Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Rahul Mundlamuri has a total of 27 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    The Evolution of AI: From Classical Machine Learning to Modern Large Language Models

    by Rahul Mundlamuri, Ganesh Reddy Gunnam, Nikhil Kumar Mysari, Jayakanth Pujuri

    This paper provides a comprehensive review of the evolution of artificial intelligence from early symbolic, rule-based systems to modern large language models (LLMs) and retrieval-augmented generation (RAG) architectures. Early AI was dominated by symbolic reasoning and expert systems using hand-crafted rules, before a paradigm shift toward data-driven learning occurred with the advent of machine learning, notably the backpropagation algorithm that enabled training of multi-layer neural networks to learn complex tasks. This breakthrough ushered in the deep learning era: convolutional neural networks (CNNs) achieved landmark results in computer vision , recurrent neural networks (RNNs such as LSTM) excelled at sequential data processing, and deep reinforcement learning (RL) systems mastered challenging decision-making tasks. In natural language processing (NLP), distributed word embedding models like Word2Vec and GloVe replaced sparse representations, capturing semantic relationships in vector space. To handle rare and out-of-vocabulary words, subword tokenization techniques such as Byte Pair Encoding (BPE) and WordPiece were introduced. These advances culminated in the Transformer architecture, based on self-attention mechanisms, which now underpins most state-of-the-art LLMs. Transformer-based LLMs, pre-trained on large text corpora, demonstrate advanced capabilities in language generation and understanding. Beyond model architectures, the review highlights how LLMs are augmented with external knowledge via RAG. Retrieval-augmented generation combines an LLM with a vector database of text embeddings, allowing the model to retrieve relevant external information to ground its outputs. This approach significantly improves factual accuracy and relevance, mitigating issues like hallucination, though ensuring complete truthfulness remains challenging. Additional obstacles include the limited context windows of LLMs (on the order of a few thousand tokens) and the computational complexity and latency of inference. The review also examines societal implications of advanced AI, such as bias amplification from training data and other ethical considerations in deployment. Finally, it outlines future directions, including extended context lengths, improved interpretability and alignment, greater efficiency, and neuro-symbolic hybrid approaches, aiming to develop more robust and trustworthy AI systems.


  • Open Access

    SDR-Fi-Z: A Wireless Local Area Network-Fingerprinting-Based Indoor Positioning Method for E911 Vertical Accuracy Mandate

    by Rahul Mundlamuri, Devasena Inupakutika, David Akopian

    The Enhanced 911 (E911) mandate of the Federal Communications Commission (FCC) drives the evolution of indoor three-dimensional (3D) location/positioning services for emergency calls. Many indoor localization systems exploit location-dependent wireless signaling signatures, often called fingerprints, and machine learning techniques for position estimation. In particular, received signal strength indicators (RSSIs) and Channel State Information (CSI) in Wireless Local Area Networks (WLANs or Wi-Fi) have gained popularity and have been addressed in the literature. While RSSI signatures are easy to collect, the fluctuation of wireless signals resulting from environmental uncertainties leads to considerable variations in RSSIs, which poses a challenge to accurate localization on a single floor, not to mention multi-floor or even three-dimensional (3D) indoor localization. Considering recent E911 mandate attention to vertical location accuracy, this study aimed to investigate CSI from Wi-Fi signals to produce baseline Z-axis location data, which has not been thoroughly addressed. To that end, we utilized CSI measurements and two representative machine learning methods, an artificial neural network (ANN) and convolutional neural network (CNN), to estimate both 3D and vertical-axis positioning feasibility to achieve E911 accuracy compliance.


  • Performance of Data Model over Computational Power in CDP Platforms

    by Rahul Mundlamuri, Hassan Siddiqi, Sohail Wadera

    In the era of big data, optimizing the performance of data models is critical to harnessing the full potential of computational power in Customer Data Platforms (CDP). This paper investigates the relationship between data model performance and computational resources within CDP environments, with a specific focus on the comparative analysis of star schema and snowflake schema. Our study examines scalability, efficiency, and processing speed across these two schema designs. We conduct a comprehensive analysis using various data models and computational configurations to identify key performance bottlenecks and their underlying causes. The study employs benchmark datasets and real-world scenarios to evaluate the impacts of different schema structures, indexing strategies, and resource allocation techniques on overall system performance. The findings reveal significant insights into how optimized data models, particularly star and snowflake schemas, can enhance computational efficiency, reduce latency, and improve throughput in CDP environments. Furthermore, we propose best practices and optimization strategies that can be adopted by data engineers and architects to achieve superior performance in large-scale data environments. This research contributes to the ongoing efforts in big data optimization, offering practical guidelines to maximize resource utilization and elevate the performance of data-intensive applications.

  • Open Access

    Abstract B021: Quitxt research study: A mobile messaging intervention to promote smoking cessation among young adult smokers in South Texas

    by Patricia Chalela, Vivian Cortez, Sandra Sivak, Armida Flores, Martha De La Mora, Zully Garcia, Byeongyeob Choi, Edgar Muñoz, Cliff Despres, Alfred L. McAlister, et al.

