Iyad Batal is a research scientist at Yahoo Labs with expertise in machine learning, data mining, and signal processing. He holds a PhD in Machine Learning from the University of Pittsburgh and has worked at GE Research and Yahoo. Batal has published extensively in top conferences and journals, and received several awards and honors for his graduate research work.

1. Curriculum Vitae
Iyad Batal
PERSONAL
Email: iyadbatal@gmail.com
Phone: 412-519-7074
Address: 350 Elan Village Ln, San Jose, CA 95134
Website: https://sites.google.com/site/iyadbatal/
U.S. Citizen
EDUCATION
PhD in Machine Learning (2012)
University of Pittsburgh, Computer Science department (GPA: 3.93). Thesis title:
Mining Predictive Patterns and Extension to Multivariate Temporal Data
BS / MS in Computer Science (2005)
University of Damascus. Thesis title: Predicting Stock Prices using Technical
Analysis
AREAS
OF
EXPERTISE
• Design and build large-scale machine learning applications to solve real-world
problems
• Machine Learning: Supervised learning, probabilistic graphical models, structured
prediction, convex optimization, Bayesian nonparametric models, dimensionality
reduction, ensemble learning, and Bayesian inference
• Data Mining: Temporal pattern mining, subgroup discovery, association rule mining,
and text analysis
• Signal processing and time series analysis

2. WORK
EXPERIENCE
• Research Scientist at Yahoo labs, Ranking Science team (09/2015 – present)
• Machine Learning Researcher at GE global research (09/2013 – 09/2015)
• Postdoctoral Associate at University of Pittsburgh (10/2012 – 09/2013)
• Intern at Siemens Corporate Research (summer of 2011)
• Intern at Yahoo (summer of 2010)
• Research Assistant at University of Pittsburgh (09/2008 – 10/2012)
• Teaching Fellow (Instructor) at University of Pittsburgh (09/2006 – 09/2008)
PROGRAMMING
SKILLS
• Languages: Python, Java, C#, C++, SQL
• Numerical computing: Matlab, R
• Big data processing: Hadoop, Pig, Hive
SELECTED
PROJECTS
• Query spelling correction (Yahoo labs): Worked on building a machine learning
to rank framework for web query spelling correction, which outperformed the
traditional noisy channel model.
• Reciprocating engine mechanical health (GE Research): Led the science team on
a project for analyzing signals obtained from vibration sensors for power generation
engines. The goal is to understand these high-frequency signals and utilizing that for
assessing the engine’s performance and detecting mechanical faults. I filed two
patents as a lead inventor and my algorithms were implemented in production.
• Surrogate model-based optimization (GE Research): Proposed and implemented
a Bayesian exploration-exploitation optimization methodology for optimization of
oil field operations.
• Multi-dimensional classification (U. PITT): Proposed and developed several
probabilistic graphical models to learn prediction models when data instances have

3. vectors of output variables (class labels). I proposed three novel methodologies
based on conditional tree-structured Bayesian networks, conditional random fields
and the mixture of adaptive probabilistic experts.
• Mining predictive temporal patterns (U. PITT): The goal of this project is
analyzing electronic medical records data for early detection and identification of
adverse medical events. I proposed and developed algorithms for efficiently finding
predicting temporal patterns and demonstrated their effectiveness on real-world
clinical problems.
• Subgroup discovery (U. PITT): This project is about mining a concise set of
patterns (rules) that are most important for predicting and explaining an output
variable. I proposed algorithms for efficiently searching and scoring patterns in order
to mine patterns that are most predictive and have low redundancy.
PUBLICATIONS
Book Chapters
I. Batal. Temporal Data Mining for Healthcare Data, in “Healthcare Data Analytics”,
CRC Press. Editors: C. Reddy and C. Aggrawal, May 2015.
Journals
[3] I. Batal, G. Cooper, D. Fradkin, J. Harrison, F. Moerchen and M. Hauskrecht. Efficient
Pattern Mining for Event Detection in Multivariate Temporal Data. Knowledge and
Information Systems, 2014.
[2] I. Batal, H. Valizadegan, G. Cooper and M. Hauskrecht. A Temporal Pattern Mining
Approach for Classifying Electronic Health Record Data. ACM Transaction on Intelligent
Systems and Technology (TIST), Special Issue on Health Informatics, 2013.
[1] M. Hauskrecht, I. Batal, M. Valko, S. Visweswaran, G. Cooper and G. Clermont.
Outlier-detection for Patient Monitoring and Alerting. Biomedical Informatics, 2013.
Refereed Conferences
[15] C. Hong, I. Batal, and M. Hauskrecht. A Generalized Mixture Framework for Multi-
label Classification. SIAM Data Mining Conference (SDM), 2015.

