Pith. sign in

REVIEW 6 cited by

Off-Policy Deep Reinforcement Learning without Exploration

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

arxiv 1812.02900 v3 pith:PTKTTPGW submitted 2018-12-07 cs.LG cs.AIstat.ML

classification cs.LGcs.AIstat.ML
keywords learningdatareinforcementbatchdeepfixedoff-policyalgorithms
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Many practical applications of reinforcement learning constrain agents to learn from a fixed batch of data which has already been gathered, without offering further possibility for data collection. In this paper, we demonstrate that due to errors introduced by extrapolation, standard off-policy deep reinforcement learning algorithms, such as DQN and DDPG, are incapable of learning with data uncorrelated to the distribution under the current policy, making them ineffective for this fixed batch setting. We introduce a novel class of off-policy algorithms, batch-constrained reinforcement learning, which restricts the action space in order to force the agent towards behaving close to on-policy with respect to a subset of the given data. We present the first continuous control deep reinforcement learning algorithm which can learn effectively from arbitrary, fixed batch data, and empirically demonstrate the quality of its behavior in several tasks.

Discussion (0). Sign in to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Weights or Skills? A Survey of Robot-Learning Techniques: from Action-Predicting Weights to Robots that Write their Own Skills

    cs.RO 2026-08 conditional novelty 6.0 of 10

    A taxonomy of robot learning on a weights-versus-skills axis, with a five-rung self-improvement ladder whose top cell (feedback plus memory plus search) holds only a few recent systems.

  2. Good Rankers, Bad Objectives: Bilinear Contrastive Critics under Expressive Policy Search

    cs.LG 2026-07 conditional novelty 6.0 of 10

    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.

  3. Do You Really Need to Pretrain Q-Functions for Online RL Fine-Tuning?

    cs.LG 2026-07 conditional novelty 6.0 of 10

    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...

  4. Conservative Query and Adaptive Regularization for Offline RL Under Uncertainty Estimation

    cs.LG 2026-07 reject novelty 6.0 of 10

    CQ2L uses Morse-network uncertainty to select in-distribution action queries and to scale CQL's regularization, reporting higher D4RL scores than the prior OAP method.

  5. FlowQ: Energy-Guided Flow Policies for Offline Reinforcement Learning

    cs.LG 2025-05 reject novelty 6.0 of 10

    FlowQ uses energy-guided flow matching to learn an offline RL policy approximating π(a|s) ∝ πβ(a|s) exp(Q(s,a)) with guidance applied during training rather than at inference.

  6. A Forget-and-Grow Strategy for Deep Reinforcement Learning Scaling in Continuous Control

    cs.LG 2025-07 conditional novelty 5.0 of 10

    Forget and Grow (FoG) combines decaying replay weights for old experiences with progressive critic-network expansion to improve continuous-control reinforcement learning, beating BRO, SimBa, and TD-MPC2 on most of 41 ...

Pith tools