Julia Taylor Rayz

Geetanjali Bihani, Julia Taylor Rayz

Neural network-based decisions tend to be overconfident, where their raw outcome probabilities do not align with the true decision probabilities. Calibration of neural networks is an essential step towards more reliable deep learning frameworks. Prior metrics of calibration error primarily utilize crisp bin membership-based measures. This exacerbates skew in model probabilities and portrays an incomplete picture of calibration error. In this work, we propose a Fuzzy Calibration Error metric (FCE) that utilizes a fuzzy binning approach to calculate calibration error. This approach alleviates the impact of probability skew and provides a tighter estimate while measuring calibration error. We compare our metric with ECE across different data populations and class memberships. Our results show that FCE offers better calibration error estimation, especially in multi-class settings, alleviating the effects of skew in model confidence scores on calibration error estimation. We make our code and supplementary materials available at: https://github.com/bihani-g/fce

Kanishka Misra, Julia Taylor Rayz, Allyson Ettinger

A characteristic feature of human semantic cognition is its ability to not only store and retrieve the properties of concepts observed through experience, but to also facilitate the inheritance of properties (can breathe) from superordinate concepts (animal) to their subordinates (dog) -- i.e. demonstrate property inheritance. In this paper, we present COMPS, a collection of minimal pair sentences that jointly tests pre-trained language models (PLMs) on their ability to attribute properties to concepts and their ability to demonstrate property inheritance behavior. Analyses of 22 different PLMs on COMPS reveal that they can easily distinguish between concepts on the basis of a property when they are trivially different, but find it relatively difficult when concepts are related on the basis of nuanced knowledge representations. Furthermore, we find that PLMs can demonstrate behavior consistent with property inheritance to a great extent, but fail in the presence of distracting information, which decreases the performance of many models, sometimes even below chance. This lack of robustness in demonstrating simple reasoning raises important questions about PLMs' capacity to make correct inferences even when they appear to possess the prerequisite knowledge.

Kanishka Misra, Julia Taylor Rayz, Allyson Ettinger

To what extent can experience from language contribute to our conceptual knowledge? Computational explorations of this question have shed light on the ability of powerful neural language models (LMs) -- informed solely through text input -- to encode and elicit information about concepts and properties. To extend this line of research, we present a framework that uses neural-network language models (LMs) to perform property induction -- a task in which humans generalize novel property knowledge (has sesamoid bones) from one or more concepts (robins) to others (sparrows, canaries). Patterns of property induction observed in humans have shed considerable light on the nature and organization of human conceptual knowledge. Inspired by this insight, we use our framework to explore the property inductions of LMs, and find that they show an inductive preference to generalize novel properties on the basis of category membership, suggesting the presence of a taxonomic bias in their representations.

Gilchan Park, Julia Rayz, Line Pouchard

Experimental research publications provide figure form resources including graphs, charts, and any type of images to effectively support and convey methods and results. To describe figures, authors add captions, which are often incomplete, and more descriptions reside in body text. This work presents a method to extract figure descriptive text from the body of scientific articles. We adopted ontological semantics to aid concept recognition of figure-related information, which generates human- and machine-readable knowledge representations from sentences. Our results show that conceptual models bring an improvement in figure descriptive sentence classification over word-based approaches.

Ahmad Esmaeili, Julia T. Rayz, Eric T. Matson

Algorithm selection and hyperparameter tuning are critical steps in both academic and applied machine learning. On the other hand, these steps are becoming ever increasingly delicate due to the extensive rise in the number, diversity, and distributedness of machine learning resources. Multi-agent systems, when applied to the design of machine learning platforms, bring about several distinctive characteristics such as scalability, flexibility, and robustness, just to name a few. This paper proposes a fully automatic and collaborative agent-based mechanism for selecting distributedly organized machine learning algorithms and simultaneously tuning their hyperparameters. Our method builds upon an existing agent-based hierarchical machine-learning platform and augments its query structure to support the aforementioned functionalities without being limited to specific learning, selection, and tuning mechanisms. We have conducted theoretical assessments, formal verification, and analytical study to demonstrate the correctness, resource utilization, and computational efficiency of our technique. According to the results, our solution is totally correct and exhibits linear time and space complexity in relation to the size of available resources. To provide concrete examples of how the proposed methodologies can effectively adapt and perform across a range of algorithmic options and datasets, we have also conducted a series of experiments using a system comprised of 24 algorithms and 9 datasets.

