Syracuse University
Syracuse University Research Facility and Collaborative EnvironmentNot a member yet
20376 research outputs found
Sort by
Harnessing LLMs to Detect Hate Speech
Get PDFHate is a sentiment, while hate speech refers to the expression of hate in a form that targets and attacks specific groups, such as race, religion, or gender. With the rise of the internet and social media, hate speech has spread rapidly, gaining wide exposure and posing threats to individual well-being, the profits of major tech companies, and social stability. As a result, both industry and academia have turned their attention to the study of hate speech. One of the most active areas is hate speech detection, which involves training models to predict whether a given piece of content is hateful. However, the choice of models and methods can vary depending on the form in which the hate speech is conveyed. In this thesis, we address two key issues: 1.In the task of hateful meme classification, many existing approaches focus on stacking model parameters to achieve better performance, but lack a deep understanding of how hateful memes are constructed. Furthermore, they have not effectively leveraged large language models (LLMs) for this task. 2.Although the definitions of hate and hate speech are well-established, individuals’ perceptions of hate can vary due to differences in cultural background. As a result, judgments about whether a piece of content is hateful may differ from person to person. However, current hate speech detection models typically rely on labels obtained through majority voting, without accounting for the cultural specificity of individual annotators. To address the first issue, we observe that creators of hateful memes often exaggerate their emotions and reinforce stereotypes, leading to a mismatch between the text and image within the meme. We refer to this phenomenon as a false claim. Based on this insight, we propose the FACT model(FAlse Claim haTeful meme classification model), a model built upon a large language model(LLM), which identifies false claims in memes to assist the classification process. To address the second issue, we first evaluate LLMs and find that they are unable to effectively utilize cultural background information to support reasoning, while historical labeling proves to be useful. Based on this, we hypothesize that an individual’s perception of hate is influenced by specific combinations of cultural background factors. To incorporate this insight, we apply matrix factorization techniques from recommender systems to learn interaction features for each cultural background combination. These features are then used to support culture-aware hate speech detection
How Toxic Communications in Chinese Online Fan Groups affect Fan Engagement
Get PDFAbstract The study used a Self-Determination Theory (SDT) theoretical framework to explore how toxicity impacts fan engagement in online fan groups in China. An online survey method in China with a sample size of 205 participants was used to measure the hypotheses 1)Toxicity causes fans to feel less autonomy in online groups, 2) Toxicity causes fans to feel less skilled (competence) in online groups, 3) Toxicity causes fans to feel less relatedness in online groups and research questions and the research questions 1) How will autonomy, competence, and relatedness relate to specific fandom behaviors? 2) What is the difference between participants’ group means of autonomy, competence, and relatedness (no longer in fan groups and still in fan groups)? The findings of the research revealed significant and negative correlation in autonomy, competence, and relatedness to support three hypotheses. In addition, the findings of the research revealed the more fans feel autonomy, competence, and relatedness, they were more likely to engage in fandom behaviors, such as “Start a new discussion,” “Response to an ongoing discussion,” and more. Also, the findings of the research revealed fans feel more relatedness and fans were more likely to stay in online fan groups. Another interesting finding is that in China, fans care more about relatedness than autonomy and competence. The importance of the study was to make the public better understand how toxicity impacts fans’ need for satisfaction, potentially helping the public understand the difference between fan behaviors, membership in fan groups, and needs for satisfaction. keywords: toxicity, fandom, fandom behaviors, Self-Determination Theory (SDT
Claire Messud, Raymond Carver Reading Series, April 23, 2025
No full textA video recording of Claire Messud reading on April 2, 2025, as part of the Syracuse University Creative Writing Program\u27s Raymond Carver Reading Series
Hari Kunzru, Raymond Carver Reading Series, March 19, 2025
No full textA video recording of Hari Kunzro reading on February 26, 2025, as part of the Syracuse University Creative Writing Program\u27s Raymond Carver Reading Series
DEVELOPMENT, VALIDATION, AND APPLICABILITY OF A RACE-SPECIFIC INSULIN SENSITIVITY ESTIMATOR (RISE) IN YOUTH AND YOUNG ADULTS
No full textParallel to the escalating rate of overweight and obesity in the youth and young adult population, the incidence and prevalence of type 2 diabetes are also mounting, presenting significant public health problems. Among the various pathophysiological factors, insulin sensitivity has emerged as a critical risk factor for impaired glucose tolerance (IGT) and type 2 diabetes in youth. Despite the pivotal role of insulin sensitivity in the development and progression of these conditions, current methodologies for its assessment, such as the hyperinsulinemic-euglycemic clamp and intravenous glucose tolerance test, are limited by their invasive methods, high cost, and unsuitability for use in large-scale clinical and observational studies. Moreover, existing predictive models, primarily derived from adult populations with a wide age range and/or limited racial/ethnic backgrounds, fail to address the unique metabolic characteristics and racial/ethnic disparities observed in insulin sensitivity among youth and young adults. This study introduces an innovative research initiative