Timezone: »
Understanding temporal dynamics of video is an essential aspect of learning better video representations. Recently, transformer-based architectural designs have been extensively explored for video tasks due to their capability to capture long-term dependency of input sequences. However, we found that these Video Transformers are still biased to learn spatial dynamics rather than temporal ones, and debiasing the spurious correlation is critical for their performance. Based on the observations, we design simple yet effective self-supervised tasks for video models to learn temporal dynamics better. Specifically, for debiasing the spatial bias, our method learns the temporal order of video frames as extra self-supervision and enforces the randomly shuffled frames to have low-confidence outputs. Also, our method learns the temporal flow direction of video tokens among consecutive frames for enhancing the correlation toward temporal dynamics. Under various video action recognition tasks, we demonstrate the effectiveness of our method and its compatibility with state-of-the-art Video Transformers.
Author Information
Sukmin Yun (KAIST)
Jaehyung Kim (KAIST)
Dongyoon Han (NAVER AI Lab)
Hwanjun Song (NAVER AI Lab)
Jung-Woo Ha (NAVER AI Lab)

Jung-Woo Ha got his BS and PhD degrees in computer science from Seoul National University in 2004 and 2015. He got the 2014 Fall semester outstanding PhD dissertation award from Computer Science Dept. of Seoul National University. He worked as a research scientist and tech lead at NAVER LABS and research head of NAVER CLOVA. Currently, he works as the head of NAVER AI Lab in NAVER Cloud. He has contributed to the AI research community as Datasets and Benchmarks Co-chair for NeurIPS and Social Co-chair for ICML 2023 and NeurIPS 2022. Also, he has joined a senior technical program committee member, such as, Area chair for NeurIPS 2023 and 2022, Area chair for ICML 2023, and Senior area chair for COLING. His research interests include large language models, generative models, multimodal representation learning and their practical applications for real-world problems. In particular, he has mainly focused on practical task definition and evaluation protocol for continual learning in various domains.
Jinwoo Shin (KAIST)
Related Events (a corresponding poster, oral, or spotlight)
-
2022 Poster: Time Is MattEr: Temporal Self-supervision for Video Transformers »
Tue. Jul 19th through Wed the 20th Room Hall E #105
More from the Same Authors
-
2021 : SmoothMix: Training Confidence-calibrated Smoothed Classifiers for Certified Adversarial Robustness »
Jongheon Jeong · Sejun Park · Minkyu Kim · Heung-Chang Lee · Doguk Kim · Jinwoo Shin -
2021 : Entropy Weighted Adversarial Training »
Minseon Kim · Jihoon Tack · Jinwoo Shin · Sung Ju Hwang -
2021 : Consistency Regularization for Adversarial Robustness »
Jihoon Tack · Sihyun Yu · Jongheon Jeong · Minseon Kim · Sung Ju Hwang · Jinwoo Shin -
2023 Poster: Prefer to Classify: Improving Text Classifiers via Auxiliary Preference Learning »
Jaehyung Kim · Jinwoo Shin · Dongyeop Kang -
2023 Poster: Modality-Agnostic Variational Compression of Implicit Neural Representations »
Jonathan Richard Schwarz · Jihoon Tack · Yee-Whye Teh · Jaeho Lee · Jinwoo Shin -
2023 Poster: Multi-View Masked World Models for Visual Robotic Manipulation »
Younggyo Seo · Junsu Kim · Stephen James · Kimin Lee · Jinwoo Shin · Pieter Abbeel -
2022 : Neural Architecture Search with Loss Flatness-aware Measure »
Joonhyun Jeong · Joonsang Yu · Dongyoon Han · YoungJoon Yoo -
2022 Poster: Dataset Condensation via Efficient Synthetic-Data Parameterization »
Jang-Hyun Kim · Jinuk Kim · Seong Joon Oh · Sangdoo Yun · Hwanjun Song · Joonhyun Jeong · Jung-Woo Ha · Hyun Oh Song -
2022 Spotlight: Dataset Condensation via Efficient Synthetic-Data Parameterization »
