Hasil untuk "Language acquisition"

E. Bates

Elinor Ochs

Shaharukh Khan, Ali Faraz, Abhinav Ravi et al.

Multimodal research has predominantly focused on single-image reasoning, with limited exploration of multi-image scenarios. Recent models have sought to enhance multi-image understanding through large-scale pretraining on interleaved image-text datasets. However, most Vision-Language Models (VLMs) are trained primarily on English datasets, leading to inadequate representation of Indian languages. To address this gap, we introduce the Chitrakshara dataset series, covering 11 Indian languages sourced from Common Crawl. It comprises (1) Chitrakshara-IL, a large-scale interleaved pretraining dataset with 193M images, 30B text tokens, and 50M multilingual documents, and (2) Chitrakshara-Cap, which includes 44M image-text pairs with 733M tokens. This paper details the data collection pipeline, including curation, filtering, and processing methodologies. Additionally, we present a comprehensive quality and diversity analysis to assess the dataset's representativeness across Indic languages and its potential for developing more culturally inclusive VLMs.

Eneko Valero, Maria Ribalta i Albado, Oscar Sainz et al.

Large Language Models (LLMs) remain heavily centered on English, with limited performance in low-resource languages. Existing adaptation approaches, such as continual pre-training, demand significant computational resources. In the case of instructed models, high-quality instruction data is also required, both of which are often inaccessible for low-resource language communities. Under these constraints, model merging offers a lightweight alternative, but its potential in low-resource contexts has not been systematically explored. In this work, we explore whether it is possible to transfer language knowledge to an instruction-tuned LLM by merging it with a language-specific base model, thereby eliminating the need of language-specific instructions and repeated fine-tuning processes whenever stronger instructed variants become available. Through experiments covering four Iberian languages (Basque, Catalan, Galician, and Spanish) and two model families, we show that merging enables effective instruction following behavior in new languages and even supports multilingual capability through the combination of multiple language-specific models. Our results indicate that model merging is a viable and efficient alternative to traditional adaptation methods for low-resource languages, achieving competitive performance while greatly reducing computational cost.

Jinghan Cao, Yu Ma, Xinjin Li et al.

Large Language Models achieve remarkable performance but incur substantial computational costs unsuitable for resource-constrained deployments. This paper presents the first comprehensive task-specific efficiency analysis comparing 16 language models across five diverse NLP tasks. We introduce the Performance-Efficiency Ratio (PER), a novel metric integrating accuracy, throughput, memory, and latency through geometric mean normalization. Our systematic evaluation reveals that small models (0.5--3B parameters) achieve superior PER scores across all given tasks. These findings establish quantitative foundations for deploying small models in production environments prioritizing inference efficiency over marginal accuracy gains.

P. Myles Eugenio, Anthony Beavers

We introduce a novel paradigm of emergent local memory. It is a continuous-learning completely-parallel content-addressable memory encoding global order. It demonstrates how local constraints on uncoordinated learning can produce topologically protected memories realizing emergent symbolic order. It is therefore a neuro-symbolic bridge. It further has the ability to produce human language without data, by exploiting its own self-organizing dynamics. It teaches us that words arise as a side-effect of emergent symbolic order, and that human language patterns at all structural levels reflect a universal mechanism of word formation (which is subregular). This work answers essential questions about the existence \& origin of all the human language data.

Tsan Tsai Chan, Xin Tong, Thi Thu Uyen Hoang et al.

Multilingual large language models (LLMs) are known to more frequently generate non-faithful output in resource-constrained languages (Guerreiro et al., 2023 - arXiv:2303.16104), potentially because these typologically diverse languages are underrepresented in their training data. To mitigate unfaithfulness in such settings, we propose using computationally light auxiliary models to rescore the outputs of larger architectures. As proof of the feasibility of such an approach, we show that monolingual 4-layer BERT models pretrained from scratch on less than 700 MB of data without fine-tuning are able to identify faithful summaries with a mean accuracy of 88.33% in three genetically unrelated languages that differ in their morphological complexity - Vietnamese, Polish and Georgian. The same hyperparameter combination moreover generalises well to three other tasks, suggesting applications for rescoring beyond improving faithfulness. In order to inform typologically aware model selection, we also investigate how morphological complexity interacts with regularisation, model depth and training objectives, ultimately demonstrating that morphologically complex languages are more likely to benefit from dropout, while across languages downstream performance is enhanced most by shallow architectures as well as training using the standard BERT objectives.

