Language Representation Model Research Articles

The role of lexical and prosodic characteristics of mothers' child-directed speech for the early vocabulary development of Italian children with cochlear implants.

Variability in the vocabulary outcomes of children with cochlear implants (CIs) is partially explained by child-directed speech (CDS) characteristics. Yet, relatively little is known about whether and how mothers adapt their lexical and prosodic characteristics to the child's hearing status (before and after implantation, and compared with groups with normal hearing (NH)) and how important they are in affecting vocabulary development in the first 12 months of hearing experience. To investigate whether mothers of children with CIs produce CDS with similar lexical and prosodic characteristics compared with mothers of age-matched children with NH, and whether they modify these characteristics after implantation. In addition, to investigate whether mothers' CDS characteristics predict children's early vocabulary skills before and after implantation. A total of 34 dyads (17 with NH, 17 with children with CIs; ages = 9-32 months), all acquiring Italian, were involved in the study. Mothers' and children's lexical quantity (tokens) and variety (types), mothers' prosodic characteristics (pitch range and variability), and children's vocabulary skills were assessed at two time points, corresponding to before and 1 year post-CI activation for children with CIs. Children's vocabulary skills were assessed using parent reports; lexical and prosodic characteristics were observed in semi-structured mother-child interactions. Results showed that mothers of children with CIs produced speech with similar lexical quantity but lower lexical variety, and with increased pitch range and variability, than mothers of children with NH. Mothers generally increased their lexical quantity and variety and their pitch range between sessions. Children with CIs showed reduced expressive vocabulary and lower lexical quantity and variety than their peers 12 months post-CI activation. Mothers' prosodic characteristics did not explain variance in children's vocabulary skills; their lexical characteristics predicted children's early vocabulary and lexical outcomes, especially in the NH group, but were not related to later language development. Our findings confirm previous studies on other languages and support the idea that the lexical characteristics of mothers' CDS have a positive effect on children's early measures of vocabulary development across hearing groups, whereas prosodic cues play a minor role. Greater input quantity and quality may assist children in the building of basic language model representations, whereas pitch cues may mainly serve attentional and emotional processes. Results emphasize the need for additional longitudinal studies investigating the input received from other figures surrounding the child and its role for children's language development. What is already known on the subject Mothers' CDS is thought to facilitate and support language acquisition in children with various language developmental trajectories, including children with CIs. Because children with CIs are at risk for language delays and have acoustic processing limitations, their mothers may have to produce a lexically simpler but prosodically richer input, compared to mothers of children with NH. Yet, the literature reports mixed findings and no study to our knowledge has concurrently addressed the role of mothers' lexical and prosodic characteristics for children's vocabulary development before implantation and in the first 12 months of hearing experience. What this study adds to the existing knowledge The study shows that mothers of children with CIs produce input of similar quantity but reduced variety, and with heightened pitch characteristics, compared to mothers of children with NH. There was also a general increase in mothers' lexical quantity and variety, and in their pitch range, between sessions. Only their lexical characteristics predicted children's early vocabulary skills. Their lexical variety predicted children's expressive vocabulary and lexical variety only in the NH group. What are the practical and clinical implications of this work? These findings expand our knowledge about the effects of maternal input and may contribute to the improvement of early family-centred intervention programmes for supporting language development in children with CIs.

A deep neural network model for Chinese toponym matching with geographic pre-training model

ABSTRACT Multiple tasks within the field of geographical information retrieval and geographical information sciences necessitate toponym matching, which involves the challenge of aligning toponyms that share a common referent. The multiple string similarity approaches struggle when confronted with the complexities associated with unofficial and/or historical variants of identical toponyms. Also, current state-of-the-art approaches/tools to supervised machine learning rely on labeled samples, and they do not adequately address the intricacies of character replacements either from transliterations or historical shifts in linguistic and cultural norms. To address these issues, this paper proposes a novel matching approach that leverages a deep neural network model empowered by geographic language representation model, known as GeoBERT, which stands for geographic Bidirectional Encoder Representations from Transformers (BERT). This model harnesses the groundbreaking capabilities of the GeoBERT framework by extending a generalized Enhanced Sequential Inference Model architecture and integrating multiple features to enhance the accuracy and robustness of the toponym matching. We present a comprehensive evaluation of the proposed method’s performance using three extensive datasets. The findings clearly illustrate that our approach outperforms the individual similarity metrics used in previous studies.

