9:00 - 9:15
Ovidiu Calin, faculty, EMU: A welcome message and a brief introduction to Machine Learning
Abstract: This is a short message for the students who would like to pursue a career in a field related to machine learning and data science. A few covered topics are: Why the need of machine learning? Why is ML that special? The future of AI.

9:20 - 9:55
Khairul Islam, faculty EMU: Unsupervised and Supervised Learning with Applications
Abstract: Machine learning refers to learning from data and making an improved prediction without being explicitly programmed. This presentation provides some applications of unsupervised and supervised machine learning approaches with implementation in R. Finally, we seek to make statistical inferences in relation to the learning of how certain factors affect an outcome (e.g., heavy drinking, abuses, etc.) using observed survey data.

10:00 - 10:30
Karthik Desingh: University of Michigan: Robots working in human environments
Abstract:   Robots working in human environments often encounter a wide range of articulated objects, such as tools, cabinets, and other jointed objects. Such articulated objects can take an infinite number of possible poses, as a point in a potentially high-dimensional continuous space. A robot must perceive this continuous pose to manipulate the object to a desired pose. This problem of perception and manipulation of articulated objects remains a challenge due to its high dimensionality and multi-modal uncertainty. Here, we describe a factored approach to estimate and track the poses of articulated objects using an efficient approach to nonparametric belief propagation. We consider inputs as geometrical models with articulation constraints, and observed RGBD sensor data. The described framework produces object-part pose beliefs iteratively. The problem is formulated as a pairwise Markov Random Field (MRF), where each hidden node (continuous pose variable) is an observed object-part's pose and the edges denote the articulation constraints between the parts. We describe articulated pose estimation and tracking by Pull Message Passing algorithm for Nonparametric Belief Propagation (PMPNBP) and evaluate its convergence properties over scenes with articulated objects. Robot experiments demonstrate the necessity of maintaining beliefs to perform goal driven manipulation tasks.

10:30 - 11:00
Keith Lambert: Neuroscience and AI, what’s the connection?
Abstract: In this presentation we will look at the historical connection between neuroscinece and AI, how the fields have lined up and differed, and what the now two distinct fields can learn from each other going forward.

11:00 - 11:35
Ovidiu Calin, faculty, EMU: Informational Aspects of Supervised and Unsupervised Learning of Images.
Abstract: An explanation of why about 2% efficiency is lost when a picture is read line by line by a FFN, instead of being read as a 2-dimensional object, as CNN do.
Pictures are stored by projecting them into a smaller dimensional space such that the lost information is minimized. We show that this space is of negative curvature when dealing with human faces.

13:15 - 13:55
Muhammad Sohaib Arif, student, EMU: Farmaid Bot
Abstract: A robotic platform for disease detection in greenhouse environments. Winner of Best Use of AI and Most Fun Social Media Video in Arm Autonomous Vehicle Competition 2018 https://www.hackster.io/contests/arm2018 Winner of Hack Harvard 2018 Most Fundable Hack https://devpost.com/software/farmaid-jr

14:20 - 14:50
Bader Al-Shamary, faculty Kuwait University: Applications of neural networks to domino portrait problem.
Abstract: Domino portrait problem involves arranging complete sets of dominos to resemble photographic portraits when seen from a distance. We use a local search algorithm to solve the problem. The cost function was modified so that important positions in the portrait such as facial features were emphasized, thus improving the results.