SenSE: Artificial Intelligence-enabled Multimodal Stress Sensing for Precision Health

Prof. Zhenan Bao and Prof. Mert Pilanci from Stanford University, and Prof. Pablo Paredes from the University of Maryland

Give freedom to global and client models!

A federated learning approach that combines personalized models for each client with a generalized global model, optimizing hyperparameters and achieving better performance than individually tuned FedAvg and FedProx.

Stress Control Personal Digital Assistant

An actor-based context-aware personal digital assistant capable of detecting stress and providing corresponding recommendations.

Wearable Personal Stress Tracker for Blind and Low Vision People

We developed an accessible wearable personal stress tracker for blind and low vision people, using a glove-based wearable device with galvanic skin response and haptic motor sensors.

SDN Based Prioritized Information Dissemination

We implemented a data collection system using the Software-Defined Networking (SDN) paradigm with the POX controller, also deploying queues in Open vSwitch and setting bandwidth limits for specific flows, thus ensuring automatic prioritization.

Building Search Engine for ICS domain

A Java-based search engine for web page queries using keywords is developed by crawling and indexing 136,604 pages, and a ranking system is implemented integrating TF-IDF, PageRank, and Cosine Similarity to improve search result relevance.

SDN Testbed based on the Raspberry PI

An actual OpenFlow SDN testbed system was implemented, utilizing Raspberry PIs at both controller and switches, and deploying Open vSwitch and FloodLight as an SDN Controller.