arXiv subjects

Language models

Find arXiv papers on language models and computational linguistics, including cs.CL metadata. Use abstracts to explore model training, evaluation and natural-language processing.

Follow this topic with RSS

At least 649 records · Page 36Linked to original sources

arXiv2025

Pastiche Novel Generation Creating: Fan Fiction You Love in Your Favorite Author's Style

Great novels create immersive worlds with rich character arcs, well-structured plots, and nuanced writing styles. However, current novel generation methods often rely on brief, simplistic story outlines and generate details using plain, generic language. To bridge this gap, we introduce the task of Pastiche Novel Generation, which requires the generated novels to imitate the distinctive features of the original work, including understanding character profiles, predicting plausible plot developments, and writing concrete details using vivid, expressive language. To achieve this, we propose WriterAgent, a novel generation system designed to master the core aspects of literary pastiche. WriterAgent is trained through a curriculum learning paradigm, progressing from low-level stylistic mastery to high-level narrative coherence. Its key tasks include language style learning, character modeling, plot planning, and stylish writing, ensuring comprehensive narrative control. To support this, WriterAgent leverages the WriterLoRA framework, an extension of LoRA with hierarchical and cumulative task-specific modules, each specializing in a different narrative aspect. We evaluate WriterAgent on multilingual classics like Harry Potter and Dream of the Red Chamber, demonstrating its superiority over baselines in capturing the target author's settings, character dynamics, and writing style to produce coherent, faithful narratives.

cs.CL↗

arXiv2025

Musical Heritage Historical Entity Linking

Linking named entities occurring in text to their corresponding entity in a Knowledge Base (KB) is challenging, especially when dealing with historical texts. In this work, we introduce Musical Heritage named Entities Recognition, Classification and Linking (MHERCL), a novel benchmark consisting of manually annotated sentences extrapolated from historical periodicals of the music domain. MHERCL contains named entities under-represented or absent in the most famous KBs. We experiment with several State-of-the-Art models on the Entity Linking (EL) task and show that MHERCL is a challenging dataset for all of them. We propose a novel unsupervised EL model and a method to extend supervised entity linkers by using Knowledge Graphs (KGs) to tackle the main difficulties posed by historical documents. Our experiments reveal that relying on unsupervised techniques and improving models with logical constraints based on KGs and heuristics to predict NIL entities (entities not represented in the KB of reference) results in better EL performance on historical documents.

cs.CL↗

arXiv2023

Boosting classification reliability of NLP transformer models in the long run

Transformer-based machine learning models have become an essential tool for many natural language processing (NLP) tasks since the introduction of the method. A common objective of these projects is to classify text data. Classification models are often extended to a different topic and/or time period. In these situations, deciding how long a classification is suitable for and when it is worth re-training our model is difficult. This paper compares different approaches to fine-tune a BERT model for a long-running classification task. We use data from different periods to fine-tune our original BERT model, and we also measure how a second round of annotation could boost the classification quality. Our corpus contains over 8 million comments on COVID-19 vaccination in Hungary posted between September 2020 and December 2021. Our results show that the best solution is using all available unlabeled comments to fine-tune a model. It is not advisable to focus only on comments containing words that our model has not encountered before; a more efficient solution is randomly sample comments from the new period. Fine-tuning does not prevent the model from losing performance but merely slows it down. In a rapidly changing linguistic environment, it is not possible to maintain model performance without regularly annotating new text.

cs.CL↗

arXiv2007

Toward a statistical mechanics of four letter words

We consider words as a network of interacting letters, and approximate the probability distribution of states taken on by this network. Despite the intuition that the rules of English spelling are highly combinatorial (and arbitrary), we find that maximum entropy models consistent with pairwise correlations among letters provide a surprisingly good approximation to the full statistics of four letter words, capturing ~92% of the multi-information among letters and even "discovering" real words that were not represented in the data from which the pairwise correlations were estimated. The maximum entropy model defines an energy landscape on the space of possible words, and local minima in this landscape account for nearly two-thirds of words used in written English.

