Mihajlo Grbovic, PhD

 

 

Mihajlo Grbovic, PhD [resume]
Research Scientist
Yahoo! Labs                             
Advertising Sciences


Contact: mihajlo at yahoo-inc dot com



View Mihajlo Grbovic's profile on LinkedIn

 

Research

Machine Learning, Data Mining and Predictive Analytics

  • Computational Advertising [1, 2]
  • Memory Constrained Data Mining [8, 18, 19, 20]
  • Constrained Convex Optimization [9, 10]
  • Fault Detection and Diagnosis [7, 9, 14, 16, 17]
  • Preference Learning and Ranking [3, 4, 11, 12]
  • Online Learning of Large Streaming Data with Concept Change [15]
  • Data Compression using Decentralized Estimation [8, 19]
  • Traffic (Machine Learning applications in parking) [13, A*]

 

Education

 

Teaching

Teaching Instructor for:

  • Network Architectures (CIS 4329), Temple University, Spring 2008, Fall 2008, Spring 2009, Fall 2011 [grades]
  • Intro to Computer and Network Security (CIS 4378), Temple University, Spring 2009
  • Networks and Communications (CIS 4319), Temple University, Spring 2008, Fall 2008

 

Publications


[1] Grbovic M., Vucetic S., Generating Ad Targeting Rules using Sparse Principal Component Analysis with Constraints, International World Wide Web Conference (WWW), 2014 [pdf]


[2] Djuric N., Grbovic M., Radosavljavic V., Bhamidipati, N., Vucetic S., Non-linear Label Ranking for Large-scale Prediction of Long-Term User Interests, AAAI Conference on Artificial Intelligence (AAAI), 2014 [pdf]

 

[3] Grbovic M.*, Djuric N.*, Vucetic S., Multi-prototype Label Ranking with Novel Pairwise to Total Rank Aggregation, International Joint Conference on Artificial Intelligence (IJCAI), 2013 [authors contributed equally] [pdf]


[4] Grbovic M., Djuric N., Guo S., Vucetic S., Supervised Clustering of Label Ranking Data using Label Preference Information, Machine Learning Journal (MLJ), 2013

 

[5] Grbovic M., Malkin J., Das H., Large scale Ad Latency Analysis IEEE International Conference on Big Data (BigData), 2013. [pdf]

 

[6] Djuric N. Grbovic M., Vucetic S., Distributed Confidence-Weighted Classification on MapReduce, IEEE International Conference on Big Data (BigData), 2013. [pdf]

 

[7] Grbovic M., Li W., Subrahmanya N. A., Usadi A. K., Vucetic S., Cold Start Approach for Data Driven Fault Detection, IEEE Transactions on Industrial Informatics, 2013. [pdf]

 

[8] Grbovic M., Vucetic S., Decentralized Estimation using Distortion Sensitive Learning Vector Quantization, Pattern Recognition Letters 2013. [pdf]

 

[9] Grbovic M., Dance. C., Vucetic S., Sparse Principal Component Analysis with Constraints, AAAI Conference on Artificial Intelligence (AAAI), 2012. [pdf]

 

[10] Djuric. N., Grbovic M., Vucetic S., Convex Kernelized Sorting, AAAI Conference on Artificial Intelligence (AAAI), 2012. [pdf] [code]

 

[11]  Grbovic M., Djuric N., Vucetic S., Supervised Clustering of Label Ranking Data, SIAM Conference on Data Mining (SDM), 2012. [pdf] [slides]

 

[12] Grbovic M., Djuric N., Vucetic S., Learning from Pairwise Preference Data using Gaussian Mixture Model, Preference Learning Workshop, European Conference on Artificial Intelligence, (ECAI), 2012. [pdf]

 

[13] Zoeter O., Dance C., Grbovic M., Guo S., Bouchard G., A General Noise Resolution Model for Parking Occupancy Sensors, Intelligent Transportation Systems World Congress, 2012. [pdf]

 

[14] Grbovic M., Li W., Peng X., Usadi A. K., Vucetic S., Decentralized Fault Detection and Diagnosis via Sparse PCA based Decomposition and Maximum Entropy Decision Fusion, Journal of Process Control, 2012. In press [pdf]

 

[15]  Grbovic M., Vucetic S., Tracking Concept Change using Incremental Boosting by Minimization of the Evolving Exponential Loss, The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD), 2011. [pdf] [video] [code]

 

[16]  Grbovic M., Li W., Peng X., Usadi A. K., Vucetic S., A Boosting Method for Process Fault Detection with Detection Delay Reduction and Label Denoising, KDD Workshop on Data Mining for Service and Maintenance, The 17th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (ACM SIGKDD), 2011. [pdf]

 

[17]  Grbovic M., Vucetic S., Statistical and Machine Learning Techniques for Fault Detection and Fault Classification in Industrial Systems, ExxonMobil Internal Technical Report, 2010.

