Understand interconnected nature of global multi-national companies via their supply chain, product and services competition, co-investments and co-ownerships as well as other dependencies between operations and revenue streams. We would like to consider the way news on any company specifically propagate down the connection graph and impact other businesses that are related in a way that is not necessarily explicit.
A Fall 2018 internship is available in the Eaton lab to work on the development and application of machine learning approaches to historical evolutionary inference. Research will involve learning to use high performance distributed computing infrastructure, performing population genetic simulations, fitting machine learning models, and writing reproducible shareable code. The ideal candidate will have experience and interest in Python coding and a reasonable understanding of linear algebra.
Using machine learning to conduct brain state classification at real-time on EEG/fNIRS/fMRI data.
DNA sequence reads from a community of microbial genomes are currently processed without considering sequence variants. The project involves building a processing pipeline of such billions of short reads, identifying closest strains they might belong to, assembling them into specific clones, calling their variants, and analyzing the dynamic nature of these bacterial strains along sampling points.
Recently Columbia University, Cornell, and NewYork-Presbyterian have agreed to integrate their clinical (healthcare) and business IT systems onto one shared platform called Epic. The motivating factors to move to Epic are to enhance the patient experience, improve and integrate care, and give our physicians an integrated technology platform that supports the mission of an academic medical center. The intern will assist with developing the “operational” analytics capabilities of Columbia University Medical Center including financial, healthcare operations and healthcare quality analytics.
Robotic grasp planning based on raw sensory data is difficult due to occlusion and incomplete scene geometry. Often one sensory modality does not provide enough context to enable reliable planning. A single depth sensor image cannot provide information about occluded regions of an object, and tactile information is incredibly sparse spatially. We are building a Deep Learning CNN that combines both 3D vision and tactile information to perform shape completion of an object seen from a single view only, and plan stable grasps on these completed models.