PERCORE: A Deep Learning-Based Framework for Persian Spelling Correction with Phonetic Analysis

This research introduces a state-of-the-art Persian spelling correction system that seamlessly integrates deep learning techniques with phonetic analysis, significantly enhancing the accuracy and efficiency of natural language processing (NLP) for Persian. Utilizing a fine-tuned language representation model, our methodology effectively combines deep contextual analysis with phonetic insights, adeptly correcting both non-word and real-word spelling errors. This strategy proves particularly effective in tackling the unique complexities of Persian spelling, including its elaborate morphology and the challenge of homophony. A thorough evaluation on a wide-ranging dataset confirms our system’s superior performance compared to existing methods, with impressive F1-Scores of 0.890 for detecting real-word errors and 0.905 for correcting them. Additionally, the system demonstrates a strong capability in non-word error correction, achieving an F1-Score of 0.891. These results illustrate the significant benefits of incorporating phonetic insights into deep learning models for spelling correction. Our contributions not only advance Persian language processing by providing a versatile solution for a variety of NLP applications but also pave the way for future research in the field, emphasizing the critical role of phonetic analysis in developing effective spelling correction system.

CTGGAN: Controllable Text Generation with Generative Adversarial Network

Controllable Text Generation (CTG) aims to modify the output of a Language Model (LM) to meet specific constraints. For example, in a customer service conversation, responses from the agent should ideally be soothing and address the user’s dissatisfaction or complaints. This imposes significant demands on controlling language model output. However, demerits exist among traditional methods. Promoting and fine-tuning language models exhibit the “hallucination” phenomenon and cannot guarantee complete adherence to constraints. Conditional language models (CLM), which map control codes into LM representations or latent space, require training the modified language models from scratch and a high amount of customized dataset is demanded. Decoding-time methods employ Bayesian Rules to modify the output of the LM or model constraints as a combination of energy functions and update the output along the low-energy direction. Both methods are confronted with the efficiency sampling problem. Moreover, there are no methods that consider the relation between constraints weights and the contexts, as is essential in actual applications such as customer service scenarios. To alleviate the problems mentioned above, we propose Controllable Text Generation with Generative Adversarial Networks (CTGGAN), which utilizes a language model with logits bias as the Generator to produce constrained text and employs the Discriminator with learnable constraint weight combinations to score and update the generation. We evaluate the method in the text completion task and Chinese customer service dialogues scenario, and our method shows superior performance in metrics such as PPL and Dist-3. In addition, CTGGAN also exhibits efficient decoding compared to other methods.

DialGNN: Heterogeneous Graph Neural Networks for Dialogue Classification

Dialogue systems have attracted growing research interests due to its widespread applications in various domains. However, most research work focus on sentence-level intent recognition to interpret user utterances in dialogue systems, while the comprehension of the whole documents has not attracted sufficient attention. In this paper, we propose DialGNN, a heterogeneous graph neural network framework tailored for the problem of dialogue classification which takes the entire dialogue as input. Specifically, a heterogeneous graph is constructed with nodes in different levels of semantic granularity. The graph framework allows flexible integration of various pre-trained language representation models, such as BERT and its variants, which endows DialGNN with powerful text representational capabilities. DialGNN outperforms on CM and ECS datasets, which demonstrates robustness and the effectiveness. Specifically, our model achieves a notable enhancement in performance, optimizing the classification of document-level dialogue text. The implementation of DialGNN and related data are shared through https://github.com/821code/DialGNN.

Identifying patients in need of psychological treatment with language representation models

AbstractEarly diagnosis of psychological disorders is very important for patients to regain their health. Research shows that many patients do not realize that they have a psychological disorder or apply to different departments for treatment. The detection of hidden psychological disorders in patients will both increase the quality of life of patients and reduce the traffic of patients who apply to the wrong department. This study aimed to determine whether patients who consult a physician for any reason need psychological treatment. For this purpose, the relationships, and similarities between the sentences of previous psychiatric patients and the sentences of newly arrived patients were analyzed. Domain-based trained ELECTRA language model was used to detect sentence similarities semantically. In the study, the dialogues of patients with physicians in 92 different specialties were analyzed using the MedDialog dataset, which consists of online physician applications, and the DAIC-WOZ dataset. As a result of the experiments, 90.49% success was achieved for the MedDialog dataset and 89.36% for the DAIC-WOZ dataset. With the proposed model, patients in need of psychological treatment were identified and the medical departments where psychological problems were revealed the most were determined. These divisions are Neurology, Sexology, Cardiology, and Plastic Surgery, respectively. With the findings obtained, complications caused by psychological problems and types of diseases that are precursors to psychological disorders were determined. To the best of our knowledge, this article is the first study that aims to analyze all psychological illness instead of focusing on any of the psychological problems (depression, OCD, schizophrenia, etc.) and validated by electronic health records.

