Agenda:
12:00 – 12:20 pm – Welcome & Meeting Overview
Welcome – Prof. Dimitris Metaxas, Center Director
State of CARTA – Prof. Dimitris Metaxas, Center Director
12:20 – 12:30 pm – NSF Overview-Ann Chrstine Von Lehmen
12:30 – 12:40 pm – New CARTA Site – Arizona State University
12:40 – 1:45 pm – New Proposals
At the of each presentation the CARTA Industry Members insert their comments
Collaborative Projects (6 minutes each):
Project 1: Ethical and Secure Medical Analytics, PIs: Drs. Karuna Joshi (UMBC), David Chapman (UMiami), Dimitris Metaxas (Rutgers)
Project 2: A Data Driven Deep learning OSSE, PIs Dr. Milt Halem (UMBC), Dr. Ben Kirtman (UMiami)
Project 3: Histopathology Analytics, PIs: Mitsunori Ogihara, Dimitris Mexatas, Yelena Yesha (UMiami)
UMBC – Status by UMBC Site Director, Prof. Karuna Joshi (5 minutes each)
Searchable Malware Tensors, (PI) Dr. Charles Nicholas
Veracity Measure to Manage Disinformation Propagation, (PI) Dr. Karuna Joshi
Trusted data model in edge computing environment using permissioned blockchain, (PI) Dr. Yaacov Yesha
Using dynamic neural networks for an effective dynamic firewall, (PI) Dr. Yaacov Yesha
NC State – Status by NCSU Site Director, Prof. Rada Chirkova (5 minutes each)
Learning from Mistakes: Real-time On-board CNN Training for Autonomous Driving, Prof. Frank Mueller
Privacy-preserving Behavioral Analysis of IoT devices on Network Edge, Dr. Anupam Das
Performance Analysis from Edges to Cloud, Prof. Xu Liu
Forecasting Anomalous Atlantic Hurricane Seasons Using Explainable AI, Dr. Lian Xie
Realtime Position, Navigation and Timing with AI Support, Prof. Khaled Harfoush
University of Miami – Status by Miami Site Director, Prof. Mitsunori Ogihara (5 minutes each)
Comprehensive Testbed for Clinical Flow – PI: Dr. Yelena Yesha, Dr. Mitsunori Ogihara
1:45 – 1:50 pm – Break
1:50 – 2:30 pm – Ongoing Projects: Lightning Updates (40min)
Each project will have 2 minutes for a student researcher to present key highlights of their existing project, specifically 1) the most important methodological problem the project seeks to address; and 2) the most important thing that has been learned about that problem.
UMBC
(Ongoing) Automatic Legal Document Analytics and Blockchain Smart Contracts, Dr. Karuna Joshi (PI), Dae-young Leroy Kim (Student)
(Ongoing) Blockchain for Supply Chain Asset Management and Data Security, Prof. Yaacov Yesha (PI), Yusen Wu, Lawrence Sebald (Students)
(Ongoing) Towards The Classification PA Chest Radiographs as Normal or Abnormal, Prof. David Chapman (PI), Sumeet Menon (Student)
Rutgers
(Ongoing Project) Cardiac Analytics: Dyssynchrony and CRT Outcomes – Meng Ye, Qi Chang
(Ongoing Project) Prediction of Liver Fibrosis Score, Progression and Clinical Deterioration using Multi-Modality Data – Anaya
(Ongoing Project) Predicting Prognostic Factors and Outcomes of COVID-19 Patients using Machine Learning Methods – Anaya
Exploiting Unlabeled Data with Vision and Language Models for Object Detection – Shiyu Zhao
NC State
(Active project) Fingerprinting IoT Devices at Real Time using Encrypted Traffic, Dr. Anupam Das
(Active project) Resource Protection in Cloud Data Centers Through Multi-Point Traffic Shaping, Dr. Muhammad Shahzad
University of Miami
Music-induced Brain Activities – Combining Naturalistic Music Listening, Neuroelectric Recordings, and Machine Learning for Conscious – PIs: Dr. Andrew Dykstra, Dr. Mitsunori Ogihara
Trends and patterns of Benzodiazepines: an overdose risk prediction model – PIs: Dr. Yelena Yesha, Dr. Smriti Prathapan
Genotype-aware Consumer Product Recommendation – PI: Dr. Vanessa Aguiar-Pulido
Cyber-Physical Security Analytics – PI: Steve Dinnis
2:30 – 2:35 pm – Break
2:35 – 3:10 pm – Research to Application: Final Reports – Take Home Points (5 minutes)
(Ending) Taking the Thunder Out of Zeus, Dr. Charles Nicholas (PI), Maksim Eren (Student)
(Ending) Multi-Domain Attack Detection for Cybersecurity, Prof. Yaacov Yesha (PI), Sarang Patil (Student)
(Ending) Evaluating and Eliminating Bias in AI for Medical Imaging, Prof. David Chapman (PI), Sourajit Saha, Nishanjan Ravin (Students)
(Ending) Progressive Neural Nets for Emulator Training of Climate Model Sub-grid Scale Physics Parametrization, Prof. Milt Halem (PI), Qingyuan Zheng (Student)
(Completed Project) Real-time Monitoring the Evolution of the Web, Dr. Alexandros Kapravelos
3:10 – 3:15 pm – Break
3:15 – 4:00 pm – Wicked Problems Panel – Industry Members
4:00 – 4:10 pm – NSF Survey Completion: IAB Reps, Faculty, and Students
4:10 – 4:40 pm – Guest Speaker: Mu Zhou
Title: Towards AI-driven Pathological Integration: Feature extraction, Outcome Prediction, and Annotation Efficiency
Abstract:
The rapid development of deep learning has made remarkable progress in healthcare applications. Yet we continue to face a plethora of medical data challenges from data scale, annotation cost, and outcome predictions. To advance clinical decision making, in this talk, I will highlight the recent surge of pathology-based AI analytics for predicting multiple cancer molecular outcomes. First, I will share our research progress on graph-based deep learning classifiers to predict genetic mutations in cancer. Second, I will present recent deep-learning works on linking whole slide images and important mutational outcomes in cancer and its validations across cancer types. We provide an image-to-genomics pipeline to allow a head-to-head comparison between mutations and biological pathway signals by leveraging public cohorts. Finally, ongoing data annotation challenges, limitations, and emerging industrial opportunities will be discussed in related areas.
Brief bio:
Dr. Mu Zhou is broadly interested in machine learning, medical image analysis, and bioinformatics. Dr. Zhou serves as head for AI drug discovery at SenseBrain. He also holds a visiting professor position at Rutgers University, New Jersey. His research focuses on data-centered approaches to analyze and process quantitative information from multi-scale biomedical data across radiology, histopathology, and omics profiles in oncology. He was a research scientist and a postdoctoral fellow at medical school, Stanford University. In collaboration with his colleagues, he led the research for linking cancer imaging and high-throughput RNA expression in lung cancer. He received his Ph.D. degree in computer science and engineering from University of South Florida, Tampa, where he pioneered image-based analysis for non-invasive outcome prediction of cancer patients.
4:40 – 5:10 pm – Closed IAB meeting
5:10 – 5:20 pm – Closing Remarks & Adjournment
Pingback: CARTA Spring 2022 IAB Meeting – CARTA