q-bio.NC↗

arXiv2023

Exploring the Potential of Machine Translation for Generating Named Entity Datasets: A Case Study between Persian and English

This study focuses on the generation of Persian named entity datasets through the application of machine translation on English datasets. The generated datasets were evaluated by experimenting with one monolingual and one multilingual transformer model. Notably, the CoNLL 2003 dataset has achieved the highest F1 score of 85.11%. In contrast, the WNUT 2017 dataset yielded the lowest F1 score of 40.02%. The results of this study highlight the potential of machine translation in creating high-quality named entity recognition datasets for low-resource languages like Persian. The study compares the performance of these generated datasets with English named entity recognition systems and provides insights into the effectiveness of machine translation for this task. Additionally, this approach could be used to augment data in low-resource language or create noisy data to make named entity systems more robust and improve them.

cs.CL↗

arXiv2025

Toward Copyright Integrity and Verifiability via Multi-Bit Watermarking for Intelligent Transportation Systems

Intelligent transportation systems (ITS) use advanced technologies such as artificial intelligence to significantly improve traffic flow management efficiency, and promote the intelligent development of the transportation industry. However, if the data in ITS is attacked, such as tampering or forgery, it will endanger public safety and cause social losses. Therefore, this paper proposes a watermarking that can verify the integrity of copyright in response to the needs of ITS, termed ITSmark. ITSmark focuses on functions such as extracting watermarks, verifying permission, and tracing tampered locations. The scheme uses the copyright information to build the multi-bit space and divides this space into multiple segments. These segments will be assigned to tokens. Thus, the next token is determined by its segment which contains the copyright. In this way, the obtained data contains the custom watermark. To ensure the authorization, key parameters are encrypted during copyright embedding to obtain cipher data. Only by possessing the correct cipher data and private key, can the user entirely extract the watermark. Experiments show that ITSmark surpasses baseline performances in data quality, extraction accuracy, and unforgeability. It also shows unique capabilities of permission verification and tampered location tracing, which ensures the security of extraction and the reliability of copyright verification. Furthermore, ITSmark can also customize the watermark embedding position and proportion according to user needs, making embedding more flexible.

cs.CR↗

arXiv2025

Before It's Too Late: A State Space Model for the Early Prediction of Misinformation and Disinformation Engagement

In today's digital age, conspiracies and information campaigns can emerge rapidly and erode social and democratic cohesion. While recent deep learning approaches have made progress in modeling engagement through language and propagation models, they struggle with irregularly sampled data and early trajectory assessment. We present IC-Mamba, a novel state space model that forecasts social media engagement by modeling interval-censored data with integrated temporal embeddings. Our model excels at predicting engagement patterns within the crucial first 15-30 minutes of posting (RMSE 0.118-0.143), enabling rapid assessment of content reach. By incorporating interval-censored modeling into the state space framework, IC-Mamba captures fine-grained temporal dynamics of engagement growth, achieving a 4.72% improvement over state-of-the-art across multiple engagement metrics (likes, shares, comments, and emojis). Our experiments demonstrate IC-Mamba's effectiveness in forecasting both post-level dynamics and broader narrative patterns (F1 0.508-0.751 for narrative-level predictions). The model maintains strong predictive performance across extended time horizons, successfully forecasting opinion-level engagement up to 28 days ahead using observation windows of 3-10 days. These capabilities enable earlier identification of potentially problematic content, providing crucial lead time for designing and implementing countermeasures. Code is available at: https://github.com/ltian678/ic-mamba. An interactive dashboard demonstrating our results is available at: https://ic-mamba.behavioral-ds.science.

cs.CL↗

arXiv2025

Hallucination Detection in Foundation Models for Decision-Making: A Flexible Definition and Review of the State of the Art

