Computation and Language (cs.CL)

Abstract: Over the last few years, large language models (LLMs) have emerged as the most important breakthroughs in natural language processing (NLP) that fundamentally transform research and developments in the field. ChatGPT represents one of the most exciting LLM systems developed recently to showcase impressive skills for language generation and highly attract public attention. Among various exciting applications discovered for ChatGPT in English, the model can process and generate texts for multiple languages due to its multilingual training data. Given the broad adoption of ChatGPT for English in different problems and areas, a natural question is whether ChatGPT can also be applied effectively for other languages or it is necessary to develop more language-specific technologies. The answer to this question requires a thorough evaluation of ChatGPT over multiple tasks with diverse languages and large datasets (i.e., beyond reported anecdotes), which is still missing or limited in current research. Our work aims to fill this gap for the evaluation of ChatGPT and similar LLMs to provide more comprehensive information for multilingual NLP applications. While this work will be an ongoing effort to include additional experiments in the future, our current paper evaluates ChatGPT on 7 different tasks, covering 37 diverse languages with high, medium, low, and extremely low resources. We also focus on the zero-shot learning setting for ChatGPT to improve reproducibility and better simulate the interactions of general users. Compared to the performance of previous models, our extensive experimental results demonstrate a worse performance of ChatGPT for different NLP tasks and languages, calling for further research to develop better models and understanding for multilingual learning.

Abstract: As a novel approach to tuning pre-trained models, prompt tuning involves freezing the parameters in downstream tasks while inserting trainable embeddings into inputs in the first layer.However,previous methods have mainly focused on the initialization of prompt embeddings. The question of how to train and utilize prompt embeddings in a reasonable way has become aa limiting factor in the effectiveness of prompt tuning. To address this issue, we introduce the Global Prompt Cell (GPC), a portable control module for prompt tuning that selectively preserves prompt information across all encoder layers. Our experimental results demonstrate a 5.8% improvement on SuperGLUE datasets compared to vanilla prompt tuning.

Abstract: We investigate in this paper how distributions of occupations with respect to gender is reflected in pre-trained language models. Such distributions are not always aligned to normative ideals, nor do they necessarily reflect a descriptive assessment of reality. In this paper, we introduce an approach for measuring to what degree pre-trained language models are aligned to normative and descriptive occupational distributions. To this end, we use official demographic information about gender--occupation distributions provided by the national statistics agencies of France, Norway, United Kingdom, and the United States. We manually generate template-based sentences combining gendered pronouns and nouns with occupations, and subsequently probe a selection of ten language models covering the English, French, and Norwegian languages. The scoring system we introduce in this work is language independent, and can be used on any combination of template-based sentences, occupations, and languages. The approach could also be extended to other dimensions of national census data and other demographic variables.

Abstract: Pre-trained language models have been known to perpetuate biases from the underlying datasets to downstream tasks. However, these findings are predominantly based on monolingual language models for English, whereas there are few investigative studies of biases encoded in language models for languages beyond English. In this paper, we fill this gap by analysing gender bias in West Slavic language models. We introduce the first template-based dataset in Czech, Polish, and Slovak for measuring gender bias towards male, female and non-binary subjects. We complete the sentences using both mono- and multilingual language models and assess their suitability for the masked language modelling objective. Next, we measure gender bias encoded in West Slavic language models by quantifying the toxicity and genderness of the generated words. We find that these language models produce hurtful completions that depend on the subject's gender. Perhaps surprisingly, Czech, Slovak, and Polish language models produce more hurtful completions with men as subjects, which, upon inspection, we find is due to completions being related to violence, death, and sickness.