Anh Nguyen, Stefan Lee

To interact effectively with humans in the real world, it is important for agents to understand language that describes the dynamics of the environment--that is, how the environment behaves--rather than just task instructions specifying "what to do". Understanding this dynamics-descriptive language is important for human-agent interaction and agent behavior. Recent work address this problem using a model-based approach: language is incorporated into a world model, which is then used to learn a behavior policy. However, these existing methods either do not demonstrate policy generalization to unseen games or rely on limiting assumptions. For instance, assuming that the latency induced by inference-time planning is tolerable for the target task or expert demonstrations are available. Expanding on this line of research, we focus on improving policy generalization from a language-conditioned world model while dropping these assumptions. We propose a model-based reinforcement learning approach, where a language-conditioned world model is trained through interaction with the environment, and a policy is learned from this model--without planning or expert demonstrations. Our method proposes Language-aware Encoder for Dreamer World Model (LED-WM) built on top of DreamerV3. LED-WM features an observation encoder that uses an attention mechanism to explicitly ground language descriptions to entities in the observation. We show that policies trained with LED-WM generalize more effectively to unseen games described by novel dynamics and language compared to other baselines in several settings in two environments: MESSENGER and MESSENGER-WM.To highlight how the policy can leverage the trained world model before real-world deployment, we demonstrate the policy can be improved through fine-tuning on synthetic test trajectories generated by the world model.

Elias-Leander Ahlers, Witold Brunsmann, Malte Schilling

Assessing language proficiency is essential for education, as it enables instruction tailored to learners needs. This paper investigates the use of Large Language Models (LLMs) for automatically classifying German texts according to the Common European Framework of Reference for Languages (CEFR) into different proficiency levels. To support robust training and evaluation, we construct a diverse dataset by combining multiple existing CEFR-annotated corpora with synthetic data. We then evaluate prompt-engineering strategies, fine-tuning of a LLaMA-3-8B-Instruct model and a probing-based approach that utilizes the internal neural state of the LLM for classification. Our results show a consistent performance improvement over prior methods, highlighting the potential of LLMs for reliable and scalable CEFR classification.

Hailay Tesfay Gebremariam

Although, written corrective feedback (hereafter referred to as CF) is applauded in many writing courses for fostering students’ quality writing, its impact on grammatical accuracy in L2 students’ writing remains a debated topic. Thus, this study looked into the effect of CF types on L2 students’ grammatical accuracy in writing. To achieve this objective, the design of this study was quasi-experiment. During the intervention of the study three groups: two experimental groups and one control group were participated with a total of 150 students. Over the intervention of 8 weeks, the students received pretest, immediate post-test and a delayed post-test was given. The data collected through the writing tests was analyzed using one-way ANOVA and Scheffe post hoc tests. The findings indicated that although CF types have positive effects during the immediate posttest scores, they did not have any positive effect on L2 students’ grammatical accuracy in writing context in the delayed posttest scores. This implies that CF alone is not sufficient for improving the grammatical accuracy of high school students in Ethiopia. Although the CF literature discussed its importance in the language acquisition, teachers are advised to focus on students’ additional exposures in writing accuracy rather than grammatical correction in their writing classes and use longer treatment to allow language learners’ engagement with the CF types provided.

Jennifer Hu, Michael C. Frank

Developmental psychologists have argued about when cognitive capacities such as language understanding or theory of mind emerge. These debates often hinge on the concept of "task demands" -- the auxiliary challenges associated with performing a particular evaluation -- that may mask the child's underlying ability. The same issues arise when measuring the capacities of language models (LMs): performance on a task is a function of the model's underlying knowledge, combined with the model's ability to interpret and perform the task given its available resources. Here, we show that for analogical reasoning, reflective reasoning, word prediction, and grammaticality judgments, evaluation methods with greater task demands yield lower performance than evaluations with reduced demands. This "demand gap" is most pronounced for models with fewer parameters and less training data. Our results illustrate that LM performance should not be interpreted as a direct indication of intelligence (or lack thereof), but as a reflection of capacities seen through the lens of researchers' design choices.

Andrew Shin, Kunitake Kaneko

Large language models (LLMs) have demonstrated remarkable performances on a wide range of natural language tasks. Yet, LLMs' successes have been largely restricted to tasks concerning words, sentences, or documents, and it remains questionable how much they understand the minimal units of text, namely characters. In this paper, we examine contemporary LLMs regarding their ability to understand character composition of words, and show that most of them fail to reliably carry out even the simple tasks that can be handled by humans with perfection. We analyze their behaviors with comparison to token level performances, and discuss the potential directions for future research.

Hellina Hailu Nigatu, Zeerak Talat

Mainstream Natural Language Processing (NLP) research has ignored the majority of the world's languages. In moving from excluding the majority of the world's languages to blindly adopting what we make for English, we first risk importing the same harms we have at best mitigated and at least measured for English. However, in evaluating and mitigating harms arising from adopting new technologies into such contexts, we often disregard (1) the actual community needs of Language Technologies, and (2) biases and fairness issues within the context of the communities. In this extended abstract, we consider fairness, bias, and inclusion in Language Technologies through the lens of the Capabilities Approach. The Capabilities Approach centers on what people are capable of achieving, given their intersectional social, political, and economic contexts instead of what resources are (theoretically) available to them. We detail the Capabilities Approach, its relationship to multilingual and multicultural evaluation, and how the framework affords meaningful collaboration with community members in defining and measuring the harms of Language Technologies.

Yiming Ai, Zhiwei He, Ziyin Zhang et al.

