Data-efficient Targeted Token-level Preference Optimization for LLM-based Text-to-Speech (2026.acl-short)
Copied to clipboard
| Challenge: | Recent work has shown that text-to-speech (TTS) models generate contextaware pronunciations from raw text without morphological analysis. |
| Approach: | They propose a preference optimization algorithm that aligns text-to-speech (TTS) outputs with human feedback. |
| Outcome: | The proposed method improves the challenging Japanese pronunciation accuracy by 39% and reduces CER by 54%. |
Similar Papers
TLPO: Token-Level Policy Optimization for Mitigating Language Confusion in Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Prior mitigation approaches that optimize entire responses operate at the level of entire responses and can lead to unintended degradation of general model capabilities. |
| Approach: | They propose a fine-tuning framework to mitigate erroneous outputs by localizing and updating the policy at a granular level. |
| Outcome: | The proposed framework outperforms baselines on multiple multilingual LLMs across diverse languages while preserving task accuracy. |
Advancing Zero-shot Text-to-Speech Intelligibility across Diverse Domains via Preference Alignment (2025.acl-long)
Copied to clipboard
| Challenge: | Existing zero-shot text-to-speech systems struggle in challenging scenarios such as tongue twisters, repeated words, code-switching, and cross-lingual synthesis. |
| Approach: | They propose a dataset that leverages preference alignment techniques to improve performance . they also extend the Direct Preference Optimization framework to accommodate diverse TTS architectures . |
| Outcome: | The proposed dataset improves intelligibility, similarity, and audio quality for multiple models across domains. |
Selective Preference Optimization via Token-Level Reward Function Estimation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for maximizing preference optimization on all available tokens are noisy and inefficient. |
| Approach: | They propose a selective alignment strategy that centers on efficient key token selection without strong, fine-grained supervision signals. |
| Outcome: | The proposed strategy outperforms baseline methods on three benchmarks with up to 60% reduction in training hours. |
T-REG: Preference Optimization with Token-Level Reward Regularization (2025.acl-long)
Copied to clipboard
| Challenge: | Reinforcement learning from human feedback (RLHF) is a dominant approach for large language models to follow instructions and produce meaningful alignment. |
| Approach: | They propose a method that leverages human feedback to optimize large language models . they propose to use sequence-level and token-level rewards to optimize preference . |
| Outcome: | The proposed method outperforms baseline methods on Alpaca Eval 2 and Arena-Hard benchmarks. |
Comparing Bad Apples to Good Oranges Aligning Large Language Models via Joint Preference Optimization (2025.findings-acl)
Copied to clipboard
| Challenge: | Recent studies have shown that acquiring human preferences by comparing generations is not effective for large language models. |
| Approach: | They propose a preference optimization objective that elicits preferences jointly over the instruction-response pairs. |
| Outcome: | The proposed approach outperforms prior preference optimizations by 5.2% and 3.3% in summarization and open-ended dialogue datasets. |
Aligning Large Language Models via Fine-grained Supervision (2024.acl-short)
Copied to clipboard
| Challenge: | Pre-trained large-scale language models often generate biased or toxic text, misaligning with human intentions. |
| Approach: | They propose to use human feedback to improve LLM alignment by fine-grained token supervision . they ask annotators to edit less preferred responses to make them more favorable . |
| Outcome: | The proposed method improves LLM alignment by up to 5.1% in terms of win rate compared with the traditional model. |
AlignDistil: Token-Level Language Model Alignment as Adaptive Policy Distillation (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for LLM alignment optimize tokens using a sparse, response-level reward or preference annotation. |
| Approach: | They propose an RLHF-equivalent distillation method for token-level reward optimization that incorporates the reward learned by DPO into the RLHG objective and builds a token-based teacher distribution. |
| Outcome: | The proposed method bridges the accuracy gap between the reward from the DPO model and the pure reward model by building a contrastive DPO reward with a normal and a reverse DPO. |
A Deep Dive into the Trade-Offs of Parameter-Efficient Preference Alignment Techniques (2024.acl-long)
Copied to clipboard
| Challenge: | Large language models are pre-trained on trillions of tokens and instruction-tuned or aligned to specific preferences. |
| Approach: | They propose guidelines to help researchers perform more effective parameter-efficient LLM alignment. |
| Outcome: | The proposed methods outperform preference optimization and outperformed pre-trained models on three key axes. |
Binary Classifier Optimization for Large Language Model Alignment (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for aligning large language models rely on preference-based approaches that require both positive and negative feedback as a pair. |
| Approach: | They propose a binary classifier optimization technique that trains a classifier using only binary feedback and a reward shift technique which minimizes the DPO loss. |
| Outcome: | The proposed method performs on a paired preference dataset and on 'likert-5 scale annotation dataset' it consistently demonstrates effective and robust alignment across four base LLMs and three different datasets, showcasing the strength of the proposed technique. |
DPO Kernels: A Semantically-Aware, Kernel-Enhanced, and Divergence-Rich Paradigm for Direct Preference Optimization (2025.findings-acl)
Copied to clipboard
Amitava Das, Suranjana Trivedy, Danush Khanna, Yaswanth Narsupalli, Basab Ghosh, Rajarshi Roy, Gurpreet Singh, Vinija Jain, Vasu Sharma, Aishwarya Naresh Reganti, Aman Chadha
| Challenge: | Direct Preference Optimization (DPO) is a cornerstone for preference alignment but is constrained by fixed divergence measures and limited feature transformations. |
| Approach: | They propose a new enhancement of Direct Preference Optimization that integrates kernel methods to overcome these challenges. |
| Outcome: | The proposed model improves divergence measures and features by using kernels . the proposed model achieves state-of-the-art generalization in factuality, safety, reasoning, and instruction following . |