Autonomous systems are soon to be ubiquitous, spanning manufacturing, agriculture, healthcare, entertainment, and other industries. Most of these systems are developed with modular sub-components for decision-making, planning, and control that may be hand-engineered or learning-based. While these approaches perform well under the situations they were specifically designed for, they can perform especially poorly in out-of-distribution scenarios that will undoubtedly arise at test-time. The rise of foundation models trained on multiple tasks with impressively large datasets has led researchers to believe that these models may provide "common sense" reasoning that existing planners are missing, bridging the gap between algorithm development and deployment. While researchers have shown promising results in deploying foundation models to decision-making tasks, these models are known to hallucinate and generate decisions that may sound reasonable, but are in fact poor. We argue there is a need to step back and simultaneously design systems that can quantify the certainty of a model's decision, and detect when it may be hallucinating. In this work, we discuss the current use cases of foundation models for decision-making tasks, provide a general definition for hallucinations with examples, discuss existing approaches to hallucination detection and mitigation with a focus on decision problems, present guidelines, and explore areas for further research in this exciting field.

cs.AI↗

arXiv2025

Performance Evaluation of Sentiment Analysis on Text and Emoji Data Using End-to-End, Transfer Learning, Distributed and Explainable AI Models

Emojis are being frequently used in todays digital world to express from simple to complex thoughts more than ever before. Hence, they are also being used in sentiment analysis and targeted marketing campaigns. In this work, we performed sentiment analysis of Tweets as well as on emoji dataset from the Kaggle. Since tweets are sentences we have used Universal Sentence Encoder (USE) and Sentence Bidirectional Encoder Representations from Transformers (SBERT) end-to-end sentence embedding models to generate the embeddings which are used to train the Standard fully connected Neural Networks (NN), and LSTM NN models. We observe the text classification accuracy was almost the same for both the models around 98 percent. On the contrary, when the validation set was built using emojis that were not present in the training set then the accuracy of both the models reduced drastically to 70 percent. In addition, the models were also trained using the distributed training approach instead of a traditional singlethreaded model for better scalability. Using the distributed training approach, we were able to reduce the run-time by roughly 15% without compromising on accuracy. Finally, as part of explainable AI the Shap algorithm was used to explain the model behaviour and check for model biases for the given feature set.

cs.CL↗

arXiv2025

The "negative end" of change in grammar: terminology, concepts and causes

The topic of "negative end" of change is, contrary to the fields of innovation and emergence, largely under-researched. Yet, it has lately started to gain an increasing attention from language scholars worldwide. The main focus of this article is threefold, namely to discuss the i) terminology; ii) concepts and iii) causes associated with the "negative end" of change in grammar. The article starts with an overview of research conducted on the topic. It then moves to situating phenomena referred to as loss, decline or obsolescence among processes of language change, before elaborating on the terminology and concepts behind it. The last part looks at possible causes for constructions to display a (gradual or rapid, but very consistent) decrease in the frequency of use over time, which continues until the construction disappears or there are only residual or fossilised forms left. Keywords: loss, obsolescence, decline, competition, higher

cs.CL↗

arXiv2020

WOLI at SemEval-2020 Task 12: Arabic Offensive Language Identification on Different Twitter Datasets

Communicating through social platforms has become one of the principal means of personal communications and interactions. Unfortunately, healthy communication is often interfered by offensive language that can have damaging effects on the users. A key to fight offensive language on social media is the existence of an automatic offensive language detection system. This paper presents the results and the main findings of SemEval-2020, Task 12 OffensEval Sub-task A Zampieri et al. (2020), on Identifying and categorising Offensive Language in Social Media. The task was based on the Arabic OffensEval dataset Mubarak et al. (2020). In this paper, we describe the system submitted by WideBot AI Lab for the shared task which ranked 10th out of 52 participants with Macro-F1 86.9% on the golden dataset under CodaLab username "yasserotiefy". We experimented with various models and the best model is a linear SVM in which we use a combination of both character and word n-grams. We also introduced a neural network approach that enhanced the predictive ability of our system that includes CNN, highway network, Bi-LSTM, and attention layers.

cs.CL↗

arXiv2025

Causal Discovery Inspired Unsupervised Domain Adaptation for Emotion-Cause Pair Extraction