Jang-Hyun Kim · Jinuk Kim · Seong Joon Oh · Sangdoo Yun · Hwanjun Song · Joonhyun Jeong · Jung-Woo Ha · Hyun Oh Song -
2022 Poster: TSPipe: Learn from Teacher Faster with Pipelines »
Hwijoon Lim · Yechan Kim · Sukmin Yun · Jinwoo Shin · Dongsu Han -
2022 Spotlight: TSPipe: Learn from Teacher Faster with Pipelines »
Hwijoon Lim · Yechan Kim · Sukmin Yun · Jinwoo Shin · Dongsu Han -
2022 Poster: Disentangling Sources of Risk for Distributional Multi-Agent Reinforcement Learning »
Kyunghwan Son · Junsu Kim · Sungsoo Ahn · Roben Delos Reyes · Yung Yi · Jinwoo Shin -
2022 Spotlight: Disentangling Sources of Risk for Distributional Multi-Agent Reinforcement Learning »
Kyunghwan Son · Junsu Kim · Sungsoo Ahn · Roben Delos Reyes · Yung Yi · Jinwoo Shin -
2021 : Contrastive Learning for Novelty Detection »
Jinwoo Shin -
2021 Poster: Self-Improved Retrosynthetic Planning »
Junsu Kim · Sungsoo Ahn · Hankook Lee · Jinwoo Shin -
2021 Spotlight: Self-Improved Retrosynthetic Planning »
Junsu Kim · Sungsoo Ahn · Hankook Lee · Jinwoo Shin -
2021 Poster: Learning to Generate Noise for Multi-Attack Robustness »
Divyam Madaan · Jinwoo Shin · Sung Ju Hwang -
2021 Spotlight: Learning to Generate Noise for Multi-Attack Robustness »
Divyam Madaan · Jinwoo Shin · Sung Ju Hwang -
2021 Poster: State Entropy Maximization with Random Encoders for Efficient Exploration »
Younggyo Seo · Lili Chen · Jinwoo Shin · Honglak Lee · Pieter Abbeel · Kimin Lee -
2021 Spotlight: State Entropy Maximization with Random Encoders for Efficient Exploration »
Younggyo Seo · Lili Chen · Jinwoo Shin · Honglak Lee · Pieter Abbeel · Kimin Lee -
2020 Poster: Self-supervised Label Augmentation via Input Transformations »
Hankook Lee · Sung Ju Hwang · Jinwoo Shin -
2020 Poster: Context-aware Dynamics Model for Generalization in Model-Based Reinforcement Learning »
Kimin Lee · Younggyo Seo · Seunghyun Lee · Honglak Lee · Jinwoo Shin -
2020 Poster: Polynomial Tensor Sketch for Element-wise Function of Low-Rank Matrix »
Insu Han · Haim Avron · Jinwoo Shin -
2020 Poster: Learning What to Defer for Maximum Independent Sets »
Sungsoo Ahn · Younggyo Seo · Jinwoo Shin -
2020 Poster: Adversarial Neural Pruning with Latent Vulnerability Suppression »
Divyam Madaan · Jinwoo Shin · Sung Ju Hwang -
2019 Poster: Spectral Approximate Inference »
Sejun Park · Eunho Yang · Se-Young Yun · Jinwoo Shin -
2019 Poster: Robust Inference via Generative Classifiers for Handling Noisy Labels »
Kimin Lee · Sukmin Yun · Kibok Lee · Honglak Lee · Bo Li · Jinwoo Shin -
2019 Poster: Learning What and Where to Transfer »
Yunhun Jang · Hankook Lee · Sung Ju Hwang · Jinwoo Shin -
2019 Oral: Spectral Approximate Inference »
Sejun Park · Eunho Yang · Se-Young Yun · Jinwoo Shin -
2019 Oral: Robust Inference via Generative Classifiers for Handling Noisy Labels »
Kimin Lee · Sukmin Yun · Kibok Lee · Honglak Lee · Bo Li · Jinwoo Shin -
2019 Oral: Learning What and Where to Transfer »
Yunhun Jang · Hankook Lee · Sung Ju Hwang · Jinwoo Shin -
2019 Poster: Training CNNs with Selective Allocation of Channels »
Jongheon Jeong · Jinwoo Shin -
2019 Oral: Training CNNs with Selective Allocation of Channels »
Jongheon Jeong · Jinwoo Shin -
2018 Poster: Bucket Renormalization for Approximate Inference »
Sungsoo Ahn · Michael Chertkov · Adrian Weller · Jinwoo Shin -
2018 Oral: Bucket Renormalization for Approximate Inference »
Sungsoo Ahn · Michael Chertkov · Adrian Weller · Jinwoo Shin -
2017 Poster: Faster Greedy MAP Inference for Determinantal Point Processes »
Insu Han · Prabhanjan Kambadur · Kyoungsoo Park · Jinwoo Shin -
2017 Poster: Confident Multiple Choice Learning »
Kimin Lee · Changho Hwang · KyoungSoo Park · Jinwoo Shin -
2017 Talk: Confident Multiple Choice Learning »
Kimin Lee · Changho Hwang · KyoungSoo Park · Jinwoo Shin -
2017 Talk: Faster Greedy MAP Inference for Determinantal Point Processes »
Insu Han · Prabhanjan Kambadur · Kyoungsoo Park · Jinwoo Shin