Geetanjali Bihani, Julia Taylor Rayz

With data privacy becoming more of a necessity than a luxury in today's digital world, research on more robust models of privacy preservation and information security is on the rise. In this paper, we take a look at Natural Language Steganography (NLS) methods, which perform information hiding in natural language systems, as a means to achieve data security as well as confidentiality. We summarize primary challenges regarding the secrecy and imperceptibility requirements of these systems and propose potential directions of improvement, specifically targeting steganographic text quality. We believe that this study will act as an appropriate framework to build more resilient models of Natural Language Steganography, working towards instilling security within natural language-based neural models.

Geetanjali Bihani, Julia Taylor Rayz

The advent of large language models (LLMs) has enabled significant performance gains in the field of natural language processing. However, recent studies have found that LLMs often resort to shortcuts when performing tasks, creating an illusion of enhanced performance while lacking generalizability in their decision rules. This phenomenon introduces challenges in accurately assessing natural language understanding in LLMs. Our paper provides a concise survey of relevant research in this area and puts forth a perspective on the implications of shortcut learning in the evaluation of language models, specifically for NLU tasks. This paper urges more research efforts to be put towards deepening our comprehension of shortcut learning, contributing to the development of more robust language models, and raising the standards of NLU evaluation in real-world scenarios.

Geetanjali Bihani, Tatiana Ringenberg, Julia Rayz

Encoding implicit language presents a challenge for language models, especially in high-risk domains where maintaining high precision is important. Automated detection of online child grooming is one such critical domain, where predators manipulate victims using a combination of explicit and implicit language to convey harmful intentions. While recent studies have shown the potential of Transformer language models like SBERT for preemptive grooming detection, they primarily depend on surface-level features and approximate real victim grooming processes using vigilante and law enforcement conversations. The question of whether these features and approximations are reasonable has not been addressed thus far. In this paper, we address this gap and study whether SBERT can effectively discern varying degrees of grooming risk inherent in conversations, and evaluate its results across different participant groups. Our analysis reveals that while fine-tuning aids language models in learning to assign grooming scores, they show high variance in predictions, especially for contexts containing higher degrees of grooming risk. These errors appear in cases that 1) utilize indirect speech pathways to manipulate victims and 2) lack sexually explicit content. This finding underscores the necessity for robust modeling of indirect speech acts by language models, particularly those employed by predators.

Damin Zhang, Julia Rayz

Understanding time is fundamental to human cognition, where temporal experience is often conceptualized through spatial metaphors grounded in sensory-motor experience. For example, "summer is approaching" parallels "We are approaching the summer". In such expressions, humans rely on a frame of reference (FoR) to interpret meaning relative to a particular viewpoint. Extending this concept to time, a temporal frame of reference (t-FoR) defines how temporal relations are perceived relative to an experiencer's moment of "now". While Large Language Models (LLMs) have shown remarkable advances in natural language understanding, their ability to interpret and reason about time remains limited. In this work, we introduce TUuD (Temporal Understanding under Deictic t-FoR), a framework that evaluates how LLMs interpret time-event and event-event relations when the reference point of "now" dynamically shifts along a timeline. Following recent work on temporal cognition \cite{li2025other}, LLMs are prompted to rate the similarity between the current moment and a target event from 0.00 (completely dissimilar) to 1.00 (highly similar), where similarity quantifies perceived temporal alignment between the two points. Our results show that four evaluated LLMs exhibit measurable adaptation to a deictic t-FoR, with similarity ratings peaking around the present and decreasing toward past and future events. The adaptation, however, weakens beyond near-term contexts, suggesting that while LLMs display partial human-like temporal cognition, their temporal reasoning remains sensitive to reference-frame shifts and temporal distance.