to develop and validate a Race-specific Insulin Sensitivity Estimator (RISE) based on physical and fasting-based metabolic characteristics using advanced machine learning techniques. Eventually, the RISE evaluated the effectiveness of High-Intensity Interval Training (HIIT) and Time-restricted eating (TRE) as lifestyle interventions designed to improve insulin sensitivity and reduce diabetes risk among young population with overweight and obesity. Further, RISE can be a tool for identifying responders and non-responders to lifestyle interventions. Taken together, this project addressed the critical need for an accurate and accessible tool to assess insulin sensitivity that considers racial/ethnic disparities in youth and young adults. The findings are clinically relevant as we not only created machine-learning-based RISE but also evaluated its applicability in a clinical trial focused on lifestyle changes in diabetes prevention (obesity treatment)
NEGOTIATING DEVELOPMENT: ADIVASI RESISTANCE, STATE DEVELOPMENTALISM, AND THE MAKING OF LESSER CITIZENS IN JHARKHAND, INDIA
No full textThis dissertation explores how Ho adivasis, an indigenous tribe in eastern India, navigate neo-liberal state-led development, extractive capitalism, and liberal democracy. While existing scholarship on tribal communities in South Asia crucially highlights the structural marginalization and identity politics of these groups, it often relies on binary frameworks such as exclusion versus inclusion or resistance versus compliance. Such framings risk flattening internal heterogeneity and ascribing static, essentialized group identities. In contrast, this study foregrounds adivasi personhood, shifting practices and claims of citizenship, and the quotidian negotiations through which democratic life and bureaucratic structures are inhabited, contested, and reconfigured. Drawing on eighteen months of ethnographic fieldwork in Nakahasa, a Ho adivasi village in southern Jharkhand, this study traces the long durée impact of limestone mining on the community. It examines how resource extraction has reshaped landscapes, disrupted agrarian livelihoods, and provoked a locally rooted yet evolving resistance movement. Through participant observation, in-depth interviews, and sustained immersion in everyday village life and political activity, the research explores how adivasi actors articulate new subjectivities and claims to citizenship—not simply by opposing the state, but by negotiating with, appropriating, and reworking the practices of neoliberal governance and development. The dissertation moves beyond political-economic analyses of extractive industries to foreground their affective dimensions—aspirations, suffering, and loss—experienced unevenly across class, gender, and age. These affects are mediated through relational ethics and enacted via strategic positionings, selectively drawing on translocal discourses of indigeneity, bureaucratic-legal idioms, and local adivasi cosmologies. These affective experiences are not merely responses to extraction but are integral to the formation of adivasi subjectivity. To theorize this, the dissertation develops a framework of relational-substantive personhood. Drawing on anthropological concepts of partible and fractal persons (Strathern 2018; Wagner 1991), it conceptualizes Ho subjectivity as constituted through dynamic social ties and enduring attachments to land, labor, and cosmological forces. Applying this framework reveals how resistance operates not merely as opposition to state power, but as a generative space for political becoming and subject formation. For instance, an emergence of protest meetings as informal pedagogical spaces for learning bureaucratic procedures and rights discourse, and birth of naya neta (new youth leaders) contesting village council elections. Through these acts, adivasi actors actively negotiate and rework postcolonial developments and democracy from below. At the same time, the study shows how stereotypes of adivasis are actively produced, reinforced, and utilized within state apparatuses to categorize and govern them as lesser citizens. The power dynamics embedded within administrative practices, the study reveals, create a reified boundary between a modern state and “primitive” adivasi society. In sum, this dissertation offers a window into the workings of the Indian state, extractive industry, and liberal democracy as they unfold in a remote village of eastern India. It contributes a new perspective on subaltern engagements with neo-liberal development and governance, one that refuses reductive binaries of resistance versus complicity or inclusion versus exclusion. Instead, it frames development as a dynamic site for shaping subjectivity and citizenship through contestation, negotiation, and reformulation. By centering ethical life, political aspiration, and everyday maneuvering of state, capital, and development, this research contributes to political anthropology and the anthropology of the state, development, citizenship, and democracy, especially in the context of subaltern and historically marginalized communities in the Global South and beyond
An Examination of the Opt Out Movement In New York State
No full textPublic state education systems are required by federal law to assess student learning in order to receive certain federal funding for the state’s education system. Typically, the metric are standardized testing regimes. New York State is no exception and requires students in grades 3-8 to take annual English Language Arts and Mathematics exams to meet that requirement. However, New York State is a unique case, as it has the highest Opt Out rate, or noncompliance on taking the exams, in the country. While the majority of the literature on the Opt Out Movement has focused on who is more likely to Opt Out and why they do so, there are still gaps existing in the literature. Using semi-structured interviews and Best’s (2021) social problems process, this study interviews both people whose children were Opted Out and those who had their children take the exams, as people are embedded in communities and do not make decisions free from community or societal influences. This study had four research questions: 1.) how has the Opt Out Movement changed/developed over time? 2.) how do parents and education professionals make meaning of education trends and the exams? 