Machine Translation Digest for Sep 06 2025
Here is today's selection of cs.CL papers exploring advancements in aligning language models and optimizing multilingual systems. The common themes include enhancing model alignment through preference optimization and contrastive learning, as well as innovations in conversational AI and quantum word embeddings for NLP.
LatinX: Aligning a Multilingual TTS Model with Direct Preference Optimization
We present LatinX, a multilingual text-to-speech (TTS) model for cascaded speech-to-speech translation that preserves the source speaker's identity across languages. LatinX is a 12-layer decoder-only Transformer trained in three stages: (i) pre-training for text-to-audio mapping, (ii) supervised fine-tuning for zero-shot voice cloning, and (iii) alignment with Direct Preference Optimization (DPO) using automatically labeled pairs based on Word Error Rate (WER) and speaker-similarity metrics. Trained on English and Romance languages with emphasis on Portuguese, LatinX with DPO consistently reduces WER and improves objective similarity over the fine-tuned baseline. Human evaluations further indicate stronger perceived speaker similarity than a strong baseline (XTTSv2), revealing gaps between objective and subjective measures. We provide cross-lingual analyses and discuss balanced preference signals and lower-latency architectures as future work.
A Survey of the State-of-the-Art in Conversational Question Answering Systems
Conversational Question Answering (ConvQA) systems have emerged as a pivotal area within Natural Language Processing (NLP) by driving advancements that enable machines to engage in dynamic and context-aware conversations. These capabilities are increasingly being applied across various domains, i.e., customer support, education, legal, and healthcare where maintaining a coherent and relevant conversation is essential. Building on recent advancements, this survey provides a comprehensive analysis of the state-of-the-art in ConvQA. This survey begins by examining the core components of ConvQA systems, i.e., history selection, question understanding, and answer prediction, highlighting their interplay in ensuring coherence and relevance in multi-turn conversations. It further investigates the use of advanced machine learning techniques, including but not limited to, reinforcement learning, contrastive learning, and transfer learning to improve ConvQA accuracy and efficiency. The pivotal role of large language models, i.e., RoBERTa, GPT-4, Gemini 2.0 Flash, Mistral 7B, and LLaMA 3, is also explored, thereby showcasing their impact through data scalability and architectural advancements. Additionally, this survey presents a comprehensive analysis of key ConvQA datasets and concludes by outlining open research directions. Overall, this work offers a comprehensive overview of the ConvQA landscape and provides valuable insights to guide future advancements in the field.
QCSE: A Pretrained Quantum Context-Sensitive Word Embedding for Natural Language Processing
Quantum Natural Language Processing (QNLP) offers a novel approach to encoding and understanding the complexity of natural languages through the power of quantum computation. This paper presents a pretrained quantum context-sensitive embedding model, called QCSE, that captures context-sensitive word embeddings, leveraging the unique properties of quantum systems to learn contextual relationships in languages. The model introduces quantum-native context learning, enabling the utilization of quantum computers for linguistic tasks. Central to the proposed approach are innovative context matrix computation methods, designed to create unique, representations of words based on their surrounding linguistic context. Five distinct methods are proposed and tested for computing the context matrices, incorporating techniques such as exponential decay, sinusoidal modulation, phase shifts, and hash-based transformations. These methods ensure that the quantum embeddings retain context sensitivity, thereby making them suitable for downstream language tasks where the expressibility and properties of quantum systems are valuable resources. To evaluate the effectiveness of the model and the associated context matrix methods, evaluations are conducted on both a Fulani corpus, a low-resource African language, dataset of small size and an English corpus of slightly larger size. The results demonstrate that QCSE not only captures context sensitivity but also leverages the expressibility of quantum systems for representing rich, context-aware language information. The use of Fulani further highlights the potential of QNLP to mitigate the problem of lack of data for this category of languages. This work underscores the power of quantum computation in natural language processing (NLP) and opens new avenues for applying QNLP to real-world linguistic challenges across various tasks and domains.
Icon$^{2}$: Aligning Large Language Models Using Self-Synthetic Preference Data via Inherent Regulation
Large Language Models (LLMs) require high quality preference datasets to align with human preferences. However, conventional methods for constructing such datasets face significant challenges: reliance on pre-collected instructions often leads to distribution mismatches with target models, while the need for sampling multiple stochastic responses introduces substantial computational overhead. In this work, we explore a paradigm shift by leveraging inherent regulation of LLMs' representation space for efficient and tailored preference dataset construction, named Icon$^{2}$. Specifically, it first extracts layer-wise direction vectors to encode sophisticated human preferences and then uses these vectors to filter self-synthesized instructions based on their inherent consistency. During decoding, bidirectional inherent control is applied to steer token representations, enabling the precise generation of response pairs with clear alignment distinctions. Experimental results demonstrate significant improvements in both alignment and efficiency. Llama3-8B and Qwen2-7B achieve an average win rate improvement of 13.89% on AlpacaEval 2.0 and 13.45% on Arena-Hard, while reducing computational costs by up to 48.1%.
Using Contrastive Learning to Improve Two-Way Reasoning in Large Language Models: The Obfuscation Task as a Case Study
This research addresses a fundamental question in AI: whether large language models truly understand concepts or simply recognize patterns. The authors propose bidirectional reasoning,the ability to apply transformations in both directions without being explicitly trained on the reverse direction, as a test for genuine understanding. They argue that true comprehension should naturally allow reversibility. For example, a model that can change a variable name like userIndex to i should also be able to infer that i represents a user index without reverse training. The researchers tested current language models and discovered what they term cognitive specialization: when models are fine-tuned on forward tasks, their performance on those tasks improves, but their ability to reason bidirectionally becomes significantly worse. To address this issue, they developed Contrastive Fine-Tuning (CFT), which trains models using three types of examples: positive examples that maintain semantic meaning, negative examples with different semantics, and forward-direction obfuscation examples. This approach aims to develop deeper understanding rather than surface-level pattern recognition and allows reverse capabilities to develop naturally without explicit reverse training. Their experiments demonstrated that CFT successfully achieved bidirectional reasoning, enabling strong reverse performance while maintaining forward task capabilities. The authors conclude that bidirectional reasoning serves both as a theoretical framework for assessing genuine understanding and as a practical training approach for developing more capable AI systems.
| Curated by yukajii.com |