REVIEW 13 cited by
OGBench: Benchmarking Offline Goal-Conditioned RL
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Offline goal-conditioned reinforcement learning (GCRL) is a major problem in reinforcement learning (RL) because it provides a simple, unsupervised, and domain-agnostic way to acquire diverse behaviors and representations from unlabeled data without rewards. Despite the importance of this setting, we lack a standard benchmark that can systematically evaluate the capabilities of offline GCRL algorithms. In this work, we propose OGBench, a new, high-quality benchmark for algorithms research in offline goal-conditioned RL. OGBench consists of 8 types of environments, 85 datasets, and reference implementations of 6 representative offline GCRL algorithms. We have designed these challenging and realistic environments and datasets to directly probe different capabilities of algorithms, such as stitching, long-horizon reasoning, and the ability to handle high-dimensional inputs and stochasticity. While representative algorithms may rank similarly on prior benchmarks, our experiments reveal stark strengths and weaknesses in these different capabilities, providing a strong foundation for building new algorithms. Project page: https://seohong.me/projects/ogbench
Forward citations
Cited by 13 Pith papers
-
Qantara: Bridge-Flow Training for Multi-Paradigm JEPA Control
A single ~21M JEPA checkpoint trained with Brownian-bridge state flow and edge-aligned action-state noise sampling serves planning, behaviour cloning, and inverse dynamics without retraining.
-
Good Rankers, Bad Objectives: Bilinear Contrastive Critics under Expressive Policy Search
Bilinear contrastive critics remain good compatibility rankers but are unsafe to maximize for action selection; cosine bounding does not fix value decalibration, while Bellman TD-Q does.
-
Do You Really Need to Pretrain Q-Functions for Online RL Fine-Tuning?
On six robot-manipulation tasks, offline Q-pretraining does not accelerate online RL fine-tuning from a pretrained policy, while seeding the replay buffer with rollouts from an ensemble of policies (IPE) improves fina...
-
Action from Adjacent Set in Physical Space Outperforms the Best Prediction in World Models
Minimum predicted cost selection fails under proposal overgeneration; reconstructing actions from adjacent low-cost prefixes (ASAR) raises Cube carry-and-release success by ~19–28 points.
-
Patch Policy: Efficient Embodied Control via Dense Visual Representations
Patch Policy shows that frozen dense ViT patch tokens, consumed through a block-causal attention mask, let lightweight robot policies beat pooled-feature policies and even a fine-tuned 7B vision-language-action model.
-
SERNF: Sample-Efficient Real-World Dexterous Policy Fine-Tuning via Action-Chunked Critics and Normalizing Flows
SERNF fine-tunes dexterous manipulation policies on real hardware by pairing normalizing-flow policies with action-chunked critics and conservative off-policy RL.
-
Compositional Diffusion with Guided Search for Long-Horizon Planning
CDGS adds population-based search and likelihood-based pruning to compositional diffusion, enabling long-horizon planning from short-horizon models across robot manipulation, panoramas, and video.
-
Benchmarking Massively Parallelized Multi-Task Reinforcement Learning for Robotics Tasks
The paper introduces MTBench, a GPU-accelerated benchmark for massively parallel multi-task RL, and reports experiments suggesting on-policy methods outperform off-policy baselines while value learning limits MTRL per...
-
VisualPatchWorld: Code World Models as Latent Structured Representations for Planning
A two-level induction procedure—active-probe sketch selection plus multi-step rollout fitting—recovers executable code world models that improve CEM planning over prior code baselines on four LeWM tasks.
-
Mollified Value Learning
Mollified Value Learning regularizes offline goal-conditioned value estimates with a Feynman-Kac expectation version of the viscous HJB equation instead of a pointwise Eikonal constraint.
-
Value Flows
Value Flows fits the full return distribution in RL with a flow-matching critic and reweights its learning objective by estimated return variance; the central theoretical guarantee does not follow from the stated equations.
-
Learning The Minimum Action Distance
A state-embedding method learns asymmetric minimum-action distances from state-only trajectories and beats existing representation methods on tested environments.
-
Information-Based Exploration via Random Features for Reinforcement Learning
Random-feature Gaussian-process information gain is turned into a closed-form exploration bonus for PPO that matches RND/VIME/#Explo on 12 control, navigation, and sparse-locomotion tasks, with error bounds on the app...
Discussion (0). Sign in to comment.