar, Time slot, cancel appointment, Car details) .
2)SCO web application - Will be used by SCO ( features : Calendar, Time slot, Repair history, Front desk ) .
3)Admin web application - Will be used by Admin.

Open-Source Programming Club

MOS Authenticator
  • Emil Kovacev emilkova
  • John Abramo jmabramo
  • Alan Soto alansoto
  • Matt Rubino mrrubino
MakeOpenSource Authenticator is an open-source authenticator app with additional quality-of-life features that easily integrates into existing applications and offers similar levels of security and convenience to paid alternatives.

MOS Authenticator is built by students at MakeOpenSource, UB's open source development club. If anyone is interested in joining, feel free to visit our website at https://makeopensource.org!

Parallel and Distributed Processing (CSE 603)

PySCoOL
  • Matt Ferrera <msferrer@buffalo.edu>
  • Raman Ghimire <ramanghi@buffalo.edu>
  • Shriram Ravi <shriramr@buffalo.edu>
SCoOL is a simple programming model designed to facilitate and accelerate the search space exploration phase of the optimization processes. Although the programming model is simple, SCoOL is written in C++, which is not! This project adds Python bindings to the SCoOL library, allowing users to define the optimization details with simple Python semantics while execution remains delegated to more rapid C++ code.

N/A
ODS Loggers
  • Aman Harsh (amanhars@buffalo.edu)
  • Deepika Ghodki (dghodki@buffalo.edu)
  • Jacob Goldverg (jacobgol@buffalo.edu )
  • Neha Mishra (nehamish@buffalo.edu )
Dynamic Optimization of File Transfers | OneDataShare.org

OneDataShare is an open-source tool for fast and secure file transfer with support for most cloud storage services and protocols like Google Drive, Amazon S3, etc. It aims to remove the burden of managing end-to-end data transfers from the shoulders of users. The file transfer is optimized based on certain parameters that remain constant throughout. In this project, we change the optimization parameters of the transfer at runtime and improve the transfer rate. The parameters are fine tuned dynamically using a Bayes Optimizer. Further, we also collect and report the state of the network to allow the user to understand the bottlenecks in their network.
PoLaRis
  • Sai Vishwanath Venkatesh - saivishw@buffalo.edu
  • Zainul Abideen Sayed - zsayed@buffalo.edu
  • Nithin Sastry Tellapuri - ntellapu@buffalo.edu
Developed a general-purpose runtime execution engine for SCoOL (a framework for scalable search space exploration and optimization) using HPX (a parallel runtime framework similar to MPI)

The runtime execution engine is implemented for a shared memory system. It takes advantage of framework features like HPX futures and HPX threads for asynchronous processing of workload. The solution also relies on phmap's implementation of flat hash set that is much more efficient than the unordered set. The executor strictly confirms to the BSP model enforced by the SCoOL framework
Processing Pandas
  • Krishna prasad Porandla - kporandl@buffalo.edu
  • Suhas Reddy Edavalli - suhasred@buffalo.edu
  • Viswa Nihar Nukala - viswanih@buffalo.edu
Developed a general-purpose runtime system for SCoOL (Scalable Common Optimization) Library using OpenMP. This library helps in solving large scale optimization problems in shared memory cluster.

The runtime execution system is implemented for solving large scale optimization problems in shared memory systems. This uses task based parallel programming techniques and is built on Intel OpenMP API but can be extended to support other shared memory programming models. The executor is akin to the classic Bulk Synchronous Parallel (BSP) model.

Embedded Controls (CSE 499)

Statera
  • Andrew Woska - agwoska
  • Daniel Maas - dwmaas
  • Imani Muhammad-Graham - imanimuh
An embedded control system that allows the robot to balance itself through a PID controller.

A physical demo will be present at the event.
Automated Blind Control
  • jessebot@buffalo.edu
  • bodhiswa@buffalo.edu
  • avitomba@buffalo.edu
  • gsyengle@buffalo.edu
The embedded control system will take into account the amount of transient light available and the external temperature to deduce if the blinds need to be drawn or not.

A typical household loses about 10% of its heat through its windows. In an industrial building, the amount of heat energy loss would be at least thrice the amount. To counteract this loss of heat energy, heater and air ventilation systems produce more heat. This energy inefficiency could be cut down sufficiently with smart blind control.
Swinging Pendulums
  • Ben Miller (bdm23)
  • Andrew Zhou (amzhou)
  • Maryam Nisar (maryamni)
A self balancing ball and beam PID system.

N/A
Self-Driving Model Car
  • Mohammad Tamzeed Moazzem - mmoazzem@buffalo.edu
  • Hasaan Saleh - hasaansa@buffalo.edu
A model car that utilizes PID Control to avoid obstacles and drive in a safe path. We made this possible by utilizing a LiDAR sensor and two ultrasonic sensors for detecting obstacles, a servo motor to control the steering angle, and a DC motor to control the rear wheel drive.

