Digital Commons Network^™

Revisiting Representation Learning For Natural Language Processing From The Fundamentals For Misinformation Detection, Vishnu S. Pendyala

Faculty Research, Scholarly, and Creative Activity

While existing language embeddings achieve high accuracy on benchmark tasks, recent studies suggest they often struggle to distinguish true from false statements. This work argues that new representation learning and aggregation strategies are needed, especially those that adapt to misinformation's evolving patterns and incorporate explainability. As a first step, we explore two directions: (1) aggregation strategies inspired by signal processing, where word embeddings are transformed into spectrogram-like representations for classification to overcome context window limitations, and (2) future extensions using dynamic neural architectures such as Neural ODEs, Liquid Time-Constant Networks (LTCs), Kolmogorov-Arnold Networks (KANs), and Legendre Memory Units (LMUs). Preliminary …

Huebner, Thomas G., San Jose State University

Emeritus and Retired Faculty Biographies

University of Hawai'i, Linguistics, Ph.D. 1982

University of Pittsburgh, Linguistics, M.A. Certificate: TEFL/TESL

University of Wisconsin‑Milwaukee. Majors: English and Secondary Education, B.S.

Exploring The Benefits Of Yoruba Indigenous Knowledge And The Herbal Medicine System, Sunday Opene Ozonuwe 7317673, Ayobami Aderonke Ogungbade, Faith Ohwofasa

Library Philosophy and Practice (e-journal)

The Yoruba people are one of the West African's main ethno-linguistic or ethnic groups. They have an impressive indigenous knowledge and herbal medicinal system. They pre-occupied Oyo, Osun, Ekiti, Kwara, Ondo, Lagos and Koji state. The researchers developed seven (7) research objectives and seven (7) research questions to investigate the nature, forms, benefits and methods of preparing and administering Yoruba Herbal Medicine. Descriptive research design was adopted for the study, interviews and observations were the instruments used for data collections. Eight (8) herbal medicine practitioners were interviewed in Oyo and Ogun state. The result of the interviews was also presented …

Digital Scholarly Ecosystems: Challenges In Manuscript Preparation And Evaluation, Hrudayananda Padhan

Library Philosophy and Practice (e-journal)

Scholarly communication in the digital era is becoming influenced by the problems of manuscript preparation, journal choice, and peer review in spite of the blistering technological progress and open-access publication development. Researchers frequently encounter difficulties with academic standards in writing, professional ethics of publication, engaging in the search of reliable journals, navigating complex editorial and peer review procedures, which are all the factors affecting the quality, credibility, and visibility of the research products. It follows a narrative literature review methodology to examine academic publications between 2015 and 2025, which will help to determine the main trends, organizational issues, and new …

Dresser, Rocio, San Jose State University

Emeritus and Retired Faculty Biographies

University of San Francisco, International and Multicultural Education, Ed.D. 2000

San Jose State University, Elementary Education (Bilingual Education Emphasis), M.A. 1996

San Jose State University, BCLAD (Bilingual, Crosscultural, Language and Academic Development Credential, 1992

San Jose State University, Creative Arts, B.A. 1990

Spartan Daily, April 22, 2026, San Jose State University, School Of Journalism And Mass Communications

Spartan Daily, 2026

Volume 166, Issue 36

Generative And Large Language Models For Library Knowledge Graphs, Partha Sarathi Mandal, Sukumar Mandal Dr

Library Philosophy and Practice (e-journal)

Structured Abstract

Purpose - Knowledge graphs play a key role in modern library and information services for information discovery, advanced search, and explainable recommendation systems. Recent progress in generative and large language models (LLMs) is changing. The present paper presents a clear and concise overview of how these models support knowledge graph construction in library and information science.

