arXiv — Machine Learning · · 3 min read

From Pixels to Temporal Correlations: Learning Informative Representations for Reinforcement Learning Pre-training

Mirrored from arXiv — Machine Learning for archival readability. Support the source by reading on the original site.

Computer Science > Machine Learning

arXiv:2607.00811 (cs)
[Submitted on 1 Jul 2026]

Title:From Pixels to Temporal Correlations: Learning Informative Representations for Reinforcement Learning Pre-training

View a PDF of the paper titled From Pixels to Temporal Correlations: Learning Informative Representations for Reinforcement Learning Pre-training, by Jinwen Wang and 6 other authors
View PDF HTML (experimental)
Abstract:Unsupervised pre-training on large-scale datasets has demonstrated significant potential for improving the sample efficiency and performance of Reinforcement Learning (RL). Given the large-scale action-free internet videos, existing methods utilize single-step transition prediction and image reconstruction to learn representations. However, these methods prefer to preserve large-proportion stationary information in the pixel space, neglecting small but crucial information. To preserve enough information in the representation, it is essential to pay equal attention to each element in videos. Specifically, we propose a temporal correlation space to distinguish each element. For implementation, we introduce the Multi-scale Temporal Contrastive Learning (MTCL) method to model multi-scale temporal correlations separately. This approach can balance the attention of different elements and yield more informative representations, effectively supporting policy learning in various downstream tasks. Experimental results demonstrate that our method improves sample efficiency and asymptotic performance across various downstream tasks.
Comments: 10 pages, 8 figures. Accepted by ACM MM 2025
Subjects: Machine Learning (cs.LG)
Cite as: arXiv:2607.00811 [cs.LG]
  (or arXiv:2607.00811v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2607.00811
arXiv-issued DOI via DataCite (pending registration)
Related DOI: https://doi.org/10.1145/3746027.3755689
DOI(s) linking to related resources

Submission history

From: Jinwen Wang [view email]
[v1] Wed, 1 Jul 2026 11:38:58 UTC (3,038 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled From Pixels to Temporal Correlations: Learning Informative Representations for Reinforcement Learning Pre-training, by Jinwen Wang and 6 other authors
  • View PDF
  • HTML (experimental)
  • TeX Source

Current browse context:

cs.LG
< prev   |   next >
Change to browse by:
cs

References & Citations

Loading...

BibTeX formatted citation

loading...
Data provided by:

Bookmark

BibSonomy Reddit
Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos

Demos

Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers

Recommenders and Search Tools

Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
IArxiv recommender toggle
IArxiv Recommender (What is IArxiv?)
About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Discussion (0)

Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.

Sign in →

No comments yet. Sign in and be the first to say something.

More from arXiv — Machine Learning