    Background: Despite major advances in tobacco control and treatment, tobacco use remains the single largest preventable cause of morbidity and mortality in the US. Smoking prevalence is highest among Texas young adults (ages 18-29), especially those with lower education and income and those living in rural areas, like Latinos. About 19.2% of Latinos ages 18-29 in the study areas are current smokers, putting them at higher risk of cancer and other tobacco-related morbidity and mortality. Young adults are heavy users of smartphones, text messaging, social media chats, and other mobile social media, providing a remarkable opportunity for innovation in the delivery of health promotion interventions to reduce health disparities in this large and rapidly growing racial/ethnic population. This research study aims to assess the effectiveness of Quitxt, our culturally appropriate mobile smoking cessation program. Quitxt – launched by our interdisciplinary research team using proven social cognitive, motivational interviewing, and brief intervention methods for promoting behavior change – blends bilingual text and social media messaging for smoking cessation tailored to the language and culture of young adult smokers in our vulnerable region of South Texas. The team also adapted the program to the Primary Care and Mays Cancer Center patient population, and it is currently being implemented. Quitxt has not been tested in a research study, as its creation as an evidence-based cancer prevention service for young adult smokers was supported by the Cancer Prevention and Research Institute of Texas. Methods: The proposed 5-year study involves a 2-group parallel randomized controlled trial with 4 assessments (baseline, 1, 3, and 6 months) and will recruit 1,200 young adult Latino smokers: 1)The intervention group will receive Quitxt (text messaging or chat); and 2) the usual care group will receive abbreviated text messaging with health promotion-related content, including smoking cessation and referral to the Texas DSHS cessation program YesQuit. Smoking cessation will be validated via biological tests for those who report smoking cessation at the study follow-ups. We hypothesize that the group receiving the Quitxt intervention will achieve significantly higher smoking cessation rates than the group receiving usual care. Results: The proposed intervention will mark the first significant study to assess the effectiveness of an innovative, bilingual, culturally, and linguistically appropriate mobile cessation intervention tailored to young adult Latinos living in South Texas by enhancing their skills development, competence, and self-efficacy to initiate and maintain cessation. Conclusion: This study will advance public health by testing the effectiveness of a scalable, evidence-based, easily disseminated, and adaptable intervention with potentially broad national reach to help young adults stop smoking and reduce smoking-related cancer and chronic disease morbidity and mortality and their associated healthcare costs. Citation Format: Patricia Chalela, Vivian Cortez, Sandra Sivak, Armida Flores, Martha De La Mora, Zully Garcia, Byeongyeob Choi, Edgar Muñoz, Cliff Despres, Alfred L. McAlister, Qurathul Ahmed, Audrey Cordova, Maria I. Sung-Cuadrado, Luis F. Velez, Sahak Kaghyan, Rahul Mundlamuri, David Akopian, Amelie G. Ramirez. Quitxt research study: A mobile messaging intervention to promote smoking cessation among young adult smokers in South Texas [abstract]. In: Proceedings of the 17th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2024 Sep 21-24; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2024;33(9 Suppl):Abstract nr B021.

  • Custom Natural Language Understanding for Healthcare Chatbots and A Case Study

    by Devasena Inupakutika, David Akopian, Ganesh Reddy, Patricia Chalela, Sahak Kaghyan, Rahul Mundlamuri

    Text message-based conversational agents or chatbots are computer programs that run on a variety of messaging platforms such as Facebook Messenger, short messaging service, and Slack, among others. In particular, chatbots as conversational agents have been establishing means of communication in healthcare to help provide automated services and improve communication between patients and healthcare professionals. Recently, there is a surge in publicly available Natural Language Understanding (NLU) and Processing (NLP) services, which facilitate creation of chatbots. However, for domain specific cases the technical customizations can be exploited for chatbot optimizations, considering constrained chatbot vocabularies. This paper investigates NLU performance for applications with a limited number of intents and utterances per intent to improve classification performance in closed domain dialog systems. To facilitate this investigation, firstly a customized NLU service is presented with two state-of-the-art deep learning models for a healthcare intervention-based chatbot: 1) a Continuous Bag of words with shallow neural network classifier (CBOW-NN), and 2) a Bidirectional Encoder Representation from Transformers (BERT)-based SciBERT-NLU. The SciBERT is fine-tuned on a domain-specific dataset. Secondly, based on this dataset and other chatbot style datasets, a validation methodology is adapted for the classification performance of the custom and the most popular existing cloud-based off- the-shelf NLU services. Experimental results confirm that the custom BERT-based service with limited vocabulary outperforms, and the CBOW-NN approach achieves comparable performance with cloud-based reference NLU services.


  • Open Access

    Assessing Performance of Cloud-Based Heterogeneous Chatbot Systems and A Case Study

    by Ganesh Reddy Gunnam, Devasena Inupakutika, Rahul Mundlamuri, Sahak Kaghyan, David Akopian

    Recently, human-machine digital assistants gained popularity and are commonly used in question-and-answer applications and similar consumer-supporting domains. A class of more sophisticated digital assistants (chatbots) employing more extended dialogs follows the trend. Chatbots have become increasingly popular in recent years. Nowadays, chatbot deployments in the cloud have become a common practice because of their benefits, including flexibility, scalability, reliability, security, remote working, cost, and power outages. However, measuring the cloud-based chatbot systems’ performance is challenging as human-machine information exchanges are performed in heterogeneous environments such as cloud hosting platforms, information processing units, and several machine-to-machine and human-machine communication channels. This paper investigates different methodologies for assessing the performance of such heterogeneous deployments and identifies performance metrics for evaluating the performance of cloud-based chatbot deployment. The study employs chatbot performance measurements with both real users (human) and automated (simulated) users. The experimental results discuss communication metrics such as response time, throughput, and load testing (connection loss) through the performance assessment of a case study deployment that utilizes an automated protocol chatbot development framework. The findings presented in this paper can further help practitioners to better understand the performance characteristics of a cloud-based chatbot and assist in making informed decisions related to the chatbot development and deployment options.