Esma Nafiye Polat

I hold a Bachelor's degree in Computer Engineering with a double major in Mathematics and a Master's degree in the same field. I recently completed my Ph.D. in Computer Engineering, specializing in Conversational AI, with a focus on Large Language Models (LLMs). With twelve years of research and practical experience in natural language processing (NLP), data science, machine learning, and deep learning, I am deeply engaged in advancing the frontiers of technology.

My recent research has concentrated on developing prompt engineering and fine-tuning techniques using transformer-based models and LLMs, specifically for Conversational Sentiment Analysis. I utilize advanced methods such as Llama 2, GPT-4, and GPT-3.5 Turbo. In the business sector, I excel at analyzing, summarizing, and deriving insights from large-scale call transcripts to enhance the personalization and effectiveness of agent-customer interactions.

I am currently developing agent profiling models aimed at generalizing agent behaviors to optimize and rerank our offer recommendation systems, with the goal of boosting company revenue. Additionally, I tackle challenges such as entity ambiguity and conduct detailed sentiment and emotional analysis in conversational data. My work also extends to applications in optimization and forecasting.

Proficient in programming languages including Python, R, Matlab, and C, I am committed to applying my expertise to solve complex challenges and create impactful solutions in the realm of Conversational AI.

Scholarship

• 10/2020 SmartCon Conference Scholarship
• 10/2019 SmartCon Conference Scholarship
• 10/2018 SmartCon Conference Scholarship
• 09/2018-06/2024 Ozyegin University, Full Scholarship for Ph.D study
• 09/2011-09/2013 Istanbul Sehir University, Full Scholarship for M.Sc study
• 09/2005-06/2010 Maltepe University, Full Scholarship for B.Sc study
• 09/2002-06/2005 Private Erciyes School, Full Scholarship for high school education

Awards

• 11/2013 ICECCO Conference, The best track paper award
• 07/2010 Maltepe University, Ranked 3'rd in Computer Engineering Bachelors
• 07/2005 Private Erciyes School, Ranked 3'rd in high school graduation

Painting

Polat, E. N., Kafescioglu N., Demiroglu C.(2024). Decoding Emotional Dynamics: A Comparative Analysis of Contextual and Non-Contextual Models in Sentiment Analysis of Turkish Couple Dialogues (Submitted to IEEE Access).

Koçak, T., Dibek, Ç., Polat, E. N., Kafescioglu N., Demiroglu C.(2023). Automatic Detection of Attachment Style in Married Couples Through Conversation Analysis. Computer Speech & Language. EURASIP Journal on Audio, Speech, and Music Processing. https://doi.org/10.1186/s13636-023-00291-w.

Polat, N., Cakmak, A., Turan, R. N. (2019). Exploring the Power of Supervised Learning Methods for Company Name Disambiguation in Microblog Posts. Turkish Journal of Electrical Engineering and Computer Sciences. DOI: 10.3906/elk-1809-167.

N. Polat, Experiments on company name disambiguation with supervised classification techniques, Published in Electronics, Computer and Computation (ICECCO), 2013 International Conference, 10.1109/ICECCO.2013.6718248, Pages 139-142. (The best track paper award)

N. Polat. (2014). Exploring the Supervised Learning Methods for Company Name Disambiguation on Microblog Posts. (M.S. Dissertation). (392490)

Decoding Emotional Dynamics: A Comparative Analysis of Contextual and Non-Contextual Models in Sentiment Analysis of Turkish Couple Dialogues

A transformer-based Capsule Network for Conversational Sentiment Analysis in Dyadic Relationships.

Predicting the attachment styles of recently married couples using acoustic features.

Multiplicative Normalizing Flows on Capsule Networks. 

Profile Similarity In Cross Social Media Platform 

Statistical Properties of Text

Sentiment Analysis on Turkish Text 

Named Entity Recognition on Turkish Text

Text Generator with Tensorflow 

AlexNet Network with Cifar10

Log Anomaly Detection on Apache Kafka 

Implementing a Bayesian Linear Regressor on PyTorch 

Implementing Variational Dropout with DenseLayer

Implement Variational Bayesian Linear Regressor on PyTorch 

PyTorch implementation of the Variational Auto-Encoder 

Bayesian Version of DenseNets with Normalizing flows

Bayesian Version of CapsulNets with Stochastic Gradient Hamiltonian Monte Carlo

Fraud Analysis with Python on Kaggle 

Data Anomaly Detection on Blockchain Data 

Sentiment Analysis with Python Handwritten Recognition with Matlab 

Solving to Company Name Ambiguity Problem in Microblog Posts Posts

Natural Language Processing 

Social Media Analysis

Deep Learning

Bayesian Statistics in Machine Learning 

Big Data

Distributed Systems and Big Data Artificial Neural Networks

Python in Data Science 

Data Science (Coursera)

Information Retrieval

Machine Learning (Coursera) 

Data Mining

Social Network Analysis 

Data Engineering

Numerical Analysis for Engineers 

Introduction to Cryptography 

Linear Dynamical System

Cenk Demiroğlu, Ozyegin University, Associate Professor, cenk.demiroglu@ozyegin.edu.tr, Istanbul, Turkey, PhD thesis advisor

Emre Yazıcı, Founder of Mina Proje Yönetim, yaziciemre@gmail.com, Istanbul, Turkey, Former Team Leader

Rabia Nuray Turan, Interos Inc., Chief Data Scientist, rnuray@gmail.com, Menlo Park, California 94025 U.S.A, Graduate thesis advisor

Ali Çakmak, Istanbul Technical University, Associate Professor, ali.cakmak@itu.edu.tr, Istanbul, Turkey, Graduate thesis co-advisor

NLP with Attention Models 

Generative AI Applications

Large Language Models

Data Science

Machine Learning

Deep Learning 

Information Retrieval

Bayesian Statistics in Deep Learning

Social Network Analysis 

Big Data Analysis.