Design/methodology/approach –This study shows a conceptual and analytical approach. It considers the latest trends in generative models and LLMs as applied to the construction of knowledge graphs. Their application in entity extraction, relation identification, schema mapping, and triple generation on unstructured text …

Book Review: Social Justice Design And Implementation In Library And Information Science Edited By Bharat Mehra, Ellie F. Kawada

School of Information Student Research Journal

Social Justice Design and Implementation in Library and Information Science, edited by Bharat Mehra, is an engaging volume that explores the pursuit of social justice in its many forms in the field of Library and Information Science (LIS). The included chapters cover case studies of social justice implementation in library services, theoretical frameworks for subversive methods of research and education, and the future of library advocacy for social justice. While the volume is centered on English-speaking countries and lacks the perspective of paraprofessional library staff, the themes and examples presented in the book are invaluable to any LIS students, professionals, …

Agentic Hallucination Risk Scoring For Medical Llms Via Uncertainty Quantification And Clinical Knowledge Injection, Mayank Kapadia, Mohammad Masum

Faculty Research, Scholarly, and Creative Activity

Large Language Models (LLMs) have witnessed significant adoption across numerous domains since 2020, but their proclivity to hallucinate creates unacceptable dangers in high-risk environments like healthcare, where wrong outputs can directly jeopardize human safety. While present systems focus on pre-generation mitigation strategies, they cannot ensure the safety of individual outputs during inference. We provide a post hoc Hallucination Risk Scoring (HRS) methodology that intercepts questionable outputs before they reach patients via an agentic pipeline. Given a medical question, a domain-specific LLM generates an initial response from which five complimentary uncertainty signals are computed, which are then separated into a decision …

Association Between Dietary Self-Perception And Sugar-Sweetened Beverage Consumption Among Young Adults Latinas, Mi Zhou, A. Susana Ramirez, Angela Xu, Jacqueline Bergman

Faculty Research, Scholarly, and Creative Activity

Objective: To investigate the relationship between self-perceived overall dietary healthfulness and self-reported sugar-sweetened beverage (SSB) consumption among young adult Latinas, accounting for socioeconomic and acculturation-related factors. Design: Cross-sectional analysis using survey data. SSB intake was assessed using the BEVQ-15, and dietary self-perception was measured via a two-item scale. Multiple linear regression models examined associations between self-perception and total daily SSB intake, adjusting for income, education, and two validated acculturation indicators. Setting: Participants were recruited from a national online panel across the United States. Participants: A total of 881 Latina women aged 18–29 years participated. After removing cases with invalid outcome …

Academic Libraries As A Platform For Indigenous And Agricultural Knowledge Sharing In Rural Communities: A Conceptual And Review-Based Approach, Jenifer Yesmin Laskar, Pradip Kumar Singha

Library Philosophy and Practice (e-journal)

Agriculture and traditional knowledge are the roots of rural economy, particularly in biologically sensitive areas with diverse cultures. Preserving traditional knowledge in terms of local farming practices is an activity that is generally associated with cultural practices in these areas. The potentiality of academic libraries in terms of bridging roles for dissemination, curation, and sharing of Indigenous knowledge as well as Agriculture knowledge in the respective areas remains largely overlooked. It is an attempt to explore, indicate, and address through the theoretical approaches based upon an understanding of library science concepts, Indigenous studies, community development, sustainability studies, the roles of …

Rosetta-Xai: An Automated Evaluation And Explainability Framework For Code Translation Models, Vishnu S. Pendyala, Neha Bais Thakur

Faculty Research, Scholarly, and Creative Activity

This paper presents Rosetta-XAI, a comprehensive software framework for evaluating and explaining Large Language Model (LLM) behavior in cross-language code conversion tasks. The system implements a four-stage automated pipeline: (1) code generation by LLMs accessed through the Ollama API inference service, (2) regex-based extraction of code blocks from markdown responses, (3) language-specific syntax and compilation validation with temporary artifact management, and (4) execution with timeout protections and CSV-based checkpoint recovery. The framework supports evaluation of 15 specialized code LLMs (1.3B–34B parameters), including DeepSeek Coder, Code Llama, CodeGemma, and Granite Code across 17 Rosetta Code programming tasks, generating 42 bidirectional conversion …

Artificial Intelligence And Chatgpt In Libraries: A Bibliometric Analysis Of Social Science Research (2022–2026), Shrikant W. Ramteke, Yogesh B. Daphal

Library Philosophy and Practice (e-journal)

Purpose

This paper will map the world of research on artificial intelligence and chatbots in libraries systematically by analysing the trends in publications, intellectual organization, pattern of collaboration, and the way the Social Science literature has evolved with the theme of artificial intelligence and chatbots published between 2022 and 2026.