BERT-5mC: an interpretable model for predicting 5-methylcytosine sites of DNA based on BERT

DNA 5-methylcytosine (5mC) is widely present in multicellular eukaryotes, which plays important roles in various developmental and physiological processes and a wide range of human diseases. Thus, it is essential to accurately detect the 5mC sites. Although current sequencing technologies can map genome-wide 5mC sites, these experimental methods are both costly and time-consuming. To achieve a fast and accurate prediction of 5mC sites, we propose a new computational approach, BERT-5mC. First, we pre-trained a domain-specific BERT (bidirectional encoder representations from transformers) model by using human promoter sequences as language corpus. BERT is a deep two-way language representation model based on Transformer. Second, we fine-tuned the domain-specific BERT model based on the 5mC training dataset to build the model. The cross-validation results show that our model achieves an AUROC of 0.966 which is higher than other state-of-the-art methods such as iPromoter-5mC, 5mC_Pred, and BiLSTM-5mC. Furthermore, our model was evaluated on the independent test set, which shows that our model achieves an AUROC of 0.966 that is also higher than other state-of-the-art methods. Moreover, we analyzed the attention weights generated by BERT to identify a number of nucleotide distributions that are closely associated with 5mC modifications. To facilitate the use of our model, we built a webserver which can be freely accessed at: http://5mc-pred.zhulab.org.cn.

Predictive Recognition of DNA-binding Proteins Based on Pre-trained Language Model BERT.

Identifying proteins is crucial for disease diagnosis and treatment. With the increase of known proteins, large-scale batch predictions are essential. However, traditional biological experiments being time-consuming and expensive are difficult to accomplish this task efficiently. Nevertheless, deep learning algorithms based on big data analysis have manifested potential in this aspect. In recent years, language representation models, especially BERT, have made significant advancements in natural language processing. In this paper, using three protein segmentation methods and three encoder numbers, nine BERT models with different sizes are constructed to predict whether known proteins are DNA-binding proteins or not. Furthermore, based on the concept of protein motifs, multi-scale convolutional networks are fused into the models to extract the local features of DNA-binding proteins. Finally, we find that the larger the number of encoders, the better the model predictions under the condition of considering each amino acid in the protein as a word. Our proposed algorithm achieves 81.88% sensitivity and 0.39 MCC value on the test set. Furthermore, it achieves 62.41% accuracy on the independent test set PDB2272. It is evident that our proposed method can be a tool to assist in the identification of DNA-binding proteins.

MMBERT: a unified framework for biomedical named entity recognition.

Named entity recognition (NER) is an important task in natural language processing (NLP). In recent years, NER has attracted much attention in the biomedical field. However, due to the lack of biomedical named entity identification datasets, the complexity and rarity of biomedical named entities and so on, biomedical NER is more difficult than general domain NER. So in this paper, we propose a framework (MMBERT) based on Transformer to solve the problems above. To address the scarcity of biomedical named entity recognition datasets, we introduce ERNIE-Health, a new Chinese language representation model pre-trained on large-scale biomedical text corpora. Because of the complexity and rarity of biomedical named entities, we use the Bert and CW-LSTM structures to get the joint feature vector of word pairs relations. In addition, we design multi-granularity 2D convolution to refine the relationship and representation between word pairs. Finally, we design a convolutional neural network (CNN) structure and a co-predictor to improve the model's generalization capability and prediction accuracy. We have conducted extensive experiments on three benchmark datasets, and the experimental results show that our model achieves the best results compared with several baseline models in the experiment.