This paper tackles the task of emotion-cause pair extraction in the unsupervised domain adaptation setting. The problem is challenging as the distributions of the events causing emotions in target domains are dramatically different than those in source domains, despite the distributions of emotional expressions between domains are overlapped. Inspired by causal discovery, we propose a novel deep latent model in the variational autoencoder (VAE) framework, which not only captures the underlying latent structures of data but also utilizes the easily transferable knowledge of emotions as the bridge to link the distributions of events in different domains. To facilitate knowledge transfer across domains, we also propose a novel variational posterior regularization technique to disentangle the latent representations of emotions from those of events in order to mitigate the damage caused by the spurious correlations related to the events in source domains. Through extensive experiments, we demonstrate that our model outperforms the strongest baseline by approximately 11.05\% on a Chinese benchmark and 2.45\% on a English benchmark in terms of weighted-average F1 score. We have released our source code and the generated dataset publicly at: https://github.com/tk1363704/CAREL-VAE.

cs.CL↗

arXiv2025

RideKE: Leveraging Low-Resource, User-Generated Twitter Content for Sentiment and Emotion Detection in Kenyan Code-Switched Dataset

Social media has become a crucial open-access platform for individuals to express opinions and share experiences. However, leveraging low-resource language data from Twitter is challenging due to scarce, poor-quality content and the major variations in language use, such as slang and code-switching. Identifying tweets in these languages can be difficult as Twitter primarily supports high-resource languages. We analyze Kenyan code-switched data and evaluate four state-of-the-art (SOTA) transformer-based pretrained models for sentiment and emotion classification, using supervised and semi-supervised methods. We detail the methodology behind data collection and annotation, and the challenges encountered during the data curation phase. Our results show that XLM-R outperforms other models; for sentiment analysis, XLM-R supervised model achieves the highest accuracy (69.2\%) and F1 score (66.1\%), XLM-R semi-supervised (67.2\% accuracy, 64.1\% F1 score). In emotion analysis, DistilBERT supervised leads in accuracy (59.8\%) and F1 score (31\%), mBERT semi-supervised (accuracy (59\% and F1 score 26.5\%). AfriBERTa models show the lowest accuracy and F1 scores. All models tend to predict neutral sentiment, with Afri-BERT showing the highest bias and unique sensitivity to empathy emotion. https://github.com/NEtori21/Ride_hailing

cs.CL↗

arXiv2025

Extracting Sentence Embeddings from Pretrained Transformer Models

Pre-trained transformer models shine in many natural language processing tasks and therefore are expected to bear the representation of the input sentence or text meaning. These sentence-level embeddings are also important in retrieval-augmented generation. But do commonly used plain averaging or prompt templates sufficiently capture and represent the underlying meaning? After providing a comprehensive review of existing sentence embedding extraction and refinement methods, we thoroughly test different combinations and our original extensions of the most promising ones on pretrained models. Namely, given 110 M parameters, BERT's hidden representations from multiple layers, and many tokens, we try diverse ways to extract optimal sentence embeddings. We test various token aggregation and representation post-processing techniques. We also test multiple ways of using a general Wikitext dataset to complement BERT's sentence embeddings. All methods are tested on eight Semantic Textual Similarity (STS), six short text clustering, and twelve classification tasks. We also evaluate our representation-shaping techniques on other static models, including random token representations. Proposed representation extraction methods improve the performance on STS and clustering tasks for all models considered. Very high improvements for static token-based models, especially random embeddings for STS tasks, almost reach the performance of BERT-derived representations. Our work shows that the representation-shaping techniques significantly improve sentence embeddings extracted from BERT-based and simple baseline models.

cs.CL↗

arXiv2025

Machine Learning Approaches for Mental Illness Detection on Social Media: A Systematic Review of Biases and Methodological Challenges

