Neural networks are notoriously overconfident in their predictions. We use architecture search to find a distribution of architectures and use it to construct a Bayesian ensemble for outlier detection
Testing image classification accuracy and considering pros and cons between Resnet 101 and Alexnet.
Intuitive modelling of supply side chains with exploratory neural networks
Fund management strategy that utilizes various datasets, including supply chains, social media data, news data, stock market time series data and others.
Detecting Alpha Signals with sentiment analysis and machine learning
Online conversational AI platform that provides virtual patients for the deliberate practice of medical students
The Deep Money Model determines recommendations to buy, sell, or hold volatile stocks to make big gains quickly while avoiding big losses.
Pigment Match aims to use deeplearning, AR and a mobile application to modernize the ways of matching foundation in the 532 billion dollar makeup industry.
crowd-sourced & encrypted patient diagnosis
RESBERT is a state of the art extractive document summarization AI, with a specially fine tuned transformer architecture for real estate news articles.
Have you ever thought about being the main character of your game? With our app, you can take yourself a photo and have a 8 bit character in only a few seconds with your facial attributes.
Is it possible to detect and classify toxic comments without the presence of a human agent and act accordingly? This project tries to help by automating this.
Creating take-home physical therapy with computer vision, using devices you already have
Track victory royales automatically in a fortnite stream
text to speech in anyone's voice with mere seconds of audio
Construct a virtual assistant that has a visual, audio, and text interface. It will be able to: 1) hold a conversation 2) give banking information 3) give navigation directions
We help owners find their lost dogs using machine learning with crowdsourced data.
Our algorithm looks at 4 components that could affect the solar energy stock price: substitutes cost, material cost, green spirit indicator and how well the solar industry is doing as a whole.
Using AI to detect a defective pilot
Since waste management is at the forefront of today's environment sustainability, we decided to create an application that facilitates the user’s waste disposal by identifying the proper category.
Play generated jazzy tunes with your hands
Getting white blood cell counts is an important but tedious part of hematology. We aim to automate this process with recent advancements in machine learning.
Providing a clean and safe internet environment for individual users and corporate workspaces
Can an algorithm diagnose Alzheimer's disease based on patients' MRIs? We trained a 3D CNN with 1500 brain MRIs of three different stages. The predictions are shown on 3D superimposing the actual MRIs