While traditional probabilistic models such as Latent Dirichlet Allocation (LDA) (Blei et al., 2003) have been foundational, their underlying bag - of - words assumption limits their ability to capture complex semantics. A recent paradigm shift towards models like BERTopic (Grootendorst, 2022), a state - of - the - art (SOTA) model which leverages contextualized embeddings from pre - trained transformers, has shown significant promise in generating more semantically coherent topics. These models can capture nuanced relationships, including domain - speci fic named entities and morphologically rich constructs, critical for linguistically complex data. However, despite this progress, two significant gaps persist in literature. First, research has overwhelmingly focused on high - resource, standardized languages, with a lot of scope left for under - resourced languages to be unexplored. This focus not only limits the generalizability of existing models but also risks perp etuating a technological bias where the nuances of smaller linguistic communities are overlooked. Models trained on standard corpora often fail to capture the unique lexical and semantic patterns of regional dialects or sociolects, leading to a superficial or even inaccurate understanding of the underlying discourse (Kamilo g lu, 2025) . Second, the predominant application domain has been structured or short - form text like news articles or social media posts (Egger et al., 2022; Schäfer et al., 2024), while the challenges of modeling unstructured, open - ended personal narratives have received less attention. Distinct from the short - form, often decontextualized nature of social media data, daily narratives provide granular, contextually - grounded accounts of lived experience.

A Large Language Model (LLM) offers versatility across domains and tasks, purportedly benefiting users with a wide variety of behaviors and preferences. We question this perception about an LLM when users have inherently subjective behaviors and preferences, as seen in their ubiquitous and idiosyncratic browsing of websites or apps. The sequential behavior logs of pages, thus generated, form something akin to each user's self-constructed "language", albeit without the structure and grammar imbued in natural languages. We ask: (i) Can a small LM represent the "language of browsing" better than a large LM? (ii) Can an LM with a single set of parameters (or, single LM) adequately capture myriad users' heterogeneous, subjective behaviors and preferences? (iii) Can a single LM with high average performance, yield low variance in performance to make alignment good at user level? We introduce clusterwise LM training, HeTLM (Heterogeneity aware Training of Language Model), appropriate for subjective behaviors. We find that (i) a small LM trained using a page-level tokenizer outperforms large pretrained or finetuned LMs; (ii) HeTLM with heterogeneous cluster specific set of parameters outperforms a single LM of the same family, controlling for the number of parameters; and (iii) a higher mean and a lower variance in generation ensues, implying improved alignment.

Legal documents pose unique challenges for text classification due to their domain-specific language and often limited labeled data. This paper proposes a hybrid approach for classifying legal texts by combining unsupervised topic and graph embeddings with a supervised model. We employ Top2Vec to learn semantic document embeddings and automatically discover latent topics, and Node2Vec to capture structural relationships via a bipartite graph of legal documents. The embeddings are combined and clustered using KMeans, yielding coherent groupings of documents. Our computations on a legal document dataset demonstrate that the combined Top2Vec+Node2Vec approach improves clustering quality over text-only or graph-only embeddings. We conduct a sensitivity analysis of hyperparameters, such as the number of clusters and the dimensionality of the embeddings, and demonstrate that our method achieves competitive performance against baseline Latent Dirichlet Allocation (LDA) and Non-Negative Matrix Factorization (NMF) models. Key findings indicate that while the pipeline presents an innovative approach to unsupervised legal document analysis by combining semantic topic modeling with graph embedding techniques, its efficacy is contingent upon the quality of initial topic generation and the representational power of the chosen embedding models for specialized legal language. Strategic recommendations include the exploration of domain-specific embeddings, more comprehensive hyperparameter tuning for Node2Vec, dynamic determination of cluster numbers, and robust human-in-the-loop validation processes to enhance legal relevance and trustworthiness. The pipeline demonstrates potential for exploratory legal data analysis and as a precursor to supervised learning tasks but requires further refinement and domain-specific adaptation for practical legal applications.

This paper presents Constrained Centroid Clustering (CCC), a method that extends classical centroid-based clustering by enforcing a constraint on the maximum distance between the cluster center and the farthest point in the cluster. Using a Lagrangian formulation, we derive a closed-form solution that maintains interpretability while controlling cluster spread. To evaluate CCC, we conduct experiments on synthetic circular data with radial symmetry and uniform angular distribution. Using ring-wise, sector-wise, and joint entropy as evaluation metrics, we show that CCC achieves more compact clusters by reducing radial spread while preserving angular structure, outperforming standard methods such as K-means and GMM. The proposed approach is suitable for applications requiring structured clustering with spread control, including sensor networks, collaborative robotics, and interpretable pattern analysis.

