Research

Research

Enabling Smart Applications at Scale for the Intelligent Edge Network

NeuTigers is an Edge-AI startup that helps businesses struggling to bring intelligent services close to the data sources and to the point of use.

We’re commercializing exclusive IPs algorithms from the Princeton University Department of Electrical and Computing Engineering.

NeuTigers is developing portable end-to-end machine learning applications based on uniquely compact, accurate, energy and latency-efficient neural network models that can be embedded in any resource-constrained connected objects.

We’re productizing the technology in our primary medical-healthcare market. The company is making strides to transform healthcare and research through remote patient monitoring, collection of real-world evidence, and advanced Edge-AI analytics to predict chronic and acute medical conditions and enable early intervention.

We believe that Edge-AI will redefine healthcare delivery and consumer wellbeing, across a range of existing and future devices and technologies.

Beyond our core business in Healthcare, our edge AI technologies are being deployed in industrial IoT and Cybersecurity.

Edge-AI Algorithms

Algorithms for Efficient Neural Networks (DNN & CNN)

NeST

NeST

NeST is a DNN synthesis tool to automate the generation of compact and accurate DNNs

ChamNet

ChamNet

Algorithms for Efficient Neural Networks (DNN & CNN)

LSTM

LSTM

Algorithms for Efficient Long Short-Term Memories

SCANN

SCANN

Synthesis of Compact and Accurate Neural Networks

Synthetic Data

TUTOR

TUTOR

Reduces the need for labeled data by 5.9x, improves accuracy by 3.4%, needs fewer samples than GANs and optimize existing DNN

CTRL: Clustering Training Losses for Label Error Detection

CTRL: Clustering Training Losses for Label Error Detection

CTRL is an effective framework for detecting noisy labels. It relies on the observation that training progresses differently for clean and noisy labels. It uses the K-means algorithm to classify labels as clean or noisy by clustering their training loss curves.

Cybersecurity

SHARKS

SHARKS

SHARKS uses ML to detect new exploits based on known attacks – software, hardware or network on CPS/IoT systems (healthcare devices and wearables, critical infrastructures, e.g., nuclear power plants, autonomous vehicles, smart cities, and smart homes)

GRAVITAS

GRAVITAS

GRAVITAS presents a defense model to lower vulnerability within a given budget

Applications

Healthcare

Trevor: Digital Companion for Sickle Cell Disease Patients

Trevor: Digital Companion for Sickle Cell Disease Patients

Toward a Conversational Agent to Support the Self-Management of Adults and Young Adults With Sickle Cell Disease: Usability and Usefulness Study

MHDeep

MHDeep

Mental Health Disorder Detection System based on Body-Area and Deep Neural Networks

DiabDeep

DiabDeep

DiabDeep: Pervasive Diabetes Diagnoses based on Wearable Medical Sensors and Efficient Neural Networks

CovidDeep

CovidDeep

SARS-CoV-2/COVID-19 Test Based on Wearable Medical Sensors and Efficient Neural Networks

YSUY

YSUY

ML model to infer physical, mental and emotional states