Design/methodology/approach

We have conducted a thorough bibliometric search where we have considered 7,660 peer-reviewed journal articles found in Scopus database. The collection was compiled according to the PRISMA requirements, and later, both performance-based metrics and science mapping techniques were applied.The whole process of visualisation and analysis was carried out through …

Clinically Aligned Long-Context Transformers For Cross-Platform Mental Health Risk Detection, Aditya Tekale, Mohammad Masum

Faculty Research, Scholarly, and Creative Activity

Social media platforms contain rich but noisy narratives of psychological distress, creating opportunities for early mental health risk detection. However, existing datasets capture heterogeneous constructs such as suicide risk severity, depression diagnosis, and DSM-5 symptom presence, and most prior models are trained and evaluated on a single corpus, limiting their clinical alignment and cross-dataset generalizability. In this study, we fine-tune a domain-specific long-document transformer, AIMH/Mental-Longformer-base-4096, for binary mental health risk detection (risk vs. no risk) using two clinically aligned Reddit datasets: the C-SSRS Reddit corpus and the eRisk 2025 depression dataset. To handle long user histories, we introduce an LLM-based …

The Illusion Of Care: Limited And Superficial Caring Discourse In Bilingual Public Library Interactions, Lili Luo

Faculty Research, Scholarly, and Creative Activity

Objective – To investigate how bilingual reference librarians in a public library express care in their interactions with community members, and to examine whether and how librarians' linguistic choices reflect authentic caring discourse in multilingual contexts. Design – A case study of bilingual reference desk transactions. Setting – Online. Subjects – A public library in a U.S.-Mexico border town. Methods – Members of the research team digitally recorded approximately 20 hours of reference desk interactions over one week and collected extensive field notes. Bilingual transcribers produced full transcripts, categorized by language use (Spanish-only, English-only, bilingual). The research team examined the …

Understanding Tradeoffs In Clinical Text Extraction: Prompting, Retrieval-Augmented Generation, And Supervised Learning On Electronic Health Records, Tanya Yadav, Aditya Tekale, Jeff Chong, Mohammad Masum

Faculty Research, Scholarly, and Creative Activity

Clinical discharge summaries contain rich patient information but remain difficult to convert into structured representations for downstream analysis. Recent advances in large language models (LLMs) have introduced new approaches for clinical text extraction, yet their relative strengths compared with supervised methods remain unclear. This study presents a controlled evaluation of three dominant strategies for structured clinical information extraction from electronic health records: prompting-based extraction using LLMs, retrieval-augmented generation for terminology canonicalization, and supervised fine-tuning of domain-specific transformer models. Using discharge summaries from the MIMIC-IV dataset, we compare zero-shot, few-shot, and verification-based prompting across closed-source and open-source LLMs, evaluate retrieval-augmented canonicalization …

Activation-Guided Layer Selection For Lora, Aditya Dawadikar, Pooja Shyamsundar, Rashmi Vishwanath Bhat, Navrati Saxena

Faculty Research, Scholarly, and Creative Activity

Low-Rank Adaptation (LoRA) has become a widely adopted parameter-efficient fine-tuning (PEFT) technique for large language models (LLMs). LoRA’s benefits stem from its light weight and modular adapters. Standard LoRA applies adapters uniformly across all Transformer layers, implicitly assuming that each layer contributes equally to task adaptation. However, LLMs are found to have internal substructures that contribute in a disproportionate manner. In this work, we provide a theoretical analysis of how LoRA weight updates are influenced by a layer’s activation magnitude. We propose Act-LoRA, a simple activation-guided layer selection strategy for selective Low-Rank Adaptation. We evaluate this strategy for both encoder-only …

Addressing Failing Water Systems: A Qualitative Protocol For Intervention Research On Water Insecurity, Patrick Thomson, Melissa Beresford, Jobayer Hossain, Cara Jacob, Carolina Jordão, Oswaldo Medina-Ramírez, Sophie Neems, Daniel Salcedo Serrano, Dave White, Amber Wutich

Faculty Research, Scholarly, and Creative Activity

It is increasingly acknowledged that high income countries do not enjoy universal drinking water coverage, with pockets of water insecurity persisting close to areas that have had high-quality, centralized water provision for half a century or more. The reasons behind this are complex, and effective ways to improve water security in these areas defy simple prescriptions. Nuanced measures that disaggregate the different facets of water insecurity are especially important as we attempt to understand and, crucially, improve water security in high-income countries.In this paper, we describe a qualitative protocol for assessing water insecurity at the household level in a study …