The global increase in mental illness requires innovative detection methods for early intervention. Social media provides a valuable platform to identify mental illness through user-generated content. This systematic review examines machine learning (ML) models for detecting mental illness, with a particular focus on depression, using social media data. It highlights biases and methodological challenges encountered throughout the ML lifecycle. A search of PubMed, IEEE Xplore, and Google Scholar identified 47 relevant studies published after 2010. The Prediction model Risk Of Bias ASsessment Tool (PROBAST) was utilized to assess methodological quality and risk of bias. The review reveals significant biases affecting model reliability and generalizability. A predominant reliance on Twitter (63.8%) and English-language content (over 90%) limits diversity, with most studies focused on users from the United States and Europe. Non-probability sampling (80%) limits representativeness. Only 23% explicitly addressed linguistic nuances like negations, crucial for accurate sentiment analysis. Inconsistent hyperparameter tuning (27.7%) and inadequate data partitioning (17%) risk overfitting. While 74.5% used appropriate evaluation metrics for imbalanced data, others relied on accuracy without addressing class imbalance, potentially skewing results. Reporting transparency varied, often lacking critical methodological details. These findings highlight the need to diversify data sources, standardize preprocessing, ensure consistent model development, address class imbalance, and enhance reporting transparency. By overcoming these challenges, future research can develop more robust and generalizable ML models for depression detection on social media, contributing to improved mental health outcomes globally.

cs.LG↗

arXiv2025

Language Shift or Maintenance? An Intergenerational Study of the Tibetan Community in Saudi Arabia

The present study provides the first-ever report on the language shift from Tibetan to Arabic among descendants of Tibetan families who migrated from the Tibet region to Saudi Arabia around 70 years ago. The aim of this study was to determine whether three age groups had adopted different practices in terms of maintaining Tibetan or shifting to Hijazi Arabic. To this end, 96 male and female members of the Tibetan community responded to a questionnaire in which they were asked about their code choice in different domains (home, neighbourhood, friends and relatives, expressing emotion, and performing religious rituals). The data revealed significant intergenerational differences between members of the community in terms of the extent of the shift to Arabic, with Tibetan rarely used by younger members and older members making only slightly more use of it. The difference between the three age groups was significant, at a p-value of .001.

cs.CL↗

arXiv2025

Answer Set Counting and its Applications

We have focused on Answer Set Programming (ASP), more specifically, answer set counting, exploring both exact and approximate methodologies. We developed an exact ASP counter, sharpASP, which utilizes a compact encoding for propositional formulas, significantly enhancing efficiency compared to existing methods that often struggle with inefficient encodings. Our evaluations indicate that sharpASP outperforms current ASP counters on several benchmarks. In addition, we proposed an approximate ASP counter, named ApproxASP, a hashing-based counter integrating Gauss-Jordan elimination within the ASP solver, clingo. As a practical application, we employed ApproxASP for network reliability estimation, demonstrating superior performance over both traditional reliability estimators and #SAT-based methods.

cs.CL↗

arXiv2025

Mind the Gaps: Logical English, Prolog, and Multi-agent Systems for Autonomous Vehicles

In this paper, we present a modular system for representing and reasoning with legal aspects of traffic rules for autonomous vehicles. We focus on a subset of the United Kingdom's Highway Code (HC) related to junctions. As human drivers and automated vehicles (AVs) will interact on the roads, especially in urban environments, we claim that an accessible, unitary, high-level computational model should exist and be applicable to both users. Autonomous vehicles introduce a shift in liability that should not bring disadvantages or increased burden on human drivers. We develop a system "in silico" of the model. The proposed system is built of three main components: a natural language interface, using Logical English, which encodes the rules; an internal representation of the rules in Prolog; and an multi-agent-based simulation environment, built in NetLogo. The three components interact: Logical English is translated into and out of Prolog (along with some support code); Prolog and NetLogo interface via predicates. Such a modular approach enables the different components to carry different "burdens" in the overall system; it also allows swapping of modules. Given NetLogo, we can visualize the effect of the modeled rules as well as validate the system with a simple dynamic running scenario. Designated agents monitor the behaviour of the vehicles for compliance and record potential violations where they occur. The information on potential violations is then utilized by Validators, to determine whether the violation is punishable, differentiating between exceptions and cases.

cs.AI↗