CnGeoPLM: Contextual knowledge selection and embedding with pretrained language representation model for the geoscience domain

CnGeoPLM: Contextual knowledge selection and embedding with pretrained language representation model for the geoscience domain

Knowledge Graph Construction for Social Customer Advocacy in Online Customer Engagement

The rise of online social networks has revolutionized the way businesses and consumers interact, creating new opportunities for customer word-of-mouth (WoM) and brand advocacy. Understanding and managing customer advocacy in the online realm has become crucial for businesses aiming to cultivate a positive brand image and engage with their target audience effectively. In this study, we propose a framework that leverages the pre-trained XLNet- (bi-directional long-short term memory) BiLSTM- conditional random field (CRF) architecture to construct a Knowledge Graph (KG) for social customer advocacy in online customer engagement (CE). The XLNet-BiLSTM-CRF model combines the strengths of XLNet, a powerful language representation model, with BiLSTM-CRF, a sequence labeling model commonly used in natural language processing tasks. This architecture effectively captures contextual information and sequential dependencies in CE data. The XLNet-BiLSTM-CRF model is evaluated against several baseline architectures, including variations of BERT integrated with other models, to compare their performance in identifying brand advocates and capturing CE dynamics. Additionally, an ablation study is conducted to analyze the contributions of different components in the model. The evaluation metrics, including accuracy, precision, recall, and F1 score, demonstrate that the XLNet-BiLSTM-CRF model outperforms the baseline architectures, indicating its superior ability to accurately identify brand advocates and label customer advocacy entities. The findings highlight the significance of leveraging pre-trained contextual embeddings, sequential modeling, and sequence labeling techniques in constructing effective models for constructing a KG for customer advocacy in online engagement. The proposed framework contributes to the understanding and management of customer advocacy by facilitating meaningful customer-brand interactions and fostering brand loyalty.

A Short-Text Similarity Model Combining Semantic and Syntactic Information

As one of the prominent research directions in the field of natural language processing (NLP), short-text similarity has been widely used in search recommendation and question-and-answer systems. Most of the existing short textual similarity models focus on considering semantic similarity while overlooking the importance of syntactic similarity. In this paper, we first propose an enhanced knowledge language representation model based on graph convolutional networks (KEBERT-GCN), which effectively uses fine-grained word relations in the knowledge base to assess semantic similarity and model the relationship between knowledge structure and text structure. To fully leverage the syntactic information of sentences, we also propose a computational model of constituency parse trees based on tree kernels (CPT-TK), which combines syntactic information, semantic features, and attentional weighting mechanisms to evaluate syntactic similarity. Finally, we propose a comprehensive model that integrates both semantic and syntactic information to comprehensively evaluate short-text similarity. The experimental results demonstrate that our proposed short-text similarity model outperforms the models proposed in recent years, achieving a Pearson correlation coefficient of 0.8805 on the STS-B dataset.

Localizing in-domain adaptation of transformer-based biomedical language models

In the era of digital healthcare, the huge volumes of textual information generated every day in hospitals constitute an essential but underused asset that could be exploited with task-specific, fine-tuned biomedical language representation models, improving patient care and management. For such specialized domains, previous research has shown that fine-tuning models stemming from broad-coverage checkpoints can largely benefit additional training rounds over large-scale in-domain resources. However, these resources are often unreachable for less-resourced languages like Italian, preventing local medical institutions to employ in-domain adaptation. In order to reduce this gap, our work investigates two accessible approaches to derive biomedical language models in languages other than English, taking Italian as a concrete use-case: one based on neural machine translation of English resources, favoring quantity over quality; the other based on a high-grade, narrow-scoped corpus natively written in Italian, thus preferring quality over quantity. Our study shows that data quantity is a harder constraint than data quality for biomedical adaptation, but the concatenation of high-quality data can improve model performance even when dealing with relatively size-limited corpora. The models published from our investigations have the potential to unlock important research opportunities for Italian hospitals and academia. Finally, the set of lessons learned from the study constitutes valuable insights towards a solution to build biomedical language models that are generalizable to other less-resourced languages and different domain settings.

