Pith Number

pith:BHQFDYQ2

pith:2020:BHQFDYQ2RTFWVTYAKRI7G74UWI

not attested not anchored not stored refs resolved

AWAC: Accelerating Online Reinforcement Learning with Offline Datasets

Abhishek Gupta, Ashvin Nair, Murtaza Dalal, Sergey Levine

AWAC combines offline data with online reinforcement learning to accelerate policy improvement for robotic control.

arxiv:2006.09359 v6 · 2020-06-16 · cs.LG · cs.RO · stat.ML

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{BHQFDYQ2RTFWVTYAKRI7G74UWI}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Our method, advantage weighted actor critic (AWAC), enables rapid learning of skills with a combination of prior demonstration data and online experience.

C2weakest assumption

That offline data (expert or sub-optimal) can be leveraged via AWAC to bootstrap online RL without the typical difficulties in transitioning from offline to online training remaining insurmountable.

C3one line summary

AWAC combines offline data with online RL via advantage-weighted actor-critic updates to enable faster acquisition of robotic skills such as dexterous manipulation.

References

68 extracted · 68 resolved · 1 Pith anchors

[1] Apprenticeship learning via inverse reinforcement learning 2004

[2] Maximum a Posteriori Policy Optimisation 2018

[3] An Optimistic Perspective on Ofﬂine Reinforce- ment Learning 2019

[4] ROBEL: Robotics Benchmarks for Learning with Low- Cost Robots 2019

[5] Robot Learning From Demonstration 1997

Formal links

2 machine-checked theorem links

Cited by

70 papers in Pith

Beyond One-Size-Fits-All: