Language Research Articles

A Literature Review on Emotional Intelligence of Large Language Models (LLMs)

Large Language Models(LLMs) are artificial intelligence models that use deep neural networks to perform Natural Language Processing (NLP) tasks. These tasks include interaction between humans and computers, enabling computers to interpret and generate human languages in a meaningful manner. Large Language models are called "large" because of the architecture’s size and the huge sets of training text data. With the emergence of transformer-based LLMs, the game of NLPs has reached another level. This is due to their ability to handle long-range text dependencies in parallel. The growing prevalence of transformer-based LLMs in human lives has necessitated evaluating the scope of the Emotional Intelligence(EI) of LLMs. This paper will discuss the need for emotional intelligence in transformer-based LLMs and the various existing studies that have evaluated this aspect. The potential challenges of the LLMs along with the future directions for research in this field will also be discussed.

Deciphering 3'UTR Mediated Gene Regulation Using Interpretable Deep Representation Learning.

The 3' untranslated regions (3'UTRs) of messenger RNAs contain many important cis-regulatory elements that are under functional and evolutionary constraints. It is hypothesized that these constraints are similar to grammars and syntaxes in human languages and can be modeled by advanced natural language techniques such as Transformers, which has been very effective in modeling complex protein sequence and structures. Here 3UTRBERT is described, which implements an attention-based language model, i.e., Bidirectional Encoder Representations from Transformers (BERT). 3UTRBERT is pre-trained on aggregated 3'UTR sequences of human mRNAs in a task-agnostic manner; the pre-trained model is then fine-tuned for specific downstream tasks such as identifying RBP binding sites, m6A RNA modification sites, and predicting RNA sub-cellular localizations. Benchmark results show that 3UTRBERT generally outperformed other contemporary methods in each of these tasks. More importantly, the self-attention mechanism within 3UTRBERT allows direct visualization of the semantic relationship between sequence elements and effectively identifies regions with important regulatory potential. It is expected that 3UTRBERT model can serve as the foundational tool to analyze various sequence labeling tasks within the 3'UTR fields, thus enhancing the decipherability of post-transcriptional regulatorymechanisms.

Even laypeople use legalese

Whereas principles of communicative efficiency and legal doctrine dictate that laws be comprehensible to the common world, empirical evidence suggests legal documents are largely incomprehensible to lawyers and laypeople alike. Here, a corpus analysis (n = 59) million words) first replicated and extended prior work revealing laws to contain strikingly higher rates of complex syntactic structures relative to six baseline genres of English. Next, two preregistered text generation experiments (n = 286) tested two leading hypotheses regarding how these complex structures enter into legal documents in the first place. In line with the magic spell hypothesis, we found people tasked with writing official laws wrote in a more convoluted manner than when tasked with writing unofficial legal texts of equivalent conceptual complexity. Contrary to the copy-and-edit hypothesis, we did not find evidence that people editing a legal document wrote in a more convoluted manner than when writing the same document from scratch. From a cognitive perspective, these results suggest law to be a rare exception to the general tendency in human language toward communicative efficiency. In particular, these findings indicate law's complexity to be derived from its performativity, whereby low-frequency structures may be inserted to signal law's authoritative, world-state-altering nature, at the cost of increased processing demands on readers. From a law and policy perspective, these results suggest that the tension between the ubiquity and impenetrability of the law is not an inherent one, and that laws can be simplified without a loss or distortion of communicative content.

Natural language processing in dermatology: A systematic literature review and state of the art.

