Paper Reviews by AI
2024
RetroLLM: Empowering Large Language Models to Retrieve Fine-grained Evidence within Generation
·3747 words·18 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Natural Language Processing
Question Answering
๐ข Renmin University of China
RetroLLM: ๊ฒ์๊ณผ ์์ฑ์ ํตํฉํ RAG ์์คํ
Nearly Zero-Cost Protection Against Mimicry by Personalized Diffusion Models
·3489 words·17 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
Image Generation
๐ข Inha University
์ค์๊ฐ ์ด๋ฏธ์ง ๋ณดํธ, ๋ฅํ์ดํฌ ๋๋น์ฑ
.
MOVIS: Enhancing Multi-Object Novel View Synthesis for Indoor Scenes
·2949 words·14 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
3D Vision
๐ข Peking University
MOVIS๋ ์ค๋ด ์ฅ๋ฉด์ ๋ํ ๋ฉํฐ-๊ฐ์ฒด novel view synthesis์์ ๊ตฌ์กฐ์ ์ธ์์ ํฅ์์์ผ ์ผ๊ด์ฑ ์๊ณ ์ฌ์ค์ ์ธ novel view๋ฅผ ์์ฑํฉ๋๋ค.
MaxInfoRL: Boosting exploration in reinforcement learning through information gain maximization
·2262 words·11 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Machine Learning
Reinforcement Learning
๐ข ETH Zurich
์ ๋ณด ์ด๋์ผ๋ก ๊ฐํ ํ์ต ํ์์ ๊ฐํ.
Just a Simple Transformation is Enough for Data Protection in Vertical Federated Learning
·2263 words·11 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Machine Learning
Federated Learning
๐ข MIPT
๊ฐ๋จํ ๋ณํ๋ง์ผ๋ก ์์ง ์ฐํฉ ํ์ต์์ ๋ฐ์ดํฐ ๋ณดํธ ๊ฐ๋ฅ.
IDArb: Intrinsic Decomposition for Arbitrary Number of Input Views and Illuminations
·3273 words·16 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
3D Vision
๐ข Chinese University of Hong Kong
IDArb: Decomposition under varied lights.
GeoX: Geometric Problem Solving Through Unified Formalized Vision-Language Pre-training
·2232 words·11 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Multimodal Learning
Vision-Language Models
๐ข Shanghai Jiao Tong University
GeoX: MLLM๋ณด๋ค ๋ฐ์ด๋ ๊ธฐํํ์ ๋ฌธ์ ํด๊ฒฐ์ฌ!
ColorFlow: Retrieval-Augmented Image Sequence Colorization
·2273 words·11 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
Image Generation
๐ข Tsinghua University
๋งํ ์ฑ์ ์๋ํ: ColorFlow๋ ID ์ผ๊ด์ฑ์ ์ ์งํ๋ฉด์ ํ๋ฐฑ ๋งํ ์ํ์ค๋ฅผ ์ฑ์ํฉ๋๋ค.
Causal Diffusion Transformers for Generative Modeling
·4953 words·24 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
Image Generation
๐ข ByteDance Research
CausalFusion์ ํ์ฐ ๋ฐ ์๊ธฐ ํ๊ท ๋ชจ๋ธ์ ๊ฒฐํฉํ์ฌ ์์ฑ ๋ชจ๋ธ๋ง์์ ์ต์ฒจ๋จ ๊ฒฐ๊ณผ๋ฅผ ๋ฌ์ฑํ๊ณ ์๋ก์ด ๊ธฐ๋ฅ์ ๊ฐ๋ฅํ๊ฒ ํฉ๋๋ค.
VividFace: A Diffusion-Based Hybrid Framework for High-Fidelity Video Face Swapping
·1707 words·9 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
Image Generation
๐ข CUHK MMLab
VividFace: ์ฒซ ๋ฒ์งธ ํ์ฐ ๊ธฐ๋ฐ ๋น๋์ค ์ผ๊ตด ๋ฐ๊พธ๊ธฐ ํ๋ ์์ํฌ๋ก ๊ณ ์ถฉ์ค๋ ๊ฒฐ๊ณผ ์ ๊ณต.
Smaller Language Models Are Better Instruction Evolvers
·4310 words·21 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Natural Language Processing
Large Language Models
๐ข Beijing University of Posts and Telecommunications
์ํ ์ธ์ด ๋ชจ๋ธ์ด ๋ ๋์ ๋ช
๋ น ์์ฑ์!