Preferred Panethnic Terms Among Latina/O And Hispanic Sexual And Gender Minority People, Alexis Ceja, Nguyen K. Tran, Juan M. Peña, David J. Kinitz, Devin Hursey, Ramon Ramirez, Lilia Cervantes, Micah E. Lubensky, Juno Obedin-Maliver, Annesa Flentje, Mitchell R. Lunn

Faculty Research, Scholarly, and Creative Activity

Importance Latino and Hispanic individuals in the US represent a diverse and growing population, including a significant number of sexual and gender minority (SGM) individuals whose identities challenge the use of a single panethnic (ie, ethnicity) label. Little is known about SGM individuals’ preferences for panethnic terms. Objective To understand Latino and Hispanic SGM individuals’ preferences and rationales for panethnic terms. Design, Setting, and Participants This cross-sectional study used survey data from SGM-identifying Latino and Hispanic adults in The PRIDE Study, an online, community-engaged cohort of SGM adults in the US completing 2021 or 2022 annual questionnaires, with qualitative analysis …

Exploring Bilingual Coding For Inclusive Computer Science Learning, Ethel Tshukudu, Neel Asheshbhai Shah, Thien Khang Kieu, Leqaa Deeb, Harshitha Venkateswaran, Aarav Ghai, Yusuf Gadelrab, Purujit Hada

Faculty Research, Scholarly, and Creative Activity

As computing education expands globally, the dominance of English in programming languages can present barriers for students who are not native English speakers. This research work investigates whether language matters in programming by examining how a bilingual programming workshop influences student attitudes, perceived understanding and comfort. Sixty bilingual participants from San Jose State University, 40 with little to no prior programming experience and 20 experienced programmers completed a two-hour workshop using the multilingual Hedy platform. Pre- and post-surveys measured changes in confidence, enjoyment, motivation, identity, and usefulness, complemented by open-ended questions about participants’ experiences. All students showed improvements across most …

A Prepositional Cycle In French, Fabienne Martin, Yining Nie, Artemis Alexiadou

Faculty Research, Scholarly, and Creative Activity

This paper discusses the morphosyntax and semantics of locative prepositions in French, with a particular focus on the prefix de-in complex prepositions (e.g., en dessous de [in from.under of] ‘below’) and in axial nominalizations (e.g., le dessous de [the from.under of] ‘the bottom part/region of’). We propose that in these formations, de-, which developed historically from an ablative marker, retains a Source semantics and lexicalizes the function PROJECT, which defines the region outward from the axial part of an object (Matushansky & Zwarts, 2019). We support our analysis with a corpus study on French child data showing that children acquiring …

A Constructive-Engaging Analysis Of Descarrtes (European) And Oyeshile (African) On Personhood, Olanrewaju Shitta-Bey

Comparative Philosophy

An examination of the nature of the human person provides a crucial foundation for understanding human life and identity. Conceptions of personhood have far-reaching implications beyond philosophy, influencing politics, ethics, religion, and epistemology. Beliefs about what constitutes a person shapes theories of rights, responsibility, and justice; moral views on duty and accountability; religious understandings of the soul and the afterlife; and epistemological reflections on the sources and limits of knowledge. Inquiry into human nature is therefore not merely abstract, but central to understanding social organization and individual conduct. This paper examines two influential accounts of human nature drawn from distinct …

A Cross-System Analysis Of Structural Commensurability Between Axiomatic Foundations Of Set Theory (Zfc) And The Principles Of Transcendent Wisdom, Mohammadjavad Maarefvand

Comparative Philosophy

Building on observed similarities in uniqueness proofs for the empty set and the Necessary Existent (Wājib al-Wujūd), this study engages the meta-methodological issue of cross-system commensurability. It identifies four instances of structural commensurability between Zermelo-Fraenkel set theory with Choice (ZFC) and Mulla Sadra's Transcendent Wisdom (Ḥikmat al-Muta'āliyah). First, identity principles: the Axiom of Extensionality provides a formal operational criterion analogous to the philosophical principle of identity through absence of distinguishing features. Second, multiplicity mechanisms: ZFC's constructor axioms (bottom-up construction) and emanation principles (top-down explanation) address analogous problems with inverse orientations. Third, transcending limitation: the Axiom of Infinity ensuring quantitative extensional …

A Critique Of Ziporyn’S Interpretation Of Early Chinese Thought: From The Constructive-Engagement Vantage Point, Xiaoqiao Liu