Natural Language Processing (NLP) is a field of both computational linguistics and artificial intelligence (AI) dedicated to analysis and interpretation of human language. This systematic review aims at exploring all the possible applications of NLP techniques in the dermatological setting. Extensive search on 'natural language processing' and 'dermatology' was performed on MEDLINE and Scopus electronic databases. Only journal articles with full text electronically available and English translation were considered. The PICO (Population, Intervention or exposure, Comparison, Outcome) algorithm was applied to our study protocol. Natural Language Processing (NLP) techniques have been utilized across various dermatological domains, including atopic dermatitis, acne/rosacea, skin infections, non-melanoma skin cancers (NMSCs), melanoma and skincare. There is versatility of NLP in data extraction from diverse sources such as electronic health records (EHRs), social media platforms and online forums. We found extensive utilization of NLP techniques across diverse dermatological domains, showcasing its potential in extracting valuable insights from various sources and informing diagnosis, treatment optimization, patient preferences and unmet needs in dermatological research and clinical practice. While NLP shows promise in enhancing dermatological research and clinical practice, challenges such as data quality, ambiguity, lack of standardization and privacy concerns necessitate careful consideration. Collaborative efforts between dermatologists, data scientists and ethicists are essential for addressing these challenges and maximizing the potential of NLP in dermatology.

The Acquisition of de: Investigating Recursive Possession in Mandarin-Speaking Kids

Recursive grammar is the sole cognitive mechanism unique to human languages. This study investigates three aspects of the acquisition of the recursive possessive marker de by Mandarin-speaking children: the feasibility of recursive possessive (RP) levels, performance in both cognition and competence, and the age at which these abilities emerge. Following a development theory of recursive possessives in language acquisition, we conducted two tasks—comprehension and production—each comprising 12 questions. The comprehension task involved a truth-value judgment, while the production task consisted of a question-and-answer session. Forty-eight preschoolers, aged three to five, were recruited from a kindergarten and divided into three age groups. Additionally, 16 adults were randomly recruited to serve as a baseline for comparison. The results revealed that significant effects of levels of recursive possessiveness were observed only in five-year-old Mandarin-speaking children, not in their three- and four-year-old counterparts. The comprehension task was significantly easier for all participants compared to the production task, suggesting that understanding the material was not challenging for any group, with no substantial differences noted among them. The results demonstrated that age effects were critical for younger children in adopting RPs. Specifically, three- and four-year-old Mandarin-speaking children demonstrated the ability to comprehend multiple levels of RP but struggled to produce them. In contrast, five-year-olds exhibited adult-like comprehension and the ability to produce Level 1 RP.

Language in Brains, Minds, and Machines.

It has long been argued that only humans could produce and understand language. But now, for the first time, artificial language models (LMs) achieve this feat. Here we survey the new purchase LMs are providing on the question of how language is implemented in the brain. We discuss why, a priori, LMs might be expected to share similarities with the human language system. We then summarize evidence that LMs represent linguistic information similarly enough to humans to enable relatively accurate brain encoding and decoding during language processing. Finally, we examine which LM properties-their architecture, task performance, or training-are critical for capturing human neural responses to language and review studies using LMs as in silico model organisms for testing hypotheses about language. These ongoing investigations bring us closer to understanding the representations and processes that underlie our ability to comprehend sentences and express thoughts in language.

A transformer-based spelling error correction framework for Bangla and resource scarce Indic languages

Spelling error correction is the task of identifying and rectifying misspelled words in texts. It is a potential and active research topic in Natural Language Processing because of numerous applications in human language understanding. The phonetically or visually similar yet semantically distinct characters make it an arduous task in any language. Earlier efforts on spelling error correction in Bangla and resource-scarce Indic languages focused on rule-based, statistical, and machine learning-based methods which we found rather inefficient. In particular, machine learning-based approaches, which exhibit superior performance to rule-based and statistical methods, are ineffective as they correct each character regardless of its appropriateness. In this paper, we propose a novel detector-purificator-corrector framework, DPCSpell based on denoising transformers by addressing previous issues. In addition to that, we present a method for large-scale corpus creation from scratch which in turn resolves the resource limitation problem of any left-to-right scripted language. The empirical outcomes demonstrate the effectiveness of our approach, which outperforms previous state-of-the-art methods by attaining an exact match (EM) score of 94.78%, a precision score of 0.9487, a recall score of 0.9478, an f1 score of 0.948, an f0.5 score of 0.9483, and a modified accuracy (MA) score of 95.16% for Bangla spelling error correction. The models and corpus are publicly available at https://tinyurl.com/DPCSpell.

Natural Language Processing Pretraining Language Model for Computer Intelligent Recognition Technology

Computer intelligent recognition technology refers to the use of computer vision, Natural Language Processing (NLP), machine learning and other technologies to enable computers to recognize, analyze, understand and answer human language and behavior. The common applications of computer intelligent recognition technology include image recognition, NLP, face recognition, target tracking, and other fields. NLP is a field of computer science, which involves the interaction between computers and natural languages. NLP technology can be used to process, analyze and generate natural language data, such as text, voice and image. Common NLP technology applications include language translation, emotion analysis, text classification, speech recognition and question answering system. Language model is a machine learning model, which uses a large number of text data for training to learn language patterns and relationships in text data. Although the language model has made great progress in the past few years, it still faces some challenges, including: poor semantic understanding, confusion in multilingual processing, slow language processing and other shortcomings. Therefore, in order to optimize these shortcomings, this article would study the pre-training language model based on NLP technology, which aimed at using NLP technology to optimize and improve the performance of the language model, thus optimizing the computer intelligent recognition technology. The model had a higher language understanding ability and more accurate prediction ability. In addition, the model could learn language rules and structures by using a large number of corpus, so as to better understand natural language. Through experiments, it could be known that the data size and total computing time of the traditional Generative Pretrained Transformer-2 (GPT-2) language model were 10 GB and 97 hours respectively. The data size and total computing time of BERT (Bidirectional Encoder Representations from Transformer) were 12 GB and 86 hours respectively. The data size and total computing time of the pre-training language model based on NLP were 18 GB and 71 hours respectively. Obviously, the pre-training language model based on NLP had a larger data size and shorter computing time. The experimental results showed that the NLP technology could better optimize the language model and effectively improve its various capabilities. This article opened up a new development direction for computer intelligent recognition technology and provided excellent technical support for the development of language models.

Linguistic Trends and Foreign Language Teaching: The Case of French

Linguistics is generally defined as the scientific study of human language. Whereas language is considered as an instrument of communication through which human experience is analyzed differently. Meanwhile, the spread of linguistics theories is propelled by the desire to find better ways to learn and acquire foreign languages, such as French. Hence, in what ways have the development of linguistics theories contributed to language learning and acquisition? An attempt to answer this question was made and some linguistic theories were explained.

Across the boundary: The formalization of the interface between episodic memory and narrow syntax computation of human language

A growing number of studies indicate that the hippocampus plays an essential role in language processing as well as episodic memory. However, there is no consensus on how it is engaged in both domains and how it connects the two domains. From a theoretical perspective, this paper delves into the intricate relationship between episodic memory and the narrow syntax of human language. To be more specific, I focus on the functions of the hippocampus in event processing and propose that hippocampal phasic activity supports the cyclic interaction between episodic memory and narrow syntax. Through such cyclic interaction, the event in the episodic memory is assigned a linguistic format that can be communicated, while narrow syntax is provided with an interpretative engine that underlies reference. This hypothesis predicts that when episodic memory is impaired, corresponding abnormalities appear in linguistic reference.

The Morphological Tapestry: A Linguistic Exploration

This article introduces morphological typology, exploring the patterns and structures underlying word formation and grammatical encoding across languages. A systematic literature review examines fundamental concepts, including distinctions between analytic and synthetic languages, features of agglutinating, fusional, and inflectional morphologies, and phenomena like suppletion and polysynthetic structures. Readers gain insights into language classification based on morphological characteristics, challenging strict categorical distinctions and emphasizing the continuum across types. The study highlights the diversity and complexity of morphological systems. Suppletion, where stems are irregularly replaced in inflectional patterns, and polysynthetic languages, encoding entire sentences within single complex words, are explored in depth. This work offers a concise overview of morphological typology by synthesizing reputable sources, including books, journals, and online resources. It is an accessible resource for language learners, linguists, and anyone interested in understanding the intricate morphological underpinnings of human language.