4. [14] M. Pakdaman, I. Batal, Z. Liu, C. Hong, and M. Hauskrecht.
An Optimization-based Framework to Learn Conditional Random Fields for Multi-label
Classification. SIAM Data Mining Conference (SDM), 2014.
[13] C. Hong, I. Batal, and M. Hauskrecht. A Mixtures-of-Trees Framework for Multi-
Label Classification. ACM Conference on Information and Knowledge Management
(CIKM), 2014.
[12] I. Batal, C. Hong and M. Hauskrecht. An Efficient Probabilistic Framework for Multi-
Dimensional Classification. ACM Conference on Information and Knowledge
Management (CIKM), 2013.
[11] I. Batal, D. Fradkin, J. Harrison, F. Moerchen and M. Hauskrecht. Mining Recent
Temporal Patterns for Event Detection in Multivariate Time Series Data. SIGKDD
Conference on Knowledge Discovery and Data Mining (KDD), 2012.
[10] I. Batal, G. Cooper and M. Hauskrecht. A Bayesian Scoring Technique for Mining
Predictive and Non-Spurious Rules. The European Conference on Machine Learning and
Principles of Knowledge Discovery in Databases (ECML/PKDD), 2012.
[9] I. Batal, H. Valizadegan, G. Cooper and M. Hauskrecht. A Pattern Mining Approach
for Classifying Multivariate Temporal Data. IEEE International Conference on
Bioinformatics and Biomedicine (BIBM), 2011.
[8] I. Batal and M. Hauskrecht. Constructing Classification Features using Minimal
Predictive Patterns. ACM Conference on Information and Knowledge Management
(CIKM), 2010.
[7] I. Batal and M. Hauskrecht. Mining Clinical Data using Minimal Predictive Rules.
American Medical Informatics Association (AMIA), 2010.
[6] M. Hauskrecht, M. Valko, I. Batal, G. Clermont, S. Visweswaran, G. Cooper.
Conditional Outlier Detection for Clinical Alerting. American Medical Informatics
Association (AMIA), 2010. [Homer Warner best paper award].
[5] I. Batal and M. Hauskrecht. A Concise Representation of Association Rules using
Minimal Predictive Rules. The European Conference on Machine Learning and Principles
of Knowledge Discovery in Databases (ECML/PKDD), 2010.
[4] I. Batal and M. Hauskrecht. A Supervised Time Series Feature Extraction Technique
using DCT and DWT. International Conference on Machine Learning and Applications

5. (ICMLA), 2009.
[3] I. Batal, L. Sacchi, R. Bellazzi, and M. Hauskrecht. A Temporal Abstraction
Framework for Classifying Clinical Temporal Data. American Medical Informatics
Association (AMIA), 2009.
[2] I. Batal and M. Hauskrecht. Boosting KNN Text Classification Accuracy by using
Supervised Term Weighting Schemes. ACM Conference on Information and Knowledge
Management (CIKM), 2009.
[1] I. Batal, L. Sacchi, R. Bellazzi, and M. Hauskrecht. Multivariate Time Series
Classification with Temporal Abstractions. The Florida Artificial Intelligence Research
Society Conference (FLAIRS), 2009.
Filed Patents
I. Batal, B. Matthews, J. Bizub. Mining Predictive Frequency Bands for Peak Fire Pressure
Location Estimation using a Knock Sensor.
I. Batal, B. Matthews, J. Bizub. Estimating Engine Pressure Value using a Knock Signal.
AWARDS
AND
HONORS
• Andrew Mellon Fellowship for Best Graduate Student, 2011-2012
• National Library of Medicine Fellowship, 2010-2011
• Runner-up for the Best Graduate Student Research Award, 2011
• Runner-up for the Best Graduate Student Research Award, 2010
ACADEMIC
ACTIVITIES
Program committee member for AAAI (2015). Reviewer: KDD (2015, 2011, 2010),
ICML (2012), AMIA (2011), TKDE, IJCAI (2009), AMIA (2009), MEDINFO (2009)
LANGUAGES
English (fluent), French (working knowledge), Arabic (native language)