Reliable, Reproducible, and Really Fast Leaderboards with Evalica
·1243 words·6 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Natural Language Processing
Large Language Models
๐ข JetBrains
Evalica: ๋ฒค์น๋งํน์ ์ฝ๊ณ ๋น ๋ฅด๊ณ ์ ๋ขฐํ ์ ์๊ฒ ๋ง๋๋ ํดํท
GaussianProperty: Integrating Physical Properties to 3D Gaussians with LMMs
·2657 words·13 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
3D Vision
๐ข Hong Kong University of Science and Technology
GaussianProperty๋ LMM์ ์ฌ์ฉํ์ฌ 3D ๊ฐ์ฐ์์์ ๋ฌผ๋ฆฌ์ ์์ฑ์ ํตํฉํ๋ ํ๋ จ ์๋ ํ๋ ์์ํฌ๋ก, ๋ฌผ๋ฆฌ ๊ธฐ๋ฐ ์๋ฎฌ๋ ์ด์
๋ฐ ๋ก๋ด ์ฅ๊ธฐ์ ๊ฐ์ ๋ค์ด์คํธ๋ฆผ ์์
์ ๊ฐ๋ฅํ๊ฒ ํฉ๋๋ค.
DynamicScaler: Seamless and Scalable Video Generation for Panoramic Scenes
·1754 words·9 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
Image Generation
๐ข Google DeepMind
DynamicScaler๋ ํ
์คํธ๋ ์ด๋ฏธ์ง์์ ๊ธด ๋๊น ์๋ ํ๋
ธ๋ผ๋ง ๋น๋์ค๋ฅผ ์์ฑํ๋ฉฐ, ํด์๋์ ์ข
ํก๋น์ ๊ด๊ณ์์ด ์ผ๊ด๋ ์์ง์์ ์ ์งํฉ๋๋ค.
TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies
·2744 words·13 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
AI Applications
Robotics
๐ข Microsoft Research
TraceVLA: ๊ณผ๊ฑฐ์ ์์ง์์ ์๊ฐ์ ์ผ๋ก ๋ณด์ฌ์ค์ผ๋ก์จ ๋ก๋ด์ ์๊ณต๊ฐ์ ์ธ์์ ํฅ์์ํต๋๋ค.
SplineGS: Robust Motion-Adaptive Spline for Real-Time Dynamic 3D Gaussians from Monocular Video
·3662 words·18 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
3D Vision
๐ข KAIST
SplineGS: ์ค์๊ฐ ๋์ 3D ์ฅ๋ฉด์ ์ํ ๊ฐ๋ ฅํ ๋ชจ์
์ ์ํ ์คํ๋ผ์ธ.
SCBench: A KV Cache-Centric Analysis of Long-Context Methods
·4642 words·22 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Natural Language Processing
Large Language Models
๐ข Microsoft Corporation
SCBench๋ ๋ฉํฐํด ๋ฐ ๋ฉํฐ๋ฆฌํ์คํธ ์๋๋ฆฌ์ค์์ ์ฅ๋ฌธ ๋งฅ๋ฝ ๋ฉ์๋๋ฅผ ํ๊ฐํ๋ ์๋ก์ด ๋ฒค์น๋งํฌ์
๋๋ค.
RLDG: Robotic Generalist Policy Distillation via Reinforcement Learning
·1911 words·9 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
AI Applications
Robotics
๐ข UC Berkeley
RLDG๋ ๊ฐํ ํ์ต์ ํตํด ์์ฑ๋ ๊ณ ํ์ง ๋ฐ์ดํฐ๋ก ๋ฒ์ฉ ๋ก๋ด ์ ์ฑ
์ ์ฑ๋ฅ์ ํฅ์์ํค๋ ํ๊ธฐ์ ์ธ ๋ฐฉ๋ฒ์
๋๋ค.
Prompt2Perturb (P2P): Text-Guided Diffusion-Based Adversarial Attacks on Breast Ultrasound Images
·1580 words·8 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
Image Generation
๐ข University of British Columbia
P2P: ํ
์คํธ ๊ธฐ๋ฐ์ ์๋ก์ด ์ ๋์ ๊ณต๊ฒฉ์ผ๋ก ์๋ฃ ์์ DNN์ ์ทจ์ฝ์ฑ ๊ณต๋ต
LinGen: Towards High-Resolution Minute-Length Text-to-Video Generation with Linear Computational Complexity
·3571 words·17 mins·
loading
·
loading
AI Generated
๐ค Daily Papers
Computer Vision
Video Understanding
๐ข Princeton University
LinGen: ๋ถ ๋จ์ ๊ณ ํด์๋ ํ
์คํธ-ํฌ-๋น๋์ค ์์ฑ, ์ ํ ๊ณ์ฐ ๋ณต์ก๋๋ก ํจ์จ์ฑ ๊ทน๋ํ