N/A

Software Engineering (CSE 442)

Liquor Looker
  • David Livadhi: davidliv@buffalo.edu
  • Naqi Haider: naquihaid@buffalo.edu
  • Michael Focacci: mcfocacc@buffalo.edu
  • Sean Timmons: smtimmon@buffalo.edu
We have built a web application that allows users to find the cheapest alcohol in their area and also provide directions, the second half of our project is signing up store accounts which can manage their inventory and prices and those would then show up on our map.

signing up by address instead of latitude or longitude
TERMS&Co
  • Tanvie Kirane <tanvieki@buffalo.edu>
  • Rami Khammash <ramikham@buffalo.edu>
  • Michael Geraci <megeraci@buffalo.edu>
  • Snigdha Motadaka <snigdham@buffalo.edu>
  • Elijah Washington <ejwashin@buffalo.edu>
A friendly Discord bot for organizing & scheduling tasks

A Discord bot that enables its users to schedule their daily tasks as well as create individualized channels to sort their tasks & priorities by. Our bot enable users to manipulate their schedule (Create-Read-Update-Delete) using features like add, edit, delete, and mark task as important or completed. The Discord bot notifies the user/s at the time their task is due by sending them a friendly message via the discord channel. A feature we take most pride in is that users can also specify the mood of the notification: casual, motivational, and normal. We have a login feature to redirect the user to their unique webpage (name appears on top right & only their tasks show up) where they can clearly see their tasks in different sections namely “in progress”, ”completed” & “overdue”.
Mosaic Maker/ Team Arson
  • Jia Hou - jhhou
  • Helen Luo - helenluo
  • Alex Yan - alexyan
  • Liam Mullen - liammull
Generate mosaicked images for the user without the need for learning or knowledge of photo-editing skills

Potentially other types of image editing: black and white, photoshop
CodeHub/Bees
  • Anthony Mendez
  • amendez8@buffalo.edu
  • Matthew Johnson
  • mrj26@buffalo.edu
  • Ronald Chen
  • rchen56@buffalo.edu
  • Senhuang Cai
  • senhuang@buffalo.edu
CodeHub's objective provides a tool that help CS students, professors, software engineers to have a easier time understanding/presenting code.

Our motivation comes from our frustration of understanding code from a big project and want to create a solution for simplify code for a wide audience.

Ethan Blanton (CSE 700)

GEP interactive musical devices
  • Matias Homar
  • matiasho
GEP
Is a project based on interactive devices that seek to enhance the creating, exploring and/or experience sound.
It is anchored on the idea of ‘Enactivism’ which, simply put, states that perception is constituted from the interactions between an entity and the environment. Both, entities and environments, mutually influence each other in a constant changing system. In this sense, GEP is aimed to enhance the experience of becoming aware of the constantly changing relations we establish with our environment.
Thus its potential to serve for artistic but also educational purposes as well. Each of the devices act as an extension of the person interacting by reading motions form the body or parts of the body. In turn, these readings will modify, create or change the sounds from the environment where the actions are taking place. Potentially speaking, this can entail learning processes that can go from learning to use the devices to create them and code the program that will ultimately define what each device will do.
The nature of the project is of being Open Source. This means, that the devices themselves can be altered, reshape, refactor, and so on. Therefore, they can be adapted for the needs of specific individuals or communities. Because they are MIDI devices they can be hooked up to any computer that is capable to run any MIDI compatible software. This gives a wide range of possibilities from paid to free applications available to download from the internet.
The project is based on the use of open source hardware, essentially Arduino and clone types of microcontrollers. This allows for affordable and accessible learning experiences. These microcontrollers open the way to include coding and hardware design as part of an artistic project as well as for educational curricula.

GEP Pedal (interactive shoe)
Wearable devices that reacts to different foot motions. It is built with 4 different sensors (light, pressure, accelerometer and distance). Each of this sensors can be used separately or combined creating a diverse palette of sound shaping effects. Given that it uses MIDI communication to the computer, it can be assigned to any parameter that is MIDI compatible in the host application.
GEP Tebukuro (interactive glove)
Similar to the shoe, a wearable device still under development. It consists of flex sensors that read the motions from three fingers, an accelerometer and a touch sensitive fader. As with the shoe, these sensors can be combined or used separately. Also using MIDI protocol, it can be connected to any application capable of reading MIDI messages.
GEP Sensing Motion Module
A modular arrangement of sensors. Each module consists of 6 distance sensors that work as three different units within the module. Each of these unites give a unique MIDI message that can be assigned to different variables. This sensing module was designed with two main purposes: a) as a module for art productions including (but not limited to) dancers that can modify the sounds from any sound sources (instruments, recordings, environmental music, etc.), and b) for educational purposes. For the latter, it entails a type of interaction that makes necessary to be aware of others and what they are doing in terms to coordinate, collaborate and engage in a communal bond. Within a classroom environment it can signify the breakdown
Matias Homar
of barriers allowing the differences to arise in a positive and constructive way. Each person involve can truly contribute with their uniqueness. And from this divers universe of possibilities it can emerge a representative work that is based in our shared humanity.

DRONES Lab Research Demonstration

DRONES Lab Space Vision Research
  • Timothy Chase - tbchase@buffalo.edu
  • Sannihith Kilaru - sannihit@buffalo.edu
  • Shivendra Srinivas - sriniva7@buffalo.edu
DRONES Lab graduate students will present their work on visual perception for space robotics applications.

N/A

Reinforcement Learning Project (CSE 546)

Solving Flatland Env using Parallel RL Techniques
  • Pragati Nagar (pragatin@buffalo.edu)
  • Aryan Saini (asaini2@buffalo.edu)
  • Divyansh Chopra (dchopra2@buffalo.edu)
Implementing Parallel DQN to solve a complex environment. Our environment is Flatland Environment.

How to efficiently manage dense traffic on complex railway networks? We aim to solve this by the use of Parallel DQN algorithm.
Fairness in Multi Agent Reinforcement Learning
  • Nandini Chinta - nandinic@buffalo.edu
Exploring how fairness can be applied to Multi Agents systems using Deep RL. Applied scenarios of 1. Resource sharing, 2. Time limit, 3. Individual Priorities on Agents

Using Deep Reinforcement Learning algorithms: Q-Learning, Deep Q-networks and Double Deep Q-networks to solve the fairness among agents in Multi Agents systems.
Stock Trading using RL (Team 33 ub learns))
  • Naman Tejaswi namantej@buffalo.edu
  • Arshabh Semwal arshabhs@buffalo.edu
We have used DDPG and TD3 with the objective to maximize profits trading stocks for a period of 100 days.

Our long term objective is to create a portfolio optimization tool where one can hedge their position using options and ride out volatile periods in the stock market, we currently have positive returns over 100 trading days and we will be further refining our DDPG and TD3 models and try out other reinforcement learning models. We also will be extending this to option trading.
Team 32
  • Prajvala Sonawane (prajvala@buffalo.edu)
  • Nikita Desai (ndesai2@buffalo.edu)
  • Baasit Sharief (baasitsh@buffalo.edu)
Empathetic Dialogue Generation using Deep Reinforcement Learning

Empathetic dialogue generation in chatbot using Offline Reinforcement Learning and Proximal Policy Optimization models.

MAE 600 / CSE 610 course project

Vertiport-1
  • Jhoel Witter (jhoelwit@buffalo.edu)
  • Prajit Krishna Kumar (prajitkr@buffalo.edu)
Our group is tasked with creating a reinforcement learning model for an air traffic control agent to allocate tasks to un-manned aerial vehicles, while navigating a terrain riddled with uncertainties.

This project started as a course endeavor for MAE 600 / CSE 610, and will likely go on to compete at conferences later this year, notably DARS and Scitech.

Independent Study

Bio-inspired Robotic Oysters
  • chetan palkar cpalkar@buffalo.edu
  • Ralfy Chettiar ralfyfra@buffalo.edu
  • Stephanie Rothenberg sjr6@buffalo.edu
  • Karthik Dantu kdantu@buffalo.edu
“Tending Ostreidae: Serenades for Settling” is multimedia operatic installation that examines how oyster settlement patterns are impacted by anthropogenic noise. The project focuses on the oyster reseeding initiatives in the waterways of New York City’s harbor — once the largest harvesting region for oysters in the 1800’s. Sound responsive robotic oysters created at different stages of their life cycle (larvae, pediveliger, reproductive adult, etc) will respond to sonic data collected from the harbor.

In this project, we designed the actuation mechanism for a 3-D printed Oyster to open and close, and a mechanism to spew water. The oyster sits in water, listens to ambient sounds and spews water based on the degree of anthropogenic noise heard. Future plans include developing a mechanism for the oyster to walk on the water bed.

This work is joint between Prof. Karthik Dantu in CSE, Prof. Stephanie Rothenberg in Dept of Art and Suzanne Thorpe from Columbia University/Billion Oyster Project
Workout Tracker
  • Jean Vigroux - jeanvigr@buffalo.edu
A simple and easy to use mobile application to track your weight training progress. This app can track weight and repetitions to show you how well you are performing at the gym.

(CSE 446)

AWS DeepRacer
  • jaskara2@buffalo.edu
  • mukeshga@buffalo.edu
AWS deepracer project using RL

https://aws.amazon.com/deepracer/

Past Demo Days

Fall 2016 Fall 2017 Fall 2018 Fall 2019 Spring 2019 Fall 2020 Fall 2021 Spring 2021 Fall 2022 Spring 2022 Fall 2023 Spring 2023 Spring 2024