Spatiotemporal whole-brain activity and functional connectivity of melodies recognition.

Music is a non-verbal human language, built on logical, hierarchical structures, that offers excellent opportunities to explore how the brain processes complex spatiotemporal auditory sequences. Using the high temporal resolution of magnetoencephalography, we investigated the unfolding brain dynamics of 70 participants during the recognition of previously memorized musical sequences compared to novel sequences matched in terms of entropy and information content. Measures of both whole-brain activity and functional connectivity revealed a widespread brain network underlying the recognition of the memorized auditory sequences, which comprised primary auditory cortex, superior temporal gyrus, insula, frontal operculum, cingulate gyrus, orbitofrontal cortex, basal ganglia, thalamus, and hippocampus. Furthermore, while the auditory cortex responded mainly to the first tones of the sequences, the activity of higher-order brain areas such as the cingulate gyrus, frontal operculum, hippocampus, and orbitofrontal cortex largely increased over time during the recognition of the memorized versus novel musical sequences. In conclusion, using a wide range of analytical techniques spanning from decoding to functional connectivity and building on previous works, our study provided new insights into the spatiotemporal whole-brain mechanisms for conscious recognition of auditory sequences.

Africa must create centres of educational excellence for innovation and development.

This paper explores the following development questions that perplex most Africans: "Why do African countries rely on foreign companies and foreign experts for almost all our development projects? Why can't we build our own roads, process our own food, and mine our own minerals, oil, and gas? Why don't we have world-class hospitals and industries? How can we have so much natural wealth and yet be so poor? Why do we invent so little?" The answer lies in our failure to implement idea number two. There are two major ideas in educational policy. Idea number one is the obligation to educate all children because it is their fundamental human right as enshrined in the 1948 United Nations Declaration of Human Rights. Idea number two is the strategy of establishing and sustaining world class schools and universities for the education of the most highly gifted and highly talented citizens. Developed countries deploy both ideas aggressively. Underdeveloped countries in Africa have not implemented idea number two. Countries that have deployed idea number two have at least one university ranked among the top 200 in the world. The presence of great universities (top 200) in a country is a 21st century indicator of the presence of high levels of innovation, technology, development and wealth in that country. According to the three major rankings of world universities (Shanghai-ARWU, THE, and QS-topuniversities.com), none of the world's top 100 great universities is in Africa. Although Africa was a pioneer among the continents in innovations such as human language, domestication of fire, making of tools, invention of agriculture, development of writing, and creation of great centers of learning in ancient times, it has fallen behind other continents over the last 500 years and it has been disrupted by enslavement and colonization, and the structural adjustment programs (SAPs) of the IMF and the World Bank. Ancient African centers of innovation included the Ancient City of Benin and Timbuktu in Western Africa, the Kingdom of Kush and ancient Egypt in northern Africa, Axum in Eastern Africa, Mapungubwe and Great Zimbabwe in Southern Africa, and the Kingdom of Kongo in central-Africa. An African renaissance will only occur when we implement idea number two by establishing world class schools and at least one great university per African country.

