Massey Documents by Type

Permanent URI for this communityhttps://mro.massey.ac.nz/handle/10179/294

Browse

Filters

2024 - 20252

Settings

Has files: NoSubject: search.filters.subject.Data processing

Search Results

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Deep learning for low-resource machine translation : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science, School of Mathematical and Computational Sciences, Massey University, Albany, Auckland, New Zealand. EMBARGOED until further notice.
    (Massey University, 2025-09-01) Gao, Yuan
    Machine translation, a key task in natural language processing, aims to automatically translate text from one language to another while preserving semantic integrity. This thesis builds upon existing research and introduces three deep-learning methods to enhance translation performance under low-resource conditions: (i) an effective transfer learning framework that leverages knowledge from high-resource language pairs, (ii) a pre-ordering-aware training method that explicitly utilizes contextualized representations of pre-ordered sentences, and (iii) a data augmentation strategy that expands the training data size. Firstly, we develop a two-step fine-tuning (TSFT) transfer learning framework for low-resource machine translation. Due to the inherent linguistic divergence between languages in parent (high-resource language pairs) and child (low-resource language pairs) translation tasks, the parent model often serves as a suboptimal initialization point for directly fine-tuning the child model. Our TSFT framework addresses this limitation by incorporating a pre-fine-tuning stage that adapts the parent model to the child source language characteristics, improving child model initialization and overall translation quality. Secondly, we propose a training method that enables the model to learn pre-ordering knowledge and encode the word reordering information within the contextualized representation of source sentences. Pre-ordering refers to rearranging source-side words to better align with the target-side word order before translation, which helps mitigate word-order differences between languages. Existing methods typically integrate the information of pre-ordered source sentences at the token level, where each token is assigned a local representation that fails to capture broader contextual dependencies. Moreover, these methods still require pre-ordered sentences during inference, which incur additional inference costs. In contrast, our method enables the model to encode the pre-ordering information in the contextualized representations of source sentences. In addition, our method eliminates the need for pre-ordering sentences at inference time while preserving its benefits in improving translation quality. Thirdly, to address data scarcity in low-resource scenarios, we propose a data augmentation strategy that employs high-quality translation models trained bidirectionally on high-resource language pairs. This strategy generates diverse, high-fidelity pseudo-training data through systematic sentence rephrasing, generating multiple target translations for each source sentence.. The increased diversity on the target side enhances the model's robustness, as demonstrated by significant performance improvements in eight pairs of low-resource languages. Finally, we conduct an empirical study to explore the potential of applying ChatGPT for machine translation. We design a set of translation prompts incorporating various auxiliary information to assist ChatGPT in generating translations. Our findings indicate that, with carefully designed prompts, ChatGPT can achieve results comparable to those of commercial translation systems for high-resource languages. Moreover, this study establishes a foundation for future research, offering insights into prompt engineering strategies for leveraging large language models in machine translation tasks.
  • Thumbnail Image
    Item
    The multimodality of creaminess perception : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Food Technology at Massey University, Manawatū Campus, Palmerston North, New Zealand. EMBARGOED until 21 August 2026.
    (Massey University, 2024-02-28) Fisher, Emily Claire
    Creaminess is a complex sensory sensation that drives consumer acceptability of milk. To date, creaminess research has focused on instrumental and compositional measures overlooking the critical consumer perspective. This research took a consumer-led approach to unlock new insights into the underlying sensory attributes driving consumer creaminess perception using perceptual modelling. Robust sensory data, from a trained panel, was combined with consumer approaches for accurate modelling. Initially, attributes and modalities perceived to drive milk creaminess were identified through discussion with consumers representative of two key dairy markets, China and New Zealand (NZ). Subsequently, a milk sample set (n=32) was developed, and an expert panel trained to profile the samples based on attributes identified by consumers. A novel methodological investigation, on the impact of panel training with Polarised Sensory Positioning (PSP) of the sample set, was also explored. Focusing on NZ consumers, participants (n=117) evaluated creaminess and liking perception of the milk samples. Critically, regression modelling was employed to identify key attributes driving creaminess perception based on expert panel data. Several novel findings were discovered. Drivers of creaminess differed to some degree between NZ and Chinese consumers indicating cultural differences across markets. Trained panel sensory data revealed multicollinearity between attributes measured to describe the sample set. Modelling approaches were able to identify key attributes required to predict creaminess. New findings that training has little impact on PSP outcomes was also ascertained. Pertinently combining four attributes, across different modalities, in an Elastic net regression model (‘yellow’, ‘watery’ flavour, ‘in-mouth thickness’ and ‘astringency’) successfully predicted creaminess (R2=0.9514), however these attributes were highly correlated with others retained in a PLS model. Each model had its relative merits. Of further note, consumer creaminess response was highly variable and cluster analysis revealed two different consumer segments with perception impacted by sensitivity to certain attributes: ‘green tinge’, ‘cardboard’, ‘salty’, ‘cooked’, ‘fat separation’, ‘grassy’, ‘buttery’, ‘melting’, ‘cream’ aroma, ‘smoothness’, and ‘astringent’. This research revealed new understanding concerning perceptual attributes contributing to consumer creaminess perception and provided clearer targets for the dairy industry to ensure milk creaminess levels align to consumer expectations and related commercial gain.