Inspired by the recent advances in self-supervised learning methods, the Samsung Project applies the research work SimCLR from the Google Research Team to retinal OCT image dataset. With the help of self-supervised contrastive learning, this deep learning algorithm is able to learn useful representations of the data in the absence of labeled data, and can be further fine tuned to adapt downstream classification tasks. We evaluate the training sample size threshold at which our framework outperforms a comparable supervised-learning architecture.
SpeakUpAi is an end-to-end model for speech enhancement. Users will be able to upload spoken audio recorded on consumer devices (phones, tablets, etc.), and SpeakUpAI will scale up the fidelity until that audio sounds like it was professionally recorded. The SpeakUpAI project team uses deep learning solutions to break up their results into an evaluation and speaker set.
Pulse Audio is a tool providing real-time speech enhancement for noisy calls. Background noise can be distracting, unprofessional, and make understanding difficult. What if we could remove the background noise, enabling you to focus and understand exactly what your teammate is sharing? We're excited to share a deployed deep learning solution for real-time speech enhancement.
Cancer Map assists a pathologist by demarcating cancerous cells within whole slide images of stained lymph node samples. Cancer Map utilizes image segmentation as well as binary classification to aid with the diagnosis, while also exploring the algorithm's ability to transfer to different cancer types as well as any possible enhancement to accuracy by including images created by GANs.
The quality of classroom relationships is one of the strongest predictors of students’ academic success. Our system visualizes engagement metrics for discussion-based classrooms. StudentStats delivers a dashboard that tracks individual students and quantifies their engagement. Post-lesson, the team provides descriptive statistics for each class period, along with automatic article and video recommendations related to the content of the given lesson.
Object detection is a fundamental function of the perception system within autonomous driving. 2D object detection methods do not provide the depth information required for driving tasks such as path planning, collision avoidance, and vehicle maneuvering. 3D object detection methods introduce a third dimension that reveals more detailed object size and location information. By reimplementing recent methods and experimenting with new approaches, the team improves on the state-of-the-art in bounding volume prediction in the self-driving car domain.
AI is transforming the HR industry; more and more HR teams are going to be using AI in their hiring process. For example, if a recruiter searches their candidate pool for people who have "waiter" in their resumes, they will miss people who have "waitress" or "server" in their resume. Job Recommender is a model that inputs job description and outputs the job title in vector space, which allows for discovery of job titles that are similar in function without relying on the spelling of the word.
We think you will be just as impressed as we were with this fantastic group of projects. If you are interested in the FourthBrain program, applications are open our next cohort is starting May 15, 2021. We look forward to seeing what you build!