HippoRAG, 그래프를 활용한 장기 기억
Hipporag: Neurobiologically inspired long-term memory for large language models, NeurIPS, 2024. From RAG to Memory: Non-Parametric Continual Learning for Large Language Models, arXiv, 2025. RAG...
Palette
Think-on-Graph, 그래프를 활용한 정보 탐색
Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation, ICLR 2025, arXiv. 복잡한 질문에 답하려면 흩어진 단서를 모으고, 단서끼리의 관계를 따라가야 해요. 사람은 이 과정을...
그룹을 이용한 개인화된 추천 학습
Self-Attentive Sequential Recommendation, ICDM 2018, arXiv. Efficient and Effective Adaptation of Multimodal Foundation Models in Sequential Recommendation, TKDE 2025, arXiv. Personalized Param...
Language Model은 잠을 자야 할까
Mamba: Linear-Time Sequence Modeling with Selective State Spaces, 2024, arXiv. Linear Transformers Are Secretly Fast Weight Programmers, ICML 2021, arXiv. Do Language Models Need Sleep? Offline...
Graph Neural Network 알아보기
GNN(Graph Neural Network)의 대표 모델인 GCN, GraphSAGE, GAT 세 논문을 따라가며, 그래프 위에서 노드 표현을 학습하는 방법이 어떻게 발전했는지 살펴봐요. Semi-Supervised Classification with Graph Convolutional Networks, ICLR 2017, arXiv. I...
SAM에서 Efficient Track Anything까지
Segment Anything, ICCV 2023. SAM 2: Segment Anything in Images and Videos, ICLR 2025. EfficientViT-SAM: Accelerated Segment Anything Model Without Accuracy Loss, CVPR 2024. Efficient Track An...
토큰을 출력할수록 정보가 손실된다면
Latent Collaboration in Multi-Agent Systems, ICML 2026(Spotlight), arXiv. Recursive Multi-Agent Systems, preprint 2026, arXiv. Multi-Agent System(MAS)의 동작을 떠올려봐요. Planner 에이전트가 “이 문제는 두 단계로 풀어야...