3.) what role does class play in the Movement & 4.) how do differing levels of bureaucracy affect Opt Out actors? Major findings were that the high Opt Out rate in New York is due the timing of factors interacting with each other to create the “perfect storm”, 2.) in agreement with much existing literature, class does play a role in the decision-making and meaning-making, but the current bifurcation of wanting to Opt Out vs not supporting it is not accurate. Parents can want to Opt Out, but do not, not everyone Opts Out for the same reasons, and some engage in the behavior, yet do not identify with the Movement, 3) class privilege made the Opt Out Movement possible, but perceptions about supports also drove people away, and 4). Despite educational bureaucracies having formal power, Opt Out supporters were able to use informal power to push back, including engaging in gossip
Holographic Scaling and Cosmological Correlators
No full textIt is well-established that the two-point correlator of a scalar field in AdS follows either of two scaling laws, ∆± =d/2 ± ν, where ν =Sqrt((d/2)^2+ m^2) is a function of the scalar bulk mass. The transition between these two scalings occurs in the renormalization group evolution from the UV to the IR of a dual 4D conformal theory deformed by a relevant double-trace deformation. In this thesis, we identify a new holographic scaling law, ∆UV = ν, in the context of the AdS/CFT correspondence where there exists a UV cutoff brane. Specifically, when ν \u3e 1, the ∆− scaling that violates the unitarity bound, ∆ ≥d/2 −1, will be replaced by the ∆UV scaling which satisfies the unitarity bound. At the transition between ∆UV and ∆+ scaling, there is a particle-like feature in the spectral density of the two-point correlator. We show that in the 4D dual description, this new scaling behavior corresponds to the two-point function of a fundamental scalar field mixing with a CFT operator O with dimension ∆ ≥ 3 through an irrelevant coupling, λϕO. The interplay between cosmological inflation and strongly coupled dynamics yields transient spectral features, which may leave behind phenomenological signatures in the spectrum of primordial fluctuations, and cosmological observables at late times. We study the spectral features of cosmological correlators in a holographic setting during inflation. AdS/CFT arguments suggest that, for large N , in an inflationary background with Hubble rate H, the spectral density for a scalar operator is a gapped continuum. We demonstrate that there can be additional particle-like peak features in phenomenologically interesting regions in parameter space. We estimate their contribution to cosmological observables and find phenomena that are potentially observable in future experiments. We find the new scaling behavior ∆UV = ν with the existence of these quasi-particles. We comment on how stabilization of conformal symmetry breaking vacua can be correlated with these spectral features and their phenomenology. Sections 1, 3, and 4 review existing literature. Sections 2 and 5 feature original wor
Molecular mechanisms governing left right axis formation during vertebrate development.
No full textThe development of a functional ciliated tissue during embryogenesis is vital for essential physiological processes, such as mucociliary clearance in the respiratory tract, cerebrospinal fluid movement in the brain, and ovum transport in the reproductive system. This developmental process of ciliated tissues requires the coordination of intracellular processes, including membrane trafficking, cilia formation, and cytoskeletal organization. While these processes have been extensively studied in cell culture systems, our understanding of their integration during vertebrate organ formation remains limited. Using high-resolution confocal imaging, optogenetic manipulation, and genetic approaches this thesis examines how these processes drive the development of the zebrafish left-right organizer (LRO), called the Kupffer\u27s vesicle (KV), providing insights into both conserved mechanisms and tissue-specific adaptations during vertebrate organogenesis. Chapter 2 examines the role of Rab GTPase family proteins in KV development, demonstrating that ciliogenesis initiates intracellularly before lumen formation, with Rab11 and Rab35 controlling cystic fibrosis transmembrane conductance regulator (CFTR) trafficking for lumen expansion while Rab8 specifically regulates cilia elongation. This chapter reveals conserved and divergent mechanisms previously characterized in mammalian systems, highlighting the importance of in vivo studies for understanding tissue-specific regulation of membrane trafficking. Chapter 3 investigates mitotic patterning during KV morphogenesis, identifying an FGF-regulated anterior enrichment of cell divisions critical for proper organ formation. This chapter establishes the essential role of spatially controlled cell division in establishing proper KV architecture. Chapter 4 provides a comprehensive analysis of protein localization and ultrastructure in KV. This chapter documents significant variability in ciliary architecture, demonstrating that cilia exhibit structural and molecular diversity that likely contributes to their specialized functions in establishing left-right asymmetry. Collectively, this dissertation bridges knowledge gaps between in vitro studies and vertebrate embryonic development seen in zebrafish. Demonstrating that while zebrafish employ many conserved cellular mechanisms during organogenesis, they also exhibit tissue-specific adaptations that could not be predicted from cell culture models. These findings enhance our understanding of the origins of congenital disorders associated with laterality defects and ciliopathies, and provide a high-resolution, cellular-level framework for further exploring the complex interplay of cellular processes and their inherent heterogeneity that drive organ formation in vivo