This paper describes the Tekio submission to the AmericasNLP 2025 shared task on machine translation metrics for Indigenous languages. We developed two primary metric approaches leveraging multilingual semantic embeddings. First, we fine-tuned the Language-agnostic BERT Sentence Encoder (LaBSE) specifically for Guarani, Bribri, and Nahuatl, significantly enhancing semantic representation quality. Next, we integrated our fine-tuned LaBSE into the semantic similarity metric YiSi-1, exploring the effectiveness of averaging multiple layers. Additionally, we trained regression-based COMET metrics (COMET-DA) using the fine-tuned LaBSE embeddings as a semantic backbone, comparing Mean Absolute Error (MAE) and Mean Squared Error (MSE) loss functions. Our YiSi-1 metric using layer-averaged embeddings chosen by having the best performance on the development set for each individual language achieved the highest average correlation across languages among our submitted systems, and our COMET models demonstrated competitive performance for Guarani.

This paper describes the NAIST submission to the AmericasNLP 2025 shared task on the creation of educational materials for Indigenous languages. We implement three systems to tackle the unique challenges of each language. The first system, used for Maya and Guarani, employs a straightforward GPT-4o few-shot prompting technique, enhanced by synthetically generated examples to ensure coverage of all grammatical variations encountered. The second system, used for Bribri, integrates dictionary-based alignment and linguistic rules to systematically manage linguisticand lexical transformations. Finally, we developed a specialized rule-based system for Nahuatl that systematically reduces sentences to their base form, simplifying the generation of correct morphology variants.

Advancements in dialogue systems powered by large language models (LLMs) have outpaced the development of reliable evaluation metrics, particularly for diverse and creative responses. We present a benchmark for evaluating the robustness of reference-free dialogue metrics against four categories of adversarial attacks: speaker tag prefixes, static responses, ungrammatical responses, and repeated conversational context. We analyze metrics such as DialogRPT, UniEval, and PromptEval—a prompt-based method leveraging LLMs—across grounded and ungrounded datasets. By examining both their correlation with human judgment and susceptibility to adversarial attacks, we find that these two axes are not always aligned; metrics that appear to be equivalent when judged by traditional benchmarks may, in fact, vary in their scores of adversarial responses. These findings motivate the development of nuanced evaluation frameworks to address real-world dialogue challenges.

Word frequency is a key variable in psycholinguistics, useful for modeling human familiarity with words even in the era of large language models (LLMs). Frequency in film subtitles has proved to be a particularly good approximation of everyday language exposure. For many languages, however, film subtitles are not easily available, or are overwhelmingly translated from English. We demonstrate that frequencies extracted from carefully processed YouTube subtitles provide an approximation comparable to, and often better than, the best currently available resources. Moreover, they are available for languages for which a high-quality subtitle or speech corpus does not exist. We use YouTube subtitles to construct frequency norms for five diverse languages, Chinese, English, Indonesian, Japanese, and Spanish, and evaluate their correlation with lexical decision time, word familiarity, and lexical complexity. In addition to being strongly correlated with two psycholinguistic variables, a simple linear regression on the new frequencies achieves a new high score on a lexical complexity prediction task in English and Japanese, surpassing both models trained on film subtitle frequencies and the LLM GPT-4. We publicly release our code, the frequency lists, fastText word embeddings, and statistical language models.

The grammatical knowledge of language models (LMs) is often measured using a benchmark of linguistic minimal pairs, where LMs are presented with a pair of acceptable and unacceptable sentences and required to judge which is more acceptable. Conventional approaches compare sentence probabilities directly, but large language models (LLMs) provide nuanced evaluation methods using prompts and templates. We therefore investigate how to derive the most accurate acceptability judgments from LLMs to comprehensively evaluate their grammatical knowledge. Through extensive experiments in both English and Chinese, we compare nine judgment methods and demonstrate that two of them, in-template LP (a probability readout method) and Yes/No probability computing (a prompting-based method), achieve higher accuracy than the conventional approach. Our analysis reveals that the top two methods excel in different linguistic phenomena, suggesting they access different aspects of the LLMs’ grammatical knowledge. We find that ensembling the two methods achieves even higher accuracy. Consequently, we recommend these techniques, either individually or ensembled, as more effective alternatives to conventional approaches for assessing grammatical knowledge in LLMs.

