o TLM

o Driver Performance Monitoring: Systems for large scale data processing, regression detection for both the real and virtual worlds. Handles data streams at various latencies.

o Evaluation: Transfer Learning, Dataset Fitting, Experiment Framework in Eval

o Smart Agents in Simulation

• Large scale data processing.

• Machine learning & data analysis.

• Ads Quality

• -- Develop quality metrics

• -- Quality scoring

• -- Quality signal engineering

• -- User signals

• Dialogue management (previous).

Team leader for a cloud based retinal image analysis system. Led conceptualization of system architecture and design. Researcher, lead developer, and manager of multiple systems for automated retinal image analysis for clinical and research studies. Collaborated and analyzed requirements with researchers, ophthalmologists, and retinal grading technicians from Singapore and Australia. Work contributed to the Singapore President's Technology Award (2014) received by the team.

Projects include:

1. Singapore Eye Vessel Assessment (SIVA) System – extracts multiple vascular measurements from retinal images

• - Developed system from prototype to commercially marketable quality.

• - Actively used for multiple medical studies involving thousands of patients. In use worldwide.

• - Designed methods to simplify user interaction, reducing grading time per image from 2 hours to < 15 minutes.

• - Introduced parallel algorithms that sped up processing by about 3x.

• - Created a new method for more accurate vessel separation that increased correlation of automated measurements with gold standard manual measurements by up to 2.75x.

• - Designed system as a modular subsystem library for use with other systems

2. Fractal Analyzer – extracts the fractal dimension of vasculature from retinal images for further study

3. OCT Analyzer – quantifies cataract from OCT images using template fitting to mark regions for measurements

Manages the subversion server and introduces new developers in the research group to coding guidelines.

Full publications list at: https://sites.google.com/site/peterlauqf

Projects in addition to the retinal imaging work carried forward as Research Associate include:

1. Deep Detect System – finds attribute incongruities (outliers) from discrete and continuous XML data

2. PICAR System – predicts cerebral aneurysm rupture from patient records.

• - Collaborated with and deployed at the Neurosurgical Dept. Huashan Hospital, Fudan University Shanghai, China

• - Designed a methodology to pick the best ensemble classifier combination for the task