About

Alexander (Sandy) Paul is a Professor of the Practice at the Georgia Institute of Technology's School of Building Construction. With over 25 years of experience in the real estate and consulting fields, his professional focus is on analyzing economic and commercial real estate trends. He develops and teaches courses on real estate market analysis, policy, trends, and ethics within Georgia Tech's Master of Real Estate Development program. Additionally, he leads the program's conference series in partnership with the Urban Land Institute (ULI). Sandy has held senior leadership roles, including senior managing director of national research at Newmark, where he led a team of more than 70 professionals and significantly expanded the firm's research and thought leadership offerings. Prior to that, he served as executive vice president at Delta Associates, managing client projects, business development, and media outreach, and was responsible for national research activities across multiple metropolitan areas. His research interests are centered on economic and commercial real estate trends, which he applies in classroom lectures and student projects. Sandy holds a Master of Public Policy from Duke University and a Bachelor of Arts from Dartmouth College. He has been recognized with professional credentials and memberships, including the CRE credential by the Counselors of Real Estate and membership in Lambda Alpha, the land economics honor society. He also serves as an adjunct lecturer at Georgetown University and has been a guest lecturer at George Mason University. His expertise and thought leadership have been featured in various industry publications.

Research topics

• Computer Science
• Artificial Intelligence
• Natural Language Processing
• Data Mining
• Computer Security
• Machine Learning
• Speech recognition
• Telecommunications
• Engineering
• Computer network

Selected publications

• An Approach to Named Entity Recognition for Aviation Text

  Communications in computer and information science · 2026-01-01

  book-chapter1st author
  PublisherDOI

• Domain-Specific Health Text Generation Through Low-Rank Adaptation of a Transformer Architecture

  2025-10-27

  article

  The growing demand for accessible and reliable health information has motivated the adaptation of domain-specific large language models (lLMs). LLMs perform well on general natural language processing (NLP) tasks but require fine-tuning for healthcare applications. In this work, Mistral-7B, a 7.3B parameter Transformer model, is fine-tuned for health text generation and noncritical symptom understanding using three parameterefficient methods-Low-Rank Adaptation (LoRA), Quantized Low-Rank Adaptation (QLoRA), and Rank-Optimized Reliable Adaptation (RoRA). A synthetic dataset comprising medical question answering, symptom descriptions, and home remedies was curated from public sources. Experimental results demonstrate that RoRA achieved the highest BLEU-4 (0.52), ROUGE-L (0.65), and F1-score (0.84), outperforming baselines such as BERT, RoBERTa, and LLaMA7B while maintaining low GPU memory usage. This work supports the use of fine-tuned LLMs for safe and efficient health communication, especially in low-resource settings. It also demonstrates that lightweight adaptation using Parameter Efficient Fine-Tuning (PEFT) can deliver high-quality outputs while minimizing computational demands.

  PublisherDOI

• A Region-Specific Nutritional Model Using LSTM Encoder and Attention-Enhanced Decoder

  2025-10-27

  article

  Malnutrition is a persistent challenge in rural India, where generic diet recommendations rarely reflect local food habits or economic realities. Many rural families in India rely on regionally available foods, and generic nutrition models typically overlook these dietary patterns. This paper presents an attention-augmented LSTM encoder-decoder model designed to generate affordable, region-specific meal plans for rural communities in Assam, India. A curated dataset was built using regional recipes, local expert input, and automated web scraping of authentic Bengali and Indian sources. By explicitly modeling cultural and seasonal food patterns, our approach ensures meal suggestions are realistic and easy to adopt in daily village life. The lightweight architecture also enables practical deployment in low-resource healthcare settings without sacrificing performance. On a held-out test set, the proposed model achieved strong results with a BLEU-4 score of 0.42, ROUGE-L of 0.54, F1-score of 0.81, and a BERTScore <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$F 1$</tex> of 0.581, outperforming both transformer and standard LSTM baselines. These results indicate that compact, locally adapted neural models can offer practical nutrition guidance in underserved settings.

