paper-with-me

Papers

Self-Supervised Alignment with Mutual Information: Learning to Follow Principles without Preference Labels

2024-04-22 · Jan-Philipp Fränken, Eric Zelikman, Rafael Rafailov, Kanishk Gandhi, Tobias Gerstenberg, Noah D. Goodman

When prompting a language model (LM), users often expect the model to adhere to a set of behavioral principles across diverse tasks, such as producing insightful content while avoiding harmful or biased language. Instilling such principles (i.e., a constitution) into a model is resource-intensive, technically challenging, and generally requires human preference labels or examples. We introduce SAMI, an iterative algorithm that finetunes a pretrained language model (without requiring preference labels or demonstrations) to increase the conditional mutual information between constitutions and self-generated responses given queries from a dataset. On single-turn dialogue and summarization, a SAMI-trained mistral-7b outperforms the initial pretrained model, with win rates between 66% and 77%. Strikingly, it also surpasses an instruction-finetuned baseline (mistral-7b-instruct) with win rates between 55% and 57% on single-turn dialogue. SAMI requires a model that writes the principles. To avoid dependence on strong models for writing principles, we align a strong pretrained model (mixtral-8x7b) using constitutions written by a weak instruction-finetuned model (mistral-7b-instruct), achieving a 65% win rate on summarization. Finally, we investigate whether SAMI generalizes to diverse summarization principles (e.g., "summaries should be scientific") and scales to stronger models (llama3-70b), finding that it achieves win rates of up to 68% for learned and 67% for held-out principles compared to the base model. Our results show that a pretrained LM can learn to follow constitutions without using preference labels, demonstrations, or human oversight.

📄 PDF Abstract BibTeX arXiv:2404.14313

Code (1)

janphilippfranken/sami 공식 구현 pytorch

Tasks

Language ModelingLanguage Modelling

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically
BASE 설명 없음
ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…

Similar Papers 제목 키워드 기반

Mutual- and Self- Prototype Alignment for Semi-supervised Medical Image Segmentation

2022-06-03 · Zhenxi Zhang, Chunna Tian, Zhicheng Jiao

Semi-supervised learning methods have been explored in medical image segmentation tasks due to the scarcity of pixel-level annotation in the real scenario. Proto-type alignment based consistency constraint is an intuitio…

Image SegmentationMedical Image SegmentationSegmentationSemantic Segmentation+1

Self-MI: Efficient Multimodal Fusion via Self-Supervised Multi-Task Learning with Auxiliary Mutual Information Maximization

2023-11-07 · Cam-Van Thi Nguyen, Ngoc-Hoa Thi Nguyen, Duc-Trong Le, Quang-Thuy Ha

Multimodal representation learning poses significant challenges in capturing informative and distinct features from multiple modalities. Existing methods often struggle to exploit the unique characteristics of each modal…

Multi-Task LearningRepresentation LearningSelf-Supervised Learning

An Exploration of Self-Supervised Mutual Information Alignment for Multi-Task Settings

2024-10-02 · Soham Govande

There is a growing need for pluralistic alignment methods that can steer language models towards individual attributes and preferences. One such method, Self-Supervised Alignment with Mutual Information (SAMI), uses cond…

8kMath

RFMI: Estimating Mutual Information on Rectified Flow for Text-to-Image Alignment

2025-03-18 · Chao Wang, Giulio Franzese, Alessandro Finamore, Pietro Michiardi

Rectified Flow (RF) models trained with a Flow matching framework have achieved state-of-the-art performance on Text-to-Image (T2I) conditional generation. Yet, multiple benchmarks show that synthetic images can still su…

Attribute

miCSE: Mutual Information Contrastive Learning for Low-shot Sentence Embeddings

2022-11-09 · Tassilo Klein, Moin Nabi

This paper presents miCSE, a mutual information-based contrastive learning framework that significantly advances the state-of-the-art in few-shot sentence embedding. The proposed approach imposes alignment between the at…

Contrastive LearningFew-Shot LearningSelf-Supervised LearningSentence+3