HomePage:
Bio:
I began my studies as a Graduate Research Assistant in DaSe (Data Semantic Laboratory) under the supervision of Professors Michelle Cheatham and Pascal Hitzler.
I have a Bachelor of Engineering in Computer and I started my PhD right after my graduation from bachelor program. The first year of my PhD program was filled with taking lots of classes; however, I had some interesting research experiences in Sentiment Analysis, Opinion Mining, NLP, and Privacy Preserving Data Mining.
I defended my master thesis Summer 2016 and then started on my PhD.
Email: amini.reihaneh [at] gmail [dot] com or amini.2 [at] wright [dot] edu
Now I am working full time at Apple Inc and I put my PhD on hold!
Research Interests:
___________________________________________________
Work Experience:
- Apple Inc, Maps | Data Science Intern
Jan 2018 - Current | San Francisco, CA
NLP data scientist and enrichement.
- DaSe (Data Semantics) Laboratory | Data Analyst (Graduate Research Student)
Sep 2014 – Dec 2017 | Dayton, OH
Data analyst and linked data developer in BioData (Clinical Trials), GeoData (Traffic Data), and Social Media (Microblogs).
- Apple Inc, Maps | Data Science Intern
May 2017 - Aug 2017 | San Francisco, CA
Working on a NLP-, Deep Learning- (LSTM Model), and Machine Learning-oriented project.
__________________________________________________
Projects:
- High-Quality Domain-Specific Data Interlinking:
Current data linking systems are not accurate enough for many real-world tasks. A preliminary
analysis of the mistakes made by such systems indicates that different approaches may be
needed depending on the domain of the data being linked. In this project I am exploring
the performance of data linking systems on datasets from various domains in an effort to
develop link quality metrics and an improved data linking algorithm. (phD primary research).
Technologies involved: Semantic Web, Linked Data, Link Quality
- Leveraging Amazon's Mechanical Turk for Ontology Alignment Verification:
Because the lack of accuracy in even the best-performing Ontology Alignment Systems, the results of these systems need to be manually verified using Crowdsourcing platforms. In this project, we are analyzing the impact of different factors such as the
presentation of the context of potential matches and the way in which the question is presented to workers on the accuracy of crowdsourcing for alignment verification.
Technologies involved: Ontology Alignment, Human Computer Interaction (HCI), Crowdsourcing
Acknowledgment: Intelligent Agent Incident Command System Augmentation and EarthCube Building Blocks: GeoLink.
- Gender Bias in RateMyProfessors.com: A Text-Based Analysis:
We utilize a new approach for sentiment analysis and document classification to analyze a large data corpus containing
students’ comments of their professors on ratemyprofessors.com for evidence of gender-based bias after controlling for confounding factors like field of study and university ranking.
Technologies involved: Sentiment Analysis, Machine Learning
- Computational Activity Ontology Design Pattern.
A critical issue that has arisen in science-related fields is the reproducibility of results.
Much research involves complex computational analysis of raw data in order to arrive at its
conclusions. In this project, we are designing an ontology design pattern to capture enough
information and context about such a computational analysis to reproduce it. We are also
developing some text processing tools to populate this schema based on reseach articles
(typically PDF documents). The tools publish this information as RDF triples according to
our data model. The information can then be used by other researchers and applications in
a seamless manner.
Technologies involved: NLP, Semantic Web
_______________________________________________