近年來，人工智能（AI），尤其是生成式AI，逐步受到重視。

本系列希望通過閱讀發表在國際會議的 paper，撰寫精簡內容，讓讀者了解近年來AI的研究方向。

本系列挑選了4個主題，每個主題各7篇，共計28篇 paper。

Retrieval

1. Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
2. Retrieval-Augmented Multimodal Language Modeling
3. Fine-grained Late-interaction Multi-modal Retrieval for Retrieval Augmented Visual Question Answering
4. Recommender Systems with Generative Retrieval
5. Recitation-Augmented Language Models
6. Benchmarking Large Language Models in Retrieval-Augmented Generation
7. Diversify Question Generation with Retrieval-Augmented Style Transfer

Reinforcement Learning (RL)

1. Reward Design with Language Models
2. Pre-Trained Language Models for Interactive Decision-Making
3. Guiding Pretraining in Reinforcement Learning with Large Language Models
4. Building Persona Consistent Dialogue Agents with Offline Reinforcement Learning
5. TRAVEL: Tag-Aware Conversational FAQ Retrieval via Reinforcement Learning
6. OntoFact: Unveiling Fantastic Fact-Skeleton of LLMs via Ontology-Driven Reinforcement Learning
7. ESRL: Efficient Sampling-Based Reinforcement Learning for Sequence Generation

Multi-modal

1. RLEG: Vision-Language Representation Learning with Diffusion-based Embedding Generation
2. Learning to Answer Questions in Dynamic Audio-Visual Scenarios
3. Evaluating Object Hallucination in Large Vision-Language Models
4. Compressing and Debiasing Vision-Language Pre-Trained Models for Visual Question Answering
5. Visual Instruction Tuning
6. ILLUME: Rationalizing Vision-Language Models through Human Interactions
7. PaLM-E: An Embodied Multimodal Language Model

Graph

1. MolCA: Molecular graph-language modeling with cross- modal projector and uni-modal adapter
2. Harnessing explanations: Llm-to-lm interpreter for enhanced text-attributed graph representation learning
3. Reasoning on graphs: Faithful and interpretable large language model reasoning
4. Graph of Thoughts: Solving Elaborate Problems with Large Language Models
5. StructGPT: A General Framework for Large Language Model to Reason over Structured Data
6. Graph Neural Prompting with Large Language Models
7. Talk like a Graph: Encoding Graphs for Large Language Models

接下來每天會以以下格式撰寫

• 摘要
• Background survey
• 發表方法
• 實驗
• 額外資料補充