CAN ARTIFICIAL INTELLIGENCE PREFERENCES BE AN ALTERNATIVE TO HUMAN LINGUISTIC CHOICES? A MULTIDIMENSIONAL ANALYSIS OF RESEARCH ABSTRACTS OF ENGLISH LINGUISTICS

Background and Purpose: With the rapid advancements in generative AI, understanding its ability to emulate human language conventions is crucial. This work aims to analyze the possibility of applying AI technology in language production by comparing the lexico-grammatical features of abstracts created with the help of ChatGPT and written by British and American researchers. Methodology: Twenty papers written by researchers affiliated with UK universities and twenty by researchers affiliated with American universities were selected from the journals listed under the first quartile of the Web of Science and Scopus. Using the titles that were from the selected works, 40 abstracts were generated from ChatGPT for comparison. Each article was introduced with its title, and ChatGPT was asked to create an abstract based on the title. Subsequently, the subjects were examined with the help of Biber’s (1991) multivariate model considering five dimensions, which include the Informational vs Involved discourse, the Narrative vs Non-narrative, the Explicit vs. Situation- dependent discourse, Overt expression of argumentation/persuasion and the Impersonal/Abstract style as opposed to the Non-impersonal/Non-abstract style. Findings: The five factors analysed in the texts give evidence that ChatGPT generates more information-centric, non-narrative, argumentative, and less abstractive discourse than human researchers. Contributions: The results of the study show the possibilities for the further development of AI that helps to create language closer to human language. Keywords: Abstracts, ChatGPT, lexico-grammatical patterns, multidimensional analysis, research articles. Cite as: Ali, M., & Ali, S. (2024). Can artificial intelligence preferences be an alternative to human linguistic choices? A multidimensional analysis of research abstracts of English linguistics. Journal of Nusantara Studies, 9(2), 514-536. http://dx.doi.org/10.24200/jonus.vol9iss2pp514-536

What Formal Approaches to Syntactic Interfaces Can Tell Us about the Syntax of Preverbal and Prenominal Constituents in Galician

An important line of research within a generative, formal approach to syntax in the early 21st century has centered on exploring phenomena related to the interface between syntax and other linguistic modules in human language. In this paper, we review the notion of interfaces and how they have been viewed within formal theoretical approaches to monolingual and bilignual competence and language acquistion, noting their relevance as they relate to language acquisition and bilingualism in the context of Galicia (Spain). We review a selection of Noun Phrase (NP) structures that implicate a syntactic interface: subject position, clitic directionality, and determiner clitic allomorphy. We provide a review of the relevant literature and the theoretical issues of interest as they relate to our understanding of these syntactic interfaces, reporting on our current theoretical understandings, persistent questions, and our view of the path forward as it relates to linguistic research on the Galician language.

Evaluation of ChatGPT as a diagnostic tool for medical learners and clinicians.

ChatGPT is a large language model (LLM) trained on over 400 billion words from books, articles, and websites. Its extensive training draws from a large database of information, making it valuable as a diagnostic aid. Moreover, its capacity to comprehend and generate human language allows medical trainees to interact with it, enhancing its appeal as an educational resource. This study aims to investigate ChatGPT's diagnostic accuracy and utility in medical education. 150 Medscape case challenges (September 2021 to January 2023) were inputted into ChatGPT. The primary outcome was the number (%) of cases for which the answer given was correct. Secondary outcomes included diagnostic accuracy, cognitive load, and quality of medical information. A qualitative content analysis was also conducted to assess its responses. ChatGPT answered 49% (74/150) cases correctly. It had an overall accuracy of 74%, a precision of 48.67%, sensitivity of 48.67%, specificity of 82.89%, and an AUC of 0.66. Most answers were considered low cognitive load 51% (77/150) and most answers were complete and relevant 52% (78/150). ChatGPT in its current form is not accurate as a diagnostic tool. ChatGPT does not necessarily give factual correctness, despite the vast amount of information it was trained on. Based on our qualitative analysis, ChatGPT struggles with the interpretation of laboratory values, imaging results, and may overlook key information relevant to the diagnosis. However, it still offers utility as an educational tool. ChatGPT was generally correct in ruling out a specific differential diagnosis and providing reasonable next diagnostic steps. Additionally, answers were easy to understand, showcasing a potential benefit in simplifying complex concepts for medical learners. Our results should guide future research into harnessing ChatGPT's potential educational benefits, such as simplifying medical concepts and offering guidance on differential diagnoses and next steps.