 

[18]  Grbovic M., Vucetic S., Regression Learning Vector Quantization, International Conference on Data Mining (ICDM), Miami, FL, 2009. [pdf]

 

[19] Grbovic M., Vucetic S., Decentralized Estimation using Learning Vector Quantization, Data Compression Conference (DCC), Snowbird, UT, 2009. [pdf]

 

[20]  Grbovic M., Vucetic S., Learning Vector Quantization with Adaptive Prototype Addition and Removal, International Joint Conference on Neural Networks, (IJCNN) Atlanta, GA, 2009. [pdf] [code]

 

Patents


 [A*] Grbovic M., Zoeter O., Dance C., Guo S., Bouchard G., A model to use data streams of occupancy that are susceptible to missing data, US patent pending

 

Reviewer and PC member

 

1.     Transactions on Pattern Recognition and Machine Intelligence (TPAMI)

2.     SIAM International Conference on Data Mining (SDM) 2013

3.     The Conference on Uncertainty in Artificial Intelligence (UAI) 2013, 2014

4.     Neural Information Processing Systems (NIPS) 2013, 2014

5.     International Conference on Machine Learning (ICML) 2013, 2014

6.     AAAI Conference on Artificial Intelligence (AAAI) 2014

7.     Journal of Machine Learning Research (JMLR) 2014

8.     International Conference on Artificial Intelligence and Statistics (AISTATS) 2014

9.     Neurocomputing

10.     Pattern Recognition Letters

 

Work Experience

 

1.     September 2012 – present. Research Scientist at Yahoo! Labs, Sunnyvale, CA.

 

AREA: Computational advertising. Ads targeting, Look-alike modeling.

PROJECT 1: Large-scale behavioral targeting and look-alike modeling on 700M Yahoo! users

PROJECT 2: Outlier detection (with applications to Ad latency)

PROJECT 3: Multi-class email classification (automatic categorization of emails into categories: shopping, travel, financial, etc.)

 

2.     June 2012 – September 2012. Research Scientist Intern at Akamai Technologies, Cambridge, MA.

 

AREA: Computational advertising. Ads modeling. Applying machine learning to online advertisement

PROJECT: Feature Selection for Display Advertising Purchase Prediction

 

3.     February 2012 – May 2012. Research Intern at ExxonMobil Research and Engineering, Annandale, NJ.

 

PROJECT: Working on clustering and segmentation of large-scale data. Concentrating on applications of Sparse Principal Component Analysis

in specialized clustering and segmentation algorithms

 

4.     May 2011 – October 2011. Research Scientist Intern at Xerox Research Center Europe, Grenoble, France

 

PROJECT: “Modeling Demand for on Street Parking” [link] [sfpark] [won best intern award]

 

[A*] A model to use data streams of occupancy that are susceptible to missing data, US patent pending

 

5.     May 2004 – October 2004. Research Intern at the Signal and Image Processing Lab, Department of Electrical Engineering, Technion, Israel Institute of Technology, Haifa, Israel

 

PROJECT:  “Image Identification using Hashing” [link]

 

 

Friends and Colleagues

 

Slobodan Vucetic

Zoran Obradovic

Nemanja Djuric

Vladan Radosavljevic

Kosta Ristovski

Vuk Malbasa

Vladimir Coric

Nebojsa Radovic

Zhuang Wang

Liang Lan

Shengbo Guo

Guillaume Bouchard

Zoeter Onno

Christopher R. Dance

Neda Salamati

Rosie Jones

Ana Milicic

Selen Uguroglu

Nenad Bozinovic

Weichang Li

Danilo Djordjevic

Prakash Comar

Dmitry Pechyony

Dilan Gorur

Dusan Ramljak