This paper presents our approach to the AmericasNLP 2024 Shared Task 2 as the JAJ (/dʒæz/) team. The task aimed at creating educational materials for indigenous languages, and we focused on Maya and Bribri. Given the unique linguistic features and challenges of these languages, and the limited size of the training datasets, we developed a hybrid methodology combining rule-based NLP methods with prompt-based techniques. This approach leverages the meta-linguistic capabilities of large language models, enabling us to blend broad, language-agnostic processing with customized solutions. Our approach lays a foundational framework that can be expanded to other indigenous languages languages in future work.

In various natural language processing tasks, such as named entity recognition and machine translation, example-based approaches have been used to improve performance by leveraging existing knowledge. However, the effectiveness of this approach for Grammatical Error Correction (GEC) is unclear. In this work, we explore how an example-based approach affects the accuracy and interpretability of the output of GEC systems and the trade-offs involved. The approach we investigate has shown great promise in machine translation by using the $k$-nearest translation examples to improve the results of a pretrained Transformer model. We find that using this technique increases precision by reducing the number of false positives, but recall suffers as the model becomes more conservative overall. Increasing the number of example sentences in the datastore does lead to better performing systems, but with diminishing returns and a high decoding cost. Synthetic data can be used as examples, but the effectiveness varies depending on the base model. Finally, we find that finetuning on a set of data may be more effective than using that data during decoding as examples.

This paper presents our approach to the BEA 2023 shared task of generating teacher responses in educational dialogues, using the Teacher-Student Chatroom Corpus. Our system prompts GPT-3.5-turbo to generate initial suggestions, which are then subjected to reranking. We explore multiple strategies for candidate generation, including prompting for multiple candidates and employing iterative few-shot prompts with negative examples. We aggregate all candidate responses and rerank them based on DialogRPT scores. To handle consecutive turns in the dialogue data, we divide the task of generating teacher utterances into two components: teacher replies to the student and teacher continuations of previously sent messages. Through our proposed methodology, our system achieved the top score on both automated metrics and human evaluation, surpassing the reference human teachers on the latter.

In this paper, we describe our NAIST-NICT submission to the WMT’23 English ↔ Japanese general machine translation task. Our system generates diverse translation candidates and reranks them using a two-stage reranking system to find the best translation. First, we generated 50 candidates each from 18 translation methods using a variety of techniques to increase the diversity of the translation candidates. We trained seven models per language direction using various combinations of hyperparameters. From these models we used various decoding algorithms, ensembling the models, and using kNN-MT (Khandelwal et al., 2021). We processed the 900 translation candidates through a two-stage reranking system to find the most promising candidate. In the first step, we compared 50 candidates from each translation method using DrNMT (Lee et al., 2021) and returned the candidate with the best score. We ranked the final 18 candidates using COMET-MBR (Fernandes et al., 2022) and returned the best score as the system output. We found that generating diverse translation candidates improved translation quality using the well-designed reranker model.

In this paper, we describe our NAIST-NICT-TIT submission to the WMT22 general machine translation task. We participated in this task for the English ↔ Japanese language pair. Our system is characterized as an ensemble of Transformer big models, k-nearest-neighbor machine translation (kNN-MT) (Khandelwal et al., 2021), and reranking.In our translation system, we construct the datastore for kNN-MT from back-translated monolingual data and integrate kNN-MT into the ensemble model. We designed a reranking system to select a translation from the n-best translation candidates generated by the translation system. We also use a context-aware model to improve the document-level consistency of the translation.