Cognitive causes of ‘like me’ race and gender biases in human language production

Natural language contains and communicates social biases, often reflecting attitudes, prejudices and stereotypes. Here we provide evidence for a novel psychological pathway for the expression of such biases, in which they arise as a consequence of the automatized mechanisms by which humans retrieve words to produce sentences. Four experiments show that, when describing events, speakers tend to mention people who are more like them first and, thus, tend to highlight the perspectives of their own social groups. This ‘like me’ effect was seen in speakers from multiple demographic groups, in both English and Chinese speakers and in both first- and second-language English speakers. Psycholinguistic manipulations pinpoint that the bias is caused by greater accessibility in memory of words that refer to in-group than out-group members. These data provide a new cognitive explanation for why people produce biased language and highlight how detailed cognitive theories can have social implications.

Auftritt der KI. Zur Darstellung posthumaner Arbeitswelten in Sibylle Bergs <em>GRM Brainfuck</em>

The article deals with Sibylle Berg's novel GRM Brainfuck (2019), which focuses on a 'working world' that is being transformed by digitization processes. This world is anchored in a 'post-human system' in which non-human objects emerge as 'active' and humans recede into the background, which is why the text is examined against the background of critical post-humanist and neo-materialist approaches. In addition, media-theoretical and philosophical examinations form the framework for an analysis that first and foremost deals with the dystopian potential of the novel. It is then to be shown that the novel makes legible a gendered amalgamation of transhumanist ideology on the one hand and processes of reification on the other in a world 'populated' by artificial intelligences. Finally, it is argued that Berg's aesthetics unfolded within the narrative evoke what Günther Anders has called a "Promethean gap", meaning the inadequacy of human language and imagination with regard to man-made (technological) entities.

Electroencephalography Measurements in Awake Marmosets Listening to Conspecific Vocalizations.

Vocal communication plays a crucial role in the social interactions of primates, particularly in survival and social organization. Humans have developed a unique and advanced vocal communication strategy in the form of language. To study the evolution of human language, it is necessary to investigate the neural mechanisms underlying vocal processing in humans, as well as to understand how brain mechanisms have evolved by comparing them with those in nonhuman primates. Herein, we developed a method to noninvasively measure the electroencephalography (EEG) of awake nonhuman primates. This recording method allows for long-term studies without harming the animals, and, importantly, allows us to directly compare nonhuman primate EEG data with human data, providing insights into the evolution of human language. In the current study, we used the scalp EEG recording method to investigate brain activity in response to species-specific vocalizations in marmosets. This study provides novel insights by using scalp EEG to capture widespread neural representations in marmosets during vocal perception, filling gaps in existing knowledge.

Artificial intelligence in scientific medical writing: Legitimate and deceptive uses and ethical concerns

The debate surrounding the integration of artificial intelligence (AI) into scientific writing has already attracted significant interest in medical and life sciences. While AI can undoubtedly expedite the process of manuscript creation and correction, it raises several criticisms. The crossover between AI and health sciences is relatively recent, but the use of AI tools among physicians and other scientists who work in the life sciences is growing very fast. Within this whirlwind, it is becoming essential to realize where we are heading and what the limits are, including an ethical perspective.Modern conversational AIs exhibit a context awareness that enables them to understand and remember any conversation beyond any predefined script. Even more impressively, they can learn and adapt as they engage with a growing volume of human language input. They all share neural networks as background mathematical models and differ from old chatbots for their use of a specific network architecture called transformer model [1]. Some of them exceed 100 terabytes (TB) (e.g., Bloom, LaMDA) or even 500 TB (e.g., Megatron-Turing NLG) of text data, the 4.0 version of ChatGPT (GPT-4) was trained with nearly 45 TB, but stays updated by the internet connection and may integrate with different plugins that enhance its functionality, making it multimodal.