Geetanjali Bihani, Julia Rayz

With the advent of social media, children are becoming increasingly vulnerable to the risk of grooming in online settings. Detecting grooming instances in an online conversation poses a significant challenge as the interactions are not necessarily sexually explicit, since the predators take time to build trust and a relationship with their victim. Moreover, predators evade detection using indirect and coded language. While previous studies have fine-tuned Transformers to automatically identify grooming in chat conversations, they overlook the impact of coded and indirect language on model predictions, and how these align with human perceptions of grooming. In this paper, we address this gap and evaluate bi-encoders on the task of classifying different degrees of grooming risk in chat contexts, for three different participant groups, i.e. law enforcement officers, real victims, and decoys. Using a fuzzy-theoretic framework, we map human assessments of grooming behaviors to estimate the actual degree of grooming risk. Our analysis reveals that fine-tuned models fail to tag instances where the predator uses indirect speech pathways and coded language to evade detection. Further, we find that such instances are characterized by a higher presence of out-of-vocabulary (OOV) words in samples, causing the model to misclassify. Our findings highlight the need for more robust models to identify coded language from noisy chat inputs in grooming contexts.

Geetanjali Bihani, Julia Rayz

The advent of pre-trained language models (PLMs) has enabled significant performance gains in the field of natural language processing. However, recent studies have found PLMs to suffer from miscalibration, indicating a lack of accuracy in the confidence estimates provided by these models. Current evaluation methods for PLM calibration often assume that lower calibration error estimates indicate more reliable predictions. However, fine-tuned PLMs often resort to shortcuts, leading to overconfident predictions that create the illusion of enhanced performance but lack generalizability in their decision rules. The relationship between PLM reliability, as measured by calibration error, and shortcut learning, has not been thoroughly explored thus far. This paper aims to investigate this relationship, studying whether lower calibration error implies reliable decision rules for a language model. Our findings reveal that models with seemingly superior calibration portray higher levels of non-generalizable decision rules. This challenges the prevailing notion that well-calibrated models are inherently reliable. Our study highlights the need to bridge the current gap between language model calibration and generalization objectives, urging the development of comprehensive frameworks to achieve truly robust and reliable language models.

Damin Zhang, Yi Zhang, Geetanjali Bihani et al.

With the impressive performance in various downstream tasks, large language models (LLMs) have been widely integrated into production pipelines, like recruitment and recommendation systems. A known issue of models trained on natural language data is the presence of human biases, which can impact the fairness of the system. This paper investigates LLMs' behavior with respect to gender stereotypes, in the context of occupation decision making. Our framework is designed to investigate and quantify the presence of gender stereotypes in LLMs' behavior via multi-round question answering. Inspired by prior works, we construct a dataset by leveraging a standard occupation classification knowledge base released by authoritative agencies. We tested three LLMs (RoBERTa-large, GPT-3.5-turbo, and Llama2-70b-chat) and found that all models exhibit gender stereotypes analogous to human biases, but with different preferences. The distinct preferences of GPT-3.5-turbo and Llama2-70b-chat may imply the current alignment methods are insufficient for debiasing and could introduce new biases contradicting the traditional gender stereotypes.

Gilchan Park, Julia Taylor Rayz, Cleveland G. Shields

This paper aims to provide an approach for automatic coding of physician-patient communication transcripts to improve patient-centered communication (PCC). PCC is a central part of high-quality health care. To improve PCC, dialogues between physicians and patients have been recorded and tagged with predefined codes. Trained human coders have manually coded the transcripts. Since it entails huge labor costs and poses possible human errors, automatic coding methods should be considered for efficiency and effectiveness. We adopted three machine learning algorithms (Naïve Bayes, Random Forest, and Support Vector Machine) to categorize lines in transcripts into corresponding codes. The result showed that there is evidence to distinguish the codes, and this is considered to be sufficient for training of human annotators.

Kanishka Misra, Allyson Ettinger, Julia Taylor Rayz

Building on research arguing for the possibility of conceptual and categorical knowledge acquisition through statistics contained in language, we evaluate predictive language models (LMs) -- informed solely by textual input -- on a prevalent phenomenon in cognitive science: typicality. Inspired by experiments that involve language processing and show robust typicality effects in humans, we propose two tests for LMs. Our first test targets whether typicality modulates LM probabilities in assigning taxonomic category memberships to items. The second test investigates sensitivities to typicality in LMs' probabilities when extending new information about items to their categories. Both tests show modest -- but not completely absent -- correspondence between LMs and humans, suggesting that text-based exposure alone is insufficient to acquire typicality knowledge.

