paper-with-me

Papers

Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation

2024-01-16 · Haoran Xu, Amr Sharaf, Yunmo Chen, Weiting Tan, Lingfeng Shen, Benjamin Van Durme, Kenton Murray, Young Jin Kim

Moderate-sized large language models (LLMs) -- those with 7B or 13B parameters -- exhibit promising machine translation (MT) performance. However, even the top-performing 13B LLM-based translation models, like ALMA, does not match the performance of state-of-the-art conventional encoder-decoder translation models or larger-scale LLMs such as GPT-4. In this study, we bridge this performance gap. We first assess the shortcomings of supervised fine-tuning for LLMs in the MT task, emphasizing the quality issues present in the reference data, despite being human-generated. Then, in contrast to SFT which mimics reference translations, we introduce Contrastive Preference Optimization (CPO), a novel approach that trains models to avoid generating adequate but not perfect translations. Applying CPO to ALMA models with only 22K parallel sentences and 12M parameters yields significant improvements. The resulting model, called ALMA-R, can match or exceed the performance of the WMT competition winners and GPT-4 on WMT'21, WMT'22 and WMT'23 test datasets.

📄 PDF Abstract BibTeX arXiv:2401.08417

Code (1)

fe1ixxu/alma 공식 구현 pytorch

Tasks

DecoderMachine TranslationTranslation

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Similar Papers 제목 키워드 기반

SONAR: Spectral-Contrastive Audio Residuals for Generalizable Deepfake Detection

2025-11-26 · Ido Nitzan HIdekel, Gal lifshitz, Khen Cohen, Dan Raviv arxiv

Deepfake (DF) audio detectors still struggle to generalize to out of distribution inputs. A central reason is spectral bias, the tendency of neural networks to learn low-frequency structure before high-frequency (HF) det…

DeepFake Detection

Aligning Visual Contrastive learning models via Preference Optimization

2024-11-12 · Amirabbas Afzali, Borna Khodabandeh, Ali Rasekh, Mahyar JafariNodeh 외

Contrastive learning models have demonstrated impressive abilities to capture semantic similarities by aligning representations in the embedding space. However, their performance can be limited by the quality of the trai…

Contrastive LearningFairness

Dual-Agent Co-Training for Health Coaching via Implicit Adversarial Preference Optimization

2026-05-07 · Da Long, Lingyi Fu, Diya Michelle Rao, Jasmine Ruales Carrera 외 arxiv

Motivational-interviewing-based health coaching is an effective approach for improving mental health and promoting healthy behavior change. However, the scarcity of trained human coaches and the high cost of coaching ser…

Anchored Preference Optimization and Contrastive Revisions: Addressing Underspecification in Alignment

2024-08-12 · Karel D'Oosterlinck, Winnie Xu, Chris Develder, Thomas Demeester 외

Large Language Models (LLMs) are often aligned using contrastive alignment objectives and preference pair datasets. The interaction between model, paired data, and objective makes alignment a complicated procedure, somet…

Contrastive Learning

DynamicPO: Dynamic Preference Optimization for Recommendation

2026-05-01 · Xingyu Hu, Kai Zhang, Jiancan Wu, Shuli Wang 외 arxiv

In large language model (LLM)-based recommendation systems, direct preference optimization (DPO) effectively aligns recommendations with user preferences, requiring multi-negative objective functions to leverage abundant…

Recommendation Systems