The Prognostics and Health Management Data Challenge (PHM) 2016 tracks the health state of components of a semiconductor wafer polishing process. The ultimate goal is to develop an ability to predict the measurement on the wafer surface wear through monitoring the components health state. This translates to cost saving in large scale production. The PHM dataset contains many time series measurements not utilized by traditional physics based approach. On the other hand task, applying a data driven approach such as deep learning to the PHM dataset is non-trivial. The main issue with supervised deep learning is that class label is not available to the PHM dataset. Second, the feature space trained by an unsupervised deep learner is not specifically targeted at the predictive ability or regression. In this work, we propose using the autoencoder based clustering whereby the feature space trained is found to be more suitable for performing regression. This is due to having a more compact distribution of samples respective to their nearest cluster means. We justify our claims by comparing the performance of our proposed method on the PHM dataset with several baselines such as the autoencoder as well as state-of-the-art approaches.

Spectral clustering is a popular clustering method. It first maps data into the spectral embedding space and then uses Kmeans to find clusters. However, the two decoupled steps prohibit joint optimization for the optimal solution. In addition, it needs to construct the similarity graph for samples, which suffers from the curse of dimensionality when the data are high-dimensional. To address these two challenges, we introduce \textbf{D}eep \textbf{S}pectral \textbf{C}lustering (\textbf{DSC}), which consists of two main modules: the spectral embedding module and the greedy Kmeans module. The former module learns to efficiently embed raw samples into the spectral embedding space using deep neural networks and power iteration. The latter module improves the cluster structures of Kmeans on the learned spectral embeddings by a greedy optimization strategy, which iteratively reveals the direction of the worst cluster structures and optimizes embeddings in this direction. To jointly optimize spectral embeddings and clustering, we seamlessly integrate the two modules and optimize them in an end-to-end manner. Experimental results on seven real-world datasets demonstrate that DSC achieves state-of-the-art clustering performance.

Unsupervised clustering of student responses to open-ended questions into behavioral and cognitive profiles using pre-trained LLM embeddings is an emerging technique, but little is known about how well this captures pedagogically meaningful information. We investigate this in the context of student responses to open-ended questions in biology, which were previously analyzed and clustered by experts into theory-driven Knowledge Profiles (KPs). Comparing these KPs to ones discovered by purely data-driven clustering techniques, we report poor discoverability of most KPs, except for the ones including the correct answers. We trace this "discoverability bias" to the representations of KPs in the pre-trained LLM embeddings space.

In recent years, incomplete multi-view clustering, which studies the challenging multi-view clustering problem on missing views, has received growing research interests. Although a series of methods have been proposed to address this issue, the following problems still exist: 1) Almost all of the existing methods are based on shallow models, which is difficult to obtain discriminative common representations. 2) These methods are generally sensitive to noise or outliers since the negative samples are treated equally as the important samples. In this paper, we propose a novel incomplete multi-view clustering network, called Cognitive Deep Incomplete Multi-view Clustering Network (CDIMC-net), to address these issues. Specifically, it captures the high-level features and local structure of each view by incorporating the view-specific deep encoders and graph embedding strategy into a framework. Moreover, based on the human cognition, i.e., learning from easy to hard, it introduces a self-paced strategy to select the most confident samples for model training, which can reduce the negative influence of outliers. Experimental results on several incomplete datasets show that CDIMC-net outperforms the state-of-the-art incomplete multi-view clustering methods.

Trees are convenient models for obtaining explainable predictions on relatively small datasets. Although there are many proposals for the end-to-end construction of such trees in supervised learning, learning a tree end-to-end for clustering without labels remains an open challenge. As most works focus on interpreting with trees the result of another clustering algorithm, we present here a novel end-to-end trained unsupervised binary tree for clustering: Kauri. This method performs a greedy maximisation of the kernel KMeans objective without requiring the definition of centroids. We compare this model on multiple datasets with recent unsupervised trees and show that Kauri performs identically when using a linear kernel. For other kernels, Kauri often outperforms the concatenation of kernel KMeans and a CART decision tree.

In this paper, we show that preprocessing data using a variant of rank transformation called 'Average Rank over an Ensemble of Sub-samples (ARES)' makes clustering algorithms robust to data representation and enable them to detect varying density clusters. Our empirical results, obtained using three most widely used clustering algorithms-namely KMeans, DBSCAN, and DP (Density Peak)-across a wide range of real-world datasets, show that clustering after ARES transformation produces better and more consistent results.