Vietnamese Sentiment Analysis: An Overview and Comparative Study of Fine-tuning Pretrained Language Models

Sentiment Analysis (SA) is one of the most active research areas in the Natural Language Processing (NLP) field due to its potential for business and society. With the development of language representation models, numerous methods have shown promising efficiency in fine-tuning pre-trained language models in NLP downstream tasks. For Vietnamese, many available pre-trained language models were also released, including the monolingual and multilingual language models. Unfortunately, all of these models were trained on different architectures, pre-trained data, and pre-processing steps; consequently, fine-tuning these models can be expected to yield different effectiveness. In addition, there is no study focusing on evaluating the performance of these models on the same datasets for the SA task up to now. This article presents a fine-tuning approach to investigate the performance of different pre-trained language models for the Vietnamese SA task. The experimental results show the superior performance of the monolingual PhoBERT model and ViT5 model in comparison with previous studies and provide new state-of-the-art performances on five benchmark Vietnamese SA datasets. To the best of our knowledge, our study is the first attempt to investigate the performance of fine-tuning Transformer-based models on five datasets with different domains and sizes for the Vietnamese SA task.

Spam/Ham email classification using BERT

Email is a popular method for communicating with each other. However, as sending email is free of charge as long as an email server and a domain name are available, spam mail is becoming a critical problem in the email network. Conventionally, the industry uses spam filters based on rules and Bayesian inference to counteract spam mail, reaching an accuracy of 98.76%, which is far from satisfactory. Hence, to better protect email users from unsolicited messages containing advertisements, sensitive content, phishing content, and viruses, a new approach is proposed, in which email content is filtered by a spam detector using bidirectional encoder representations from transformers (BERT). BERT is a new language representation model published by Google that has achieved great success because of its powerful capabilities in understanding natural language. After the model is trained on a corpus from Kaggle, the spam detector equipped with the BERT model reaches a binary accuracy of 99.40% when classifying spam mail.

An ALBERT-based TextCNN-Hatt hybrid model enhanced with topic knowledge for sentiment analysis of sudden-onset disasters

Sudden-onset disasters put forward new requirements for on the state authorities’ ability to analyze public opinion sentiment. However, traditional sentiment analysis methods ignore the contextual semantic relationships and out-of-vocabulary words, and their computational resource utilization is excessive compared to their expected accuracy. In this paper, an ALBERT-based model combined with a text convolution neural network, a hierarchical attention mechanism and the latent Dirichlet allocation is proposed to create a hybrid model enhanced with topic knowledge for sentiment analysis of sudden-onset disasters. Weibo text data from a rainstorm disaster in China are used to evaluate the model’s performance. Compared with the XLNet, DistilBERT and RoBERTa models, the experimental results demonstrate that the proposed approach is capable of achieving better performance by incorporating external topic knowledge into the language representation model to compensate for the limited vocabulary data.

MP70-07 HIGH-PERFORMING SCALABLE EXTRACTION OF ELEMENTS OF URINARY TRACT DILATION CLASSIFICATION USING NATURAL LANGUAGE PROCESSING ALGORITHMS FOR NEONATAL ULTRASOUND REPORTS

You have accessJournal of UrologyCME1 Apr 2023MP70-07 HIGH-PERFORMING SCALABLE EXTRACTION OF ELEMENTS OF URINARY TRACT DILATION CLASSIFICATION USING NATURAL LANGUAGE PROCESSING ALGORITHMS FOR NEONATAL ULTRASOUND REPORTS Hsin-Hsiao Wang, Andrew Zhang, Caleb Nelson, Anudeep Mukkamala, Michael Li, and Carlos Estrada Hsin-Hsiao WangHsin-Hsiao Wang More articles by this author , Andrew ZhangAndrew Zhang More articles by this author , Caleb NelsonCaleb Nelson More articles by this author , Anudeep MukkamalaAnudeep Mukkamala More articles by this author , Michael LiMichael Li More articles by this author , and Carlos EstradaCarlos Estrada More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003338.07AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: The urinary tract dilation (UTD) classification system provides objective assessment for hydronephrosis in young children. However, the lack of specificity regarding UTD categorization in radiology reports causes difficulty in both clinical management and research. We seek to extract UTD elements and classification from early postnatal ultrasound (US) reports using the cutting-edge natural language processing (NLP) machine learning algorithms. METHODS: Radiology records from our institution were reviewed to identify infants age 0-90 days undergoing early ultrasound for antenatal UTD. The report and images were reviewed by the study team physicians to create the ground truth of UTD classification and components (primary outcome). Data were split into training/testing sets by 85:15 ratio. Bidirectional Encoder Representations from Transformers (BERT) language representation models were used as the basis of classification model. The model was fine-tuned with a head consisting of a drop layer, a fully connected layer and a binary classification layer. The model performance was evaluated with out-of-sample testing set. RESULTS: 2500 early (0-90 days) US reports were included. The model performance for the out-of-sample testing dataset is very high (AUC>0.9 for all). Accuracy, F1 score and AUC for UTD features in out-of-sample testing dataset are listed in Table 1. CONCLUSIONS: By applying deep state-of-the-art NLP neural networks, we developed a high-performing and scalable solution to extract UTD components from unstructured ultrasound reports. This can potentially help reduce miscommunications and facilitate large-scale computer vision research for children with hydronephrosis. Source of Funding: Internal © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e1005 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Hsin-Hsiao Wang More articles by this author Andrew Zhang More articles by this author Caleb Nelson More articles by this author Anudeep Mukkamala More articles by this author Michael Li More articles by this author Carlos Estrada More articles by this author Expand All Advertisement PDF downloadLoading ...