  PublisherDOI

• Improving the Efficiency of Pattern Matching Algorithm in Image Mining

  Lecture notes in networks and systems · 2024-01-01 · 2 citations

  book-chapter
  PublisherDOI

• 14 Medical Image Fusion Method by Deep Learning

  2024-09-09 · 1 citations

  book-chapterSenior author
  PublisherDOI

• 21 IoMT-based data aggregation using quantum learning

  2023-07-24

  book-chapterSenior author

  In this chapter, quantum-assisted learning is introduced to aggregate data from the IoMT sensors for possible data processing and storage. The model is designed in such a way that it reduces redundant information and improves the flexibility of processing data. The entire modeling is conducted in a simulation environment to test the effectiveness of data aggregation from IoMT and its associated complexity. Experimental validation shows a reduced complex data aggregation task than with other methods.

  PublisherDOI

• Deep Learning Algorithms in Mobile Edge with Real-Time Abnormal Event Detection for 5G-IoT Devices

  International Journal of Interactive Mobile Technologies (iJIM) · 2023-09-14 · 1 citations

  articleOpen access

  IoT is becoming increasingly popular due to its quick expansion and variety of applications. In addition, 5G technology helps with communication and network connectivity. This work integrates C-RAN with IoT networks to provide an experimental 5G testbed. In a 5G IoT environment, this experience is utilized to enhance both perpendicular and flat localization (3D localization). DRCaG, an acronym for the proposed model, stands for a deep, complicated network with a gated layer on top. The performance of the proposed model has been demonstrated through extensive simulations in terms of learning reduction, accuracy, and matrix disorientation, with a variable signal-to-noise ratio (SNR) spanning from 20 dB to + 20 dB, which illustrates the superiority of DRCaG compared to others. An online, end-to-end solution based on deep learning techniques is presented in this study for the fast, precise, reliable, and automatic detection of diverse petty crime types. By detecting tiny crimes like hostility, bag snatching, and vandalism, the suggested system may not only identify unusual passenger behavior like vandalism and accidents but also improve passenger security. The solution performs admirably in a variety of use cases and environmental settings.

  PublisherOA PDFDOI

• Information-Based Image Extraction with Data Mining Techniques for Quality Retrieval

  Lecture notes in networks and systems · 2023-01-01 · 1 citations

  book-chapter
  PublisherDOI

• Scientific document retrieval using structure encoded string with trie indexing

  Information Services & Use · 2022-03-29

  articleOpen accessSenior author

  Retrieving mathematical expressions from scientific documents is a challenging task as mathematical expressions or formulae are quite different from the traditional text. Mathematical expressions are highly symbolic and complex. Moreover, the structure of a mathematical formula conveys a semantic meaning which cannot be overlooked. This paper proposes a scientific document retrieval system based on mathematical formula query. The paper explores the concept of Structure Encoded String (SES), which has been employed for mathematical expressions to capture the relations among the formula structures. A pattern based trie indexing scheme has been proposed for faster retrieval. The Jaro-Winkler Similarity has been adopted for matching and ranking. Experiments are conducted, results are reported using standard evaluation measures and compared with similar existing systems.

  PublisherDOI

• An Efficient Authentication Using Monitoring Scheme for Node Misbehaviour Detection in MANET

  EAI/Springer Innovations in Communication and Computing · 2022 · 6 citations

  1st authorCorresponding
  • Computer Science
  • Computer network
  • Computer Science
  PublisherDOI

Frequent coauthors

• Umakishore Ramachandran

  Georgia Institute of Technology

  19 shared

• Nissim Harel

  6 shared

• Subodh Adhikari

  Georgia Institute of Technology

  6 shared

• Kenneth MacKenzie

  University of Glasgow

  6 shared

• Bikash Agarwalla

  Georgia Institute of Technology

  5 shared

• Sudipta Roy

  5 shared

• Suresh Venkatasubramanian

  4 shared

• Sourish Dhar

  Assam University

  4 shared

Awards & honors

• CRE credential by the Counselors of Real Estate (2012)
• Membership in Lambda Alpha, the land economics honor society…