Geetanjali Bihani, Julia Taylor Rayz

Contextual word representation models have shown massive improvements on a multitude of NLP tasks, yet their word sense disambiguation capabilities remain poorly explained. To address this gap, we assess whether contextual word representations extracted from deep pretrained language models create distinguishable representations for different senses of a given word. We analyze the representation geometry and find that most layers of deep pretrained language models create highly anisotropic representations, pointing towards the existence of representation degeneration problem in contextual word representations. After accounting for anisotropy, our study further reveals that there is variability in sense learning capabilities across different language models. Finally, we propose LASeR, a 'Low Anisotropy Sense Retrofitting' approach that renders off-the-shelf representations isotropic and semantically more meaningful, resolving the representation degeneration problem as a post-processing step, and conducting sense-enrichment of contextualized representations extracted from deep neural language models.

Geetanjali Bihani, Julia Taylor Rayz

Current intent classification approaches assign binary intent class memberships to natural language utterances while disregarding the inherent vagueness in language and the corresponding vagueness in intent class boundaries. In this work, we propose a scheme to address the ambiguity in single-intent as well as multi-intent natural language utterances by creating degree memberships over fuzzified intent classes. To our knowledge, this is the first work to address and quantify the impact of the fuzzy nature of natural language utterances over intent category memberships. Additionally, our approach overcomes the sparsity of multi-intent utterance data to train classification models by using a small database of single intent utterances to generate class memberships over multi-intent utterances. We evaluate our approach over two task-oriented dialog datasets, across different fuzzy membership generation techniques and approximate string similarity measures. Our results reveal the impact of lexical overlap between utterances of different intents, and the underlying data distributions, on the fuzzification of intent memberships. Moreover, we evaluate the accuracy of our approach by comparing the defuzzified memberships to their binary counterparts, across different combinations of membership functions and string similarity measures.

Kanishka Misra, Julia Taylor Rayz

Humans often communicate by using imprecise language, suggesting that fuzzy concepts with unclear boundaries are prevalent in language use. In this paper, we test the extent to which models trained to capture the distributional statistics of language show correspondence to fuzzy-membership patterns. Using the task of natural language inference, we test a recent state of the art model on the classical case of temperature, by examining its mapping of temperature data to fuzzy-perceptions such as "cool", "hot", etc. We find the model to show patterns that are similar to classical fuzzy-set theoretic formulations of linguistic hedges, albeit with a substantial amount of noise, suggesting that models trained solely on language show promise in encoding fuzziness.

Xiaonan Jing, Yi Zhang, Qingyuan Hu et al.

Twitter can be viewed as a data source for Natural Language Processing (NLP) tasks. The continuously updating data streams on Twitter make it challenging to trace real-time topic evolution. In this paper, we propose a framework for modeling fuzzy transitions of topic clusters. We extend our previous work on crisp cluster transitions by incorporating fuzzy logic in order to enrich the underlying structures identified by the framework. We apply the methodology to both computer generated clusters of nouns from tweets and human tweet annotations. The obtained fuzzy transitions are compared with the crisp transitions, on both computer generated clusters and human labeled topic sets.

Xiaonan Jing, Qingyuan Hu, Yi Zhang et al.

Twitter serves as a data source for many Natural Language Processing (NLP) tasks. It can be challenging to identify topics on Twitter due to continuous updating data stream. In this paper, we present an unsupervised graph based framework to identify the evolution of sub-topics within two weeks of real-world Twitter data. We first employ a Markov Clustering Algorithm (MCL) with a node removal method to identify optimal graph clusters from temporal Graph-of-Words (GoW). Subsequently, we model the clustering transitions between the temporal graphs to identify the topic evolution. Finally, the transition flows generated from both computational approach and human annotations are compared to ensure the validity of our framework.

Qingyuan Hu, Yi Zhang, Kanishka Misra et al.

