Mehrnaz Najafi


I'm a first year PhD student in Computer Science at the University of Massachusetts Amherst (please visit my new webpage at or my personal website at I received my Master's degree in Computer Science from University of Toronto. My Master's thesis was "Towards an Algorithm for Property- and Structure- Preserving Slicing of Statecharts".

I got my Master's and Bachelor's degrees in Computer Engineering from Shahid Beheshti University. My Master's thesis was "An Approach to Animate Object-Z Specifications".


Recent Research Projects

Towards an algorithm for property- and structure- preserving slicing of Statecharts
[Joint work with Professor Marsha Chechik]
Software development and maintenance typically includes changes that are driven either from new requirements or bugs encountered during software usage. To introduce changes in the system, developers need to work (write or modify) on a portion of the software with respect to the changes. Slicing is a reduction technique; it can be used to extract one of computations of software which developers are interested in and leave the others behind; so, slicing can be used to extract relevant portion of the software to the changes. As my M.Sc. thesis, I presented a novel structure- and property- preserving algorithm for slicing Statecharts (a visual language for describing state-based models) with respect to a temporal safety property of interest (it is driven from the changes). The impact of my work is that it enables developers to save time and effort by working on a smaller model scale (portion, or slice) instead of the whole model.

[Joint work with Dr. Patricia C. Arocena , Professor Boris Glavic and Professor Renee J. Miller]
The iBench project proposes a new data integration metadata generator that can be used to create realistic synthetic data integration scenarios [1]. iBench's metadata generator is able to create arbitrarily large and complex schemas, mappings and constraints, with the ultimate goal of facilitating the empirical evaluation of data integration solutions. As a graduate student, my role in this project was to design and implement a number of features that would allow the benchmark to expand its current repertoire of features. Two important features are (i) the reuse of schema elements among multiple instances of randomized data integration scenarios, and (ii) the generation of inclusion dependencies (e.g., foreign keys), which can be randomly created for a given source or target schema, or prescribed by the semantics of a given scenario. These two features add to iBench's pre-existing highly customizable design.

[1] P. C. Arocena, M. D'Angelo, B. Glavic and R. J. Miller, iBench First Cut, Technical Report, Department of Computer Science, University of Toronto, 2013.

A usability case study using iBench
[Joint work with Dr. Patricia C. Arocena , Professor Boris Glavic and Professor Renee J. Miller]
In this project we demonstrate how iBench can be used to empirically evaluate a data exchange mapping operator called MapMerge [2]. The MapMerge operator, proposed by Alexe, Hernandez, Popa, and Tan [2], allows a user to combine a number of independently created schema mappings into a single, simpler mapping. As a graduate student, my role in this project was to study and propose a new set of empirical experiments that would showcase how effective MapMerge can be with varying degrees of schema reuse and inclusion dependencies. The experimental plan I designed goes beyond the existing MapMerge evaluation [2], and covers a wide range of synthetic mappings. My work in this project demonstrates that iBench is well positioned for supporting empirical data integration research.

[2] B. Alexe, M. Hernandez, L. Popa and W. Tan, MapMerge: correlating independent schema mappings, The VLDB Journal, Vol. 21, No. 2, pp. 191-211, 2012.



  • Chunk-Based Software Evolution
    M. Najafi, R. Salay and M. Chechik.
    NECSIS 2014.


  • Verification in Chunk-Based Software Evolution, NECSIS 2014

Relevant Courses

Teaching Assistantships