Paper Reviews by AI
2024
Fourier Position Embedding: Enhancing Attention's Periodic Extension for Length Generalization
·1717 words·9 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Tsinghua University
FoPE: μ£Όνμ μμ νΉμ§ κ°μ μΌλ‘ κΈ΄ λ¬Έλ§₯ κΈΈμ΄ μΌλ°ν λ¬μ±!
DRT-o1: Optimized Deep Reasoning Translation via Long Chain-of-Thought
·366 words·2 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Machine Translation
π’ Tencent AI Lab
DRT-01 λͺ¨λΈμ μ₯λ¬Έμ μ¬κ³ κ³Όμ μ νμ©νμ¬ λ¬Έν λ²μμ μ νλμ μ μ°½μ±μ ν¬κ² ν₯μμμΌ°μ΅λλ€.
Diving into Self-Evolving Training for Multimodal Reasoning
·2584 words·13 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Hong Kong University of Science and Technology
M-STAR: λ€λͺ¨λ¬ μΆλ‘ μ μν μκΈ° μ§ν νλ ¨μ μλ‘μ΄ νλ μμν¬λ₯Ό μ μ!
Deliberation in Latent Space via Differentiable Cache Augmentation
·2751 words·13 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Google DeepMind
λκ·λͺ¨ μΈμ΄ λͺ¨λΈμ μΆλ‘ μ±λ₯μ ν₯μμν€λ μλ‘μ΄ λ°©λ²μΈ βμ°¨λ³ κ°λ₯ν μΊμ μ¦κ°β κΈ°λ² μ μ!
B-STaR: Monitoring and Balancing Exploration and Exploitation in Self-Taught Reasoners
·1797 words·9 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Hong Kong University of Science and Technology
B-STAR: μκΈ° νμ΅ μΆλ‘ μμμ νμκ³Ό νμ©μ κ· νμ λͺ¨λν°λ§νκ³ μ‘°μ νμ¬ μ±λ₯μ ν₯μμν€λ μλ‘μ΄ νλ μμν¬
Revisiting In-Context Learning with Long Context Language Models
·3818 words·18 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Google DeepMind
μ₯λ¬Έ 컨ν
μ€νΈ μΈμ΄ λͺ¨λΈμμ μ κ΅ν μν μ ν μ λ΅λ³΄λ€ 무μμ μνλ§μ΄ ICL μ±λ₯ ν₯μμ λ ν¨κ³Όμ μ΄λ©°, λ°μ΄ν° μ¦κ°μ ν΅ν΄ μ μμ μμ
μ±λ₯μ 5% ν₯μμμΌ°λ€λ λλΌμ΄ μ°κ΅¬ κ²°κ³Όλ₯Ό λ°ν!
OpenRFT: Adapting Reasoning Foundation Model for Domain-specific Tasks with Reinforcement Fine-Tuning
·1880 words·9 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Beijing Jiaotong University
OpenRFTλ μ νλ λλ©μΈ νΉμ λ°μ΄ν°λ₯Ό μ¬μ©νμ¬ μΌλ°μ μΈ μΆλ‘ λͺ¨λΈμ λ―ΈμΈ μ‘°μ νλ μλ‘μ΄ λ°©λ²μ μ μν©λλ€.
Distilled Decoding 1: One-step Sampling of Image Auto-regressive Models with Flow Matching
·3113 words·15 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Computer Vision
Image Generation
π’ Tsinghua University
λ¨μΌ λ¨κ³ μνλ§μΌλ‘ μ΄λ―Έμ§ μλ νκ· λͺ¨λΈ μλλ₯Ό νκΈ°μ μΌλ‘ ν₯μμν¨ μ¦λ₯ λμ½λ©(DD) κΈ°λ² μ μ!
NILE: Internal Consistency Alignment in Large Language Models
·2709 words·13 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Chinese University of Hong Kong
NILE νλ μμν¬λ LLMμ λ΄λΆ μ§μκ³Ό IFT λ°μ΄ν°μ
μ μΈκ³ μ§μ κ° μΌκ΄μ±μ λμ¬ LLM μ±λ₯μ μ΅λ 68.5%κΉμ§ ν₯μμν΅λλ€.
LearnLM: Improving Gemini for Learning
·3761 words·18 mins·
loading
·
loading
AI Generated
π€ Daily Papers
AI Applications
Education
π’ Google DeepMind
LearnLMμ κ΅μ‘μ λ§₯λ½μμ μμ±ν AIμ νλ€κ³ μ§(Pedagogy)λ₯Ό ν₯μμν¨ λͺ¨λΈμ
λλ€. κ΅μ¬λ κ°λ°μκ° μνλ νλ€κ³ μ§μ νΉμ±μ λͺ¨λΈμ μ£Όμ
νλ μλ‘μ΄ νλ μμν¬λ₯Ό ν΅ν΄ κΈ°μ‘΄ λͺ¨λΈλ³΄λ€ νμ΅ ν¨κ³Όλ₯Ό 31% ν₯μμμΌ°μ΅λλ€.
Toward Robust Hyper-Detailed Image Captioning: A Multiagent Approach and Dual Evaluation Metrics for Factuality and Coverage
·2414 words·12 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Computer Vision
Visual Question Answering
π’ Seoul National University
μ΄μ λ° μ΄λ―Έμ§ μΊ‘μ
μμ±μ νκ° λ¬Έμ ν΄κ²°μ μν΄, LLM-MLLM νμ
κΈ°λ°μ λ€μ€ μμ΄μ νΈ μμ€ν
(CapMAS)μ μ μνμ¬ μ¬μ€μ±κ³Ό ν¬κ΄μ±μ λμμ΅λλ€.
Multi-LLM Text Summarization
·2623 words·13 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Text Summarization
π’ UC Santa Cruz
λ€μμ κ±°λ μΈμ΄ λͺ¨λΈ(LLM)μ νμ©ν νμ μ μΈ μ₯λ¬Έ μμ½ νλ μμν¬κ° μ μλμ΄ μμ½ νμ§μ μ΅λ 3λ°° ν₯μμμΌ°μ΅λλ€!
MotiF: Making Text Count in Image Animation with Motion Focal Loss
·2819 words·14 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Computer Vision
Video Understanding
π’ Brown University
MotiF: μμ§μμ μ΄μ μ λ§μΆ μμ€ ν¨μλ‘ ν
μ€νΈ κΈ°λ° μ΄λ―Έμ§ μ λλ©μ΄μ
κ°μ
Ensembling Large Language Models with Process Reward-Guided Tree Search for Better Complex Reasoning
·4085 words·20 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Microsoft Research
λκ·λͺ¨ μΈμ΄ λͺ¨λΈλ€μ μμλΈμ ν΅ν΄ 볡μ‘ν μΆλ‘ λ¬Έμ λ₯Ό λμ± ν¨κ³Όμ μΌλ‘ ν΄κ²°νλ μλ‘μ΄ νλ μμν¬, LE-MCTSλ₯Ό μ μν©λλ€!
CLEAR: Conv-Like Linearization Revs Pre-Trained Diffusion Transformers Up
·3581 words·17 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Computer Vision
Image Generation
π’ National University of Singapore
CLEAR: μ ννλ μ΄ν
μ
μΌλ‘ κ³ ν΄μλ μ΄λ―Έμ§ μμ± μλλ₯Ό νκΈ°μ μΌλ‘ λμ΄λ€!
UIP2P: Unsupervised Instruction-based Image Editing via Cycle Edit Consistency
·2616 words·13 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Computer Vision
Image Generation
π’ ETH Zurich
λΉμ§λ νμ΅ κΈ°λ° μν νΈμ§ μΌκ΄μ±(CEC) νμ©, μ§μμ΄ κΈ°λ° μ΄λ―Έμ§ νΈμ§μ μλ‘μ΄ μ§νμ μ΄λ€!
TOMG-Bench: Evaluating LLMs on Text-based Open Molecule Generation
·3930 words·19 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Hong Kong Polytechnic University
TOMG-Bench: LLM κΈ°λ° μ€ν λΆμ μμ± λ²€μΉλ§ν¬ μ μ! 25κ° LLM νκ° λ° μλ‘μ΄ instruction tuning λ°μ΄ν°μ
OpenMolIns 곡κ°λ‘, μ€νμμ€ LLMμ μ±λ₯ ν₯μ λ° λΆμ λ°κ²¬μ μλ‘μ΄ κ°λ₯μ± μ μ!
Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis
·1340 words·7 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Multimodal Learning
Multimodal Generation
π’ University of Illinois Urbana-Champaign
κ³ νμ§ λΉλμ€-μ€λμ€ ν©μ±μ μν νμ μ μΈ λ€μ€ λͺ¨λ μ‘°μΈνΈ νμ΅ νλ μμν¬ MMAudio μ μ!
RobustFT: Robust Supervised Fine-tuning for Large Language Models under Noisy Response
·2295 words·11 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Peking University
ROBUSTFTλ μ‘μμ΄ ν¬ν¨λ μλ΅ μλμμ λκ·λͺ¨ μΈμ΄ λͺ¨λΈμ κ°κ±΄ν μ§λ νμ΅ λ―ΈμΈ μ‘°μ μ μν νλ μμν¬λ‘, μ‘μ κ°μ§ λ° μ¬λΌλ²¨λ§μ ν΅ν΄ νλ₯ μμ
μ±λ₯μ ν₯μμν΅λλ€.
ReMoE: Fully Differentiable Mixture-of-Experts with ReLU Routing
·4863 words·23 mins·
loading
·
loading
AI Generated
π€ Daily Papers
Natural Language Processing
Large Language Models
π’ Tsinghua University
ReLU λΌμ°ν
μ μ¬μ©νλ μμ λ―ΈλΆ κ°λ₯ν MoE μν€ν
μ² ReMoEλ₯Ό ν΅ν΄ λκ·λͺ¨ μΈμ΄ λͺ¨λΈμ νμ₯μ±κ³Ό ν¨μ¨μ±μ νκΈ°μ μΌλ‘ κ°μ νμ΅λλ€!