Abstract: Few-shot dense retrieval (DR) aims to effectively generalize to novel search scenarios by learning a few samples. Despite its importance, there is little study on specialized datasets and standardized evaluation protocols. As a result, current methods often resort to random sampling from supervised datasets to create "few-data" setups and employ inconsistent training strategies during evaluations, which poses a challenge in accurately comparing recent progress. In this paper, we propose a customized FewDR dataset and a unified evaluation benchmark. Specifically, FewDR employs class-wise sampling to establish a standardized "few-shot" setting with finely-defined classes, reducing variability in multiple sampling rounds. Moreover, the dataset is disjointed into base and novel classes, allowing DR models to be continuously trained on ample data from base classes and a few samples in novel classes. This benchmark eliminates the risk of novel class leakage, providing a reliable estimation of the DR model's few-shot ability. Our extensive empirical results reveal that current state-of-the-art DR models still face challenges in the standard few-shot scene. Our code and data will be open-sourced at https://github.com/OpenMatch/ANCE-Tele.

Abstract: Homographs, words with the same spelling but different meanings, remain challenging in Neural Machine Translation (NMT). While recent works leverage various word embedding approaches to differentiate word sense in NMT, they do not focus on the pivotal components in resolving ambiguities of homographs in NMT: the hidden states of an encoder. In this paper, we propose a novel approach to tackle homographic issues of NMT in the latent space. We first train an encoder (aka "HDR-encoder") to learn universal sentence representations in a natural language inference (NLI) task. We further fine-tune the encoder using homograph-based synset sentences from WordNet, enabling it to learn word-level homographic disambiguation representations (HDR). The pre-trained HDR-encoder is subsequently integrated with a transformer-based NMT in various schemes to improve translation accuracy. Experiments on four translation directions demonstrate the effectiveness of the proposed method in enhancing the performance of NMT systems in the BLEU scores (up to +2.3 compared to a solid baseline). The effects can be verified by other metrics (F1, precision, and recall) of translation accuracy in an additional disambiguation task. Visualization methods like heatmaps, T-SNE and translation examples are also utilized to demonstrate the effects of the proposed method.

Abstract: Rehabilitation research focuses on determining the components of a treatment intervention, the mechanism of how these components lead to recovery and rehabilitation, and ultimately the optimal intervention strategies to maximize patients' physical, psychologic, and social functioning. Traditional randomized clinical trials that study and establish new interventions face several challenges, such as high cost and time commitment. Observational studies that use existing clinical data to observe the effect of an intervention have shown several advantages over RCTs. Electronic Health Records (EHRs) have become an increasingly important resource for conducting observational studies. To support these studies, we developed a clinical research datamart, called ReDWINE (Rehabilitation Datamart With Informatics iNfrastructure for rEsearch), that transforms the rehabilitation-related EHR data collected from the UPMC health care system to the Observational Health Data Sciences and Informatics (OHDSI) Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) to facilitate rehabilitation research. The standardized EHR data stored in ReDWINE will further reduce the time and effort required by investigators to pool, harmonize, clean, and analyze data from multiple sources, leading to more robust and comprehensive research findings. ReDWINE also includes deployment of data visualization and data analytics tools to facilitate cohort definition and clinical data analysis. These include among others the Open Health Natural Language Processing (OHNLP) toolkit, a high-throughput NLP pipeline, to provide text analytical capabilities at scale in ReDWINE. Using this comprehensive representation of patient data in ReDWINE for rehabilitation research will facilitate real-world evidence for health interventions and outcomes.

Abstract: Methods such as chain-of-thought prompting and self-consistency have pushed the frontier of language model reasoning performance with no additional training. To further improve performance, we propose a prompt ensembling method for large language models, which uses a small dataset to construct a set of few shot prompts that together comprise a ``boosted prompt ensemble''. The few shot examples for each prompt are chosen in a stepwise fashion to be ``hard'' examples on which the previous step's ensemble is uncertain. We show that this outperforms single-prompt output-space ensembles and bagged prompt-space ensembles on the GSM8k and AQuA datasets, among others. We propose both train-time and test-time versions of boosted prompting that use different levels of available annotation and conduct a detailed empirical study of our algorithm.