Natural Language Inference (NLI) or Recognizing Textual Entailment (RTE) is the task of predicting the entailment relation between a pair of sentences (premise and hypothesis). This task has been described as a valuable testing ground for the development of semantic representations, and is a key component in natural language understanding evaluation benchmarks. Models that understand entailment should encode both, the premise and the hypothesis. However, experiments by Poliak et al. revealed a strong preference of these models towards patterns observed only in the hypothesis, based on a 10 dataset comparison. Their results indicated the existence of statistical irregularities present in the hypothesis that bias the model into performing competitively with the state of the art. While recast datasets provide large scale generation of NLI instances due to minimal human intervention, the papers that generate them do not provide fine-grained analysis of the potential statistical patterns that can bias NLI models. In this work, we analyze hypothesis-only models trained on one of the recast datasets provided in Poliak et al. for word-level patterns. Our results indicate the existence of potential lexical biases that could contribute to inflating the model performance.

Xiaonan Jing, Julia Taylor Rayz

Graph structures are powerful tools for modeling the relationships between textual elements. Graph-of-Words (GoW) has been adopted in many Natural Language tasks to encode the association between terms. However, GoW provides few document-level relationships in cases when the connections between documents are also essential. For identifying sub-events on social media like Twitter, features from both word- and document-level can be useful as they supply different information of the event. We propose a hybrid Graph-of-Tweets (GoT) model which combines the word- and document-level structures for modeling Tweets. To compress large amount of raw data, we propose a graph merging method which utilizes FastText word embeddings to reduce the GoW. Furthermore, we present a novel method to construct GoT with the reduced GoW and a Mutual Information (MI) measure. Finally, we identify maximal cliques to extract popular sub-events. Our model showed promising results on condensing lexical-level information and capturing keywords of sub-events.

Yifei Hu, Xiaonan Jing, Youlim Ko et al.

Spelling irregularities, known now as spelling mistakes, have been found for several centuries. As humans, we are able to understand most of the misspelled words based on their location in the sentence, perceived pronunciation, and context. Unlike humans, computer systems do not possess the convenient auto complete functionality of which human brains are capable. While many programs provide spelling correction functionality, many systems do not take context into account. Moreover, Artificial Intelligence systems function in the way they are trained on. With many current Natural Language Processing (NLP) systems trained on grammatically correct text data, many are vulnerable against adversarial examples, yet correctly spelled text processing is crucial for learning. In this paper, we investigate how spelling errors can be corrected in context, with a pre-trained language model BERT. We present two experiments, based on BERT and the edit distance algorithm, for ranking and selecting candidate corrections. The results of our experiments demonstrated that when combined properly, contextual word embeddings of BERT and edit distance are capable of effectively correcting spelling errors.

Geetanjali Bihani, Julia Taylor Rayz

Static word embeddings encode word associations, extensively utilized in downstream NLP tasks. Although prior studies have discussed the nature of such word associations in terms of biases and lexical regularities captured, the variation in word associations based on the embedding training procedure remains in obscurity. This work aims to address this gap by assessing attributive word associations across five different static word embedding architectures, analyzing the impact of the choice of the model architecture, context learning flavor and training corpora. Our approach utilizes a semi-supervised clustering method to cluster annotated proper nouns and adjectives, based on their word embedding features, revealing underlying attributive word associations formed in the embedding space, without introducing any confirmation bias. Our results reveal that the choice of the context learning flavor during embedding training (CBOW vs skip-gram) impacts the word association distinguishability and word embeddings' sensitivity to deviations in the training corpora. Moreover, it is empirically shown that even when trained over the same corpora, there is significant inter-model disparity and intra-model similarity in the encoded word associations across different word embedding models, portraying specific patterns in the way the embedding space is created for each embedding architecture.

Shih-Feng Yang, Julia Taylor Rayz

Twitter is one of the most popular microblogging services in the world. The great amount of information within Twitter makes it an important information channel for people to learn and share news. Twitter hashtag is an popular feature that can be viewed as human-labeled information which people use to identify the topic of a tweet. Many researchers have proposed event-detection approaches that can monitor Twitter data and determine whether special events, such as accidents, extreme weather, earthquakes, or crimes take place. Although many approaches use hashtags as one of their features, few of them explicitly focus on the effectiveness of using hashtags on event detection. In this study, we proposed an event detection approach that utilizes hashtags in tweets. We adopted the feature extraction used in STREAMCUBE and applied a clustering K-means approach to it. The experiments demonstrated that the K-means approach performed better than STREAMCUBE in the clustering results. A discussion on optimal K values for the K-means approach is also provided.