In this study, we delve into the validity of conventional personality questionnaires in capturing the human-like personality traits of Large Language Models (LLMs). Our objective is to assess the congruence between the personality traits LLMs claim to possess and their demonstrated tendencies in real-world scenarios. By conducting an extensive examination of LLM outputs against observed human response patterns, we aim to understand the disjunction between self-knowledge and action in LLMs.

M. Guasti

Magdalena Szyszka

The qualitative study presented in this paper aimed to collect beliefs about learning and teaching English as a foreign language (EFL) from individual representatives of the generation frequently referred to as the millennials. The participants were 47 pre-service EFL trainee teachers from four socio-cultural contexts: Finnish, Israeli, Polish, and Spanish. Their voices have been considered because beliefs are dynamically related to actions and soon the millennial EFL teachers may implement them in the course of their teaching. The contextual approach, followed in this research, provided opportunities for discussing similarities and differences in the beliefs of Finnish, Israeli, Polish, and Spanish pre-service teachers. The identified similarities lead to outlining a tentative picture of a universal, future, post-pandemic EFL classroom.

Marina Bondi, Matteo Di Cristofaro

The article discusses the on-going process for the creation of the MoReThesisCorpus, outlining its major characteristics and offering an account of the considerations and issues involved so far. The corpus, composed of the theses submitted to the University of Modena and Reggio Emilia between 2011 and 2020, is being developed as part of the project CAP (‘Comunicazione Accademica e Professionale;’ Academic and Professional Communication), and is meant to foster research into academic language in a cross-disciplinary discourse perspective, as well as to facilitate the production of educational materials aimed at university students. It aims at supporting the acquisition of discipline-related vocabularies and styles to improve the learning of academic writing through corpus tools and resources, following a data-driven learning approach. Technical details surrounding the acquisition and subsequent processing of the data are discussed, along with considerations on a number of issues pertaining both to computer science and linguistics, directly impinging on the capability of the corpus to correctly support an investigation of academic discourse across different languages and disciplines.

Amel Jardak, Casey Lew-Williams, Krista Byers-Heinlein et al.

Bilingual children regularly hear sentences that contain words from both languages, also known as code-switching. Investigating how bilinguals process code-switching is important for understanding bilingual language acquisition, because young bilinguals have been shown to experience processing costs and reduced comprehension when encountering code-switched nouns. Studies have yet to inves-tigate if processing costs are present when children encounter code-switches at other parts of speech within a sentence. The current study examined how 30 young bilinguals (age range: 37 – 48 months) processed sentences with code-switches at an uninformative determiner-adjective pair before the target noun (e.g., “Can you find le bon [the good] duck?) compared to single-language sentences (e.g., “Can you find the good duck?”). Surprisingly, bilingual children accurately identified the target object in both sentence types, contrasting with previous findings that sentences containing code-switching lead to processing difficulties. Indeed, children showed similar (and in some cases, better) comprehension of sentences with a code-switch at an uninformative adjective phrase, relative to single-language sentenc-es. We conclude that functional information conveyed by a code-switch may contribute to bilingual children’s sentence processing.

Bertaria Sohnata Hutauruk, Runi Fazalani, Darul Ilmi et al.

The acquisition of children's language at the phonological level in psycholinguistic studies is very interesting to be discussed by experts and researchers. The study of language acquisition in the field of phonology is important in the field of phonology in children. This research was conducted to analyze children's language acquisition. This research is qualitative because it is appropriate to systematically, factually, and accurately describe language acquisition. This type of research is descriptive qualitative. The subjects in this study were children aged 2 to 4 years. The data collection method in this study is observation. Observation or direct observation of the research object is intended to get a clear picture of the existence of the research object and the activities carried out. After the data is collected, a discussion is carried out using the distribution method. The data analysis technique uses descriptive qualitative analysis. The data were analyzed based on the forms and functions in the language of children aged 2-4 years. The study results are that children's language acquisition at the age of 2-4 years is different for each child, but this is considered reasonable because the language acquisition that occurs in each child is not the same. Children aged 2-4 years can use the sound of pronouncing words and sentences in consonant acquisition; children aged 2-4 years can pronounce vowels; while in acquiring syntax, children aged 2-4 years can use the best words and sentences.

Theodore R. Sumers, Shunyu Yao, Karthik Narasimhan et al.

Recent efforts have augmented large language models (LLMs) with external resources (e.g., the Internet) or internal control flows (e.g., prompt chaining) for tasks requiring grounding or reasoning, leading to a new class of language agents. While these agents have achieved substantial empirical success, we lack a systematic framework to organize existing agents and plan future developments. In this paper, we draw on the rich history of cognitive science and symbolic artificial intelligence to propose Cognitive Architectures for Language Agents (CoALA). CoALA describes a language agent with modular memory components, a structured action space to interact with internal memory and external environments, and a generalized decision-making process to choose actions. We use CoALA to retrospectively survey and organize a large body of recent work, and prospectively identify actionable directions towards more capable agents. Taken together, CoALA contextualizes today's language agents within the broader history of AI and outlines a path towards language-based general intelligence.