Comparative Philosophy

Sameness and difference are among the most fundamental and yet most tension-laden categories in philosophy, bearing directly on the conditions under which judgment, classification, and meaning become possible. This article critically reconstructs and compares two influential but methodologically divergent approaches to this issue in contemporary comparative philosophy. One is Brook Ziporyn’s idea of coherence, which seeks to articulate a non-substantialist metaphysics as a response to tensions internal to the Western metaphysical tradition. The other is Bo Mou’s enhanced account of relative identity, which revises both absolute and modern relative identity theories through a bottom-up, semantically grounded framework oriented toward cross-tradition …

Challenging The Double-Negation Principle: An Analysis Of Strategic Theories Of Sun Zi, Carnot, And Clausewitz Through Intuitionist Logic, Antonino Drago

Comparative Philosophy

To study a subject similarly represented in Eastern and Western cultures, I examine some classical theories of war strategy according to a new analytical method. This method is based on the recognition that the most important strategic theories are organized as theories essentially not deduced from axioms according to classical logic. Such theories reason through doubly negated propositions (DNPs), whose meanings differ from those of their affirmative counterparts. This corresponds to the use of intuitionist logic instead of classical logic. An inspection of the works of the texts of Sun Zi, Lazare Carnot, and Clausewitz reveals many such propositions, and …

Henze, Rosemary C., San Jose State University

Emeritus and Retired Faculty Biographies

Stanford University, Education (Language, Literacy & Culture), Ph.D. 1988

San Francisco State University, English (TESL), M.A. 1980

University of California, Santa Cruz, Art, B.A. 1974

Emotional Patterns In Conversational Social Media Using Graph-Based Context, Sai Sanjay Yerunkar

Master's Projects

Most social media datasets for emotion detection have not been constructed to account for conversational structure‚ so we investigate whether it carries signal for emotion prediction. Using Sentiment140 and GoEmotions Reddit threads‚ we construct their thread-based conversation graphs and compute the aggregated features of neighbors as well as the transformer and TF-IDF representations of comments. Connected comments are 4.5× more similar in emotions than expected by chance. Pairwise emotional similarity decays exponentially with geodesic distance (e.g.‚ after 2 hops). Pairs separated by a 30s timestamp difference have the highest emotional similarity. These results suggest that emotion is structured locally and …

Cost-Aware Predictive Routing For Vision-Language Models, Sahil Sait Naveed

Master's Projects

Vision-language models (VLMs) are increasingly used for multimodal tasks, but their inference costs vary widely across model tiers. This work presents a predictive routing framework that assigns each query to the most cost-effective VLM by using multimodal embeddings, clustering, and per-cluster model error estimates. Evaluated on six commercial VLMs, the selected router reduces average cost by about 49 percent relative to the quality-first setting on both validation and test. This comes with accuracy losses of 0.4 percentage points on validation and 0.9 on held-out test data. It also outperforms K-NN and ZeroRouter routing baselines while providing lowercost operating points than …

Safe-R2r: A Safety-Aware And Budget-Optimized Retrieval Controller For Rag Systems, Mandar Sunil Gondane

Master's Projects

Retrieval-Augmented Generation (RAG) improves factual grounding in large language models by incorporating external evidence during inference. However, most RAG systems rely on fixed retrieval strategies that ignore query difficulty, computational cost, and prediction uncertainty. This project introduces SAFE-R2R (Safety-Aware and Budget-Optimized Reason-to-Retrieve), a retrieval controller that treats retrieval as a query-dependent decision problem. SAFE-R2R organizes retrieval into a multi-rung ladder ranging from no retrieval to deeper retrieval with reranking. At each rung, the system generates an answer, computes reliability signals, and combines them into a risk score used within a conformal calibration framework to decide whether to accept the answer …

Multimodal Emotion Detection System, Shubhankar Sameer Munshi

Master's Projects

Trying to understand emotion from speech is a problem that is present in human computer interaction. Nevertheless, there are still some shortcomings in current SER methods. Text-based systems may miss vital vocal cues, such as sarcasm, tone changes, and delivery. On the other hand, purely audio-based systems are prone to noise and unstable acoustic features. The combination of linguistic and acoustic features in multimodal approaches partially solves this problem, but many existing approaches use inflexible multimodal fusion techniques that cannot adjust their behaviors according to the quality of input signals. In this work, we propose a multimodal approach based on …