Toward Prompt-Enhanced Sentiment Analysis with Mutual Describable Information Between Aspects

ABSTRACT Sentiment analysis aims to attain the sentiment polarity of the text, which is a coarse-grained approach and does not focus on the targets . On the other hand, an aspect-based sentiment analysis (ABSA) has recently gained boosting interest. The ABSA is a fine-grained sentiment assignment to determine the sentiment tendency toward a specific aspect. Most previous methods employ Recurrent Neural Network (RNN) coupled with attention mechanisms to accomplish this task. However, such RNN-based models tend to be complex and require much training. Recently, a growing number of BERT-style models have been emerging and presenting better results in ABSA tasks . Nevertheless, these methods cannot well distinguish the various logical relationships between aspects that exist in the data and thus do not model the relationship between aspects. In the manuscript, a prompt-enhanced sentiment analysis (PESA) is proposed. Hence, a novel and efficient approach could retrieve the training set that is most similar to the input text and represents the mutual information between aspects by utilizing the [MASK] token. Moreover, the proposed model only needs to forward once if a sentiment analysis of multiple aspects is required in the inference stage. The language representation model BERT is employed to boost the performance of the proposed method. Comprehensive experiments and conducted analysis indicate the efficiency of the proposed methodology.

Tuning Language Processing Approaches for Pashto Texts Classification

Nowadays, text classification for different purposes becomes a basic task for concerned people. Hence, much research has been done to develop automatic text classification for the majority of national and international languages. However, the need for an automated text classification system for local languages is felt. The main purpose of this study is to establish a novel automatic classification system of Pashto text. In order to follow this up, we established a collection of Pashto documents and constructed the dataset. In addition, this study includes several models containing statistical techniques and neural network neural machine learning including DistilBERT-base-multilingual-cased, Multilayer Perceptron, Support Vector Machine, K Nearest Neighbor, decision tree, Gaussian naïve Bayes, multinomial naïve Bayes, random forest, and logistic regression to discover the most effective approach. Moreover, this investigation evaluates two different feature extraction methods including bag of words, and Term Frequency Inverse Document Frequency. Subsequently, this research obtained an average testing accuracy rate of 94% using the MLP classification algorithm and TFIDF feature extraction method in single label multi-class classification. Similarly, MLP+TFIDF with F1-measure of 0.81 showed the best result. Experiments on the use of pre-trained language representation models (such as DistilBERT) for classifying Pashto texts show that we need a specific tokenizer for a particular language to obtain reasonable results.

Can language representation models think in bets?

In recent years, transformer-based language representation models (LRMs) have achieved state-of-the-art results on difficult natural language understanding problems, such as question answering and text summarization. As these models are integrated into real-world applications, evaluating their ability to make rational decisions is an important research agenda, with practical ramifications. This article investigates LRMs' rational decision-making ability through a carefully designed set of decision-making benchmarks and experiments. Inspired by classic work in cognitive science, we model the decision-making problem as a bet. We then investigate an LRM's ability to choose outcomes that have optimal, or at minimum, positive expected gain. Through a robust body of experiments on four established LRMs, we show that a model is able to 'think in bets' if it is first fine-tuned on bet questions with an identical structure. Modifying the bet question's structure, while still retaining its fundamental characteristics, decreases an LRM's performance by more than 25%, on average, although absolute performance remains well above random. LRMs are also found to be more rational when selecting outcomes with non-negative expected gain, rather than optimal or strictly positive expected gain. Our results suggest that LRMs could potentially be applied to tasks that rely on cognitive decision-making skills, but that more research is necessary before these models can robustly make rational decisions.