오늘 주제는 ViT Distillation입니다. 그런데 Singular Nullspace를 곁들인. SiNGER: Clearer Voice Distills ViT Further

발표일: 2026.03.05

발표자: 유근혁

제목: SiNGER : A Clearer Voice Distills Vision Transformers Further

게재: ICLR 2026

발표 자료: https://docs.google.com/presentation/d/1JWNuQ43BL9FueuEg6A_a-38ZN_xQ1Yff/edit?usp=sharing&ouid=118148277292201825613&rtpof=true&sd=true

AIRLab: http://airlab.khu.ac.kr/

#seminar #AIRLAB #ViT #transformer #distillation

Видео 오늘 주제는 ViT Distillation입니다. 그런데 Singular Nullspace를 곁들인. SiNGER: Clearer Voice Distills ViT Further канала AIRLab

Комментарии отсутствуют

Информация о видео

9 марта 2026 г. 8:00:48

00:14:40

AIRLab

Правообладателям

Жалоба на материал Недопустимый материал Нарушение авторских прав

Комментарии

Другие видео канала

너에겐 점이지만 나에겐 선이야📐📏SOFT2: Stereo Visual Odometry Based on a Point-to-Epipolar-Line Metric(T-RO 2022)

오일러!🤜 라그랑지안!🤛 크로스!!!🤝 ELPINN: Eulerian Lagrangian Physics-Informed Neural Network

반사된 정보도 여기에 써먹어?!😮😮Intensity Assisted Automatic Extrinsic Calibration of LiDAR and Panoramic Camera

이젠 로봇 관절마저 렌더링 해버림 ㄷㄷ 🩻🩻 Differentiable Robot Rendering (CoRL 2024 Oral)

[AIRLAB@KHU 논문세미나] Denoising Diffusion Probabilistic Models

오늘 주제는 ViT Distillation입니다. 그런데 Singular Nullspace를 곁들인. SiNGER: Clearer Voice Distills ViT Further

너에겐 점이지만 나에겐 선이야📐📏SOFT2: Stereo Visual Odometry Based on a Point-to-Epipolar-Line Metric(T-RO 2022)

오일러!🤜 라그랑지안!🤛 크로스!!!🤝 ELPINN: Eulerian Lagrangian Physics-Informed Neural Network

반사된 정보도 여기에 써먹어?!😮😮Intensity Assisted Automatic Extrinsic Calibration of LiDAR and Panoramic Camera

이젠 로봇 관절마저 렌더링 해버림 ㄷㄷ 🩻🩻 Differentiable Robot Rendering (CoRL 2024 Oral)

[AIRLAB@KHU 논문세미나] Denoising Diffusion Probabilistic Models

[AIRLAB@KHU 논문세미나] POCO: Point Convolution for Surface Reconstruction

싱글샷→풀가구 실화냐🪑SINGAPO: Single Image Controlled Generation of Articulated Parts in Objects(ICLR 2025)

[AIRLAB@KHU 논문세미나] Point Cloud Upsampling via Disentangled Refinement

[AIRLAB@KHU 논문세미나] CVPR (2022) Unified Transformer Tracker for Object Tracking

네트워크를 뒤집어놓으셔따..!🫢 SegNeXt: Rethinking Convolution Attention for Semantic Segmentation (NIPS 2022)

[AIRLAB@KHU 논문세미나] End-to-End Semi-Supervised Object Detection with Soft Teacher

나도 봐줘..ㅠ😿 Cut and Learn for Unsupervised Object Detection and Instance Segmentation (CVPR 2023)

3시간 걸릴 data를 10분만에 모아서 가능👍 MimicPlay: Long-Horizon Imitation Learning by Watching Human Play

자, 잘 봐. 이게 화살표라는거야↗️↙️↘️⬇️↙️. 이대로만 하면 돼💫 Robotic Visual Instruction (CVPR 2025)

다 덮어버릴 거야 😝 Voronoi Multi-Robot Exploration in Unknown Environments via Reinforcement (TVT 2020)

[AIRLab X HYPER] The First Public Seminar

Diffusion? "느려..." CARP: Visuomotor Policy Learning via Coarse-to-Fine AutoRegressive Prediction

눈 두 개 있는 파운데이션 모델 등장. (두둥) 👁️👁️FoundationStereo: Zero-Shot Stereo Matching(CVPR 2025)

이게 왜 개..?😵‍💫 Domain Adaptation via Prompt Learning (arXiv 2022)

어느 세월에 그걸 다 하고 있니??⏱️ Accelerating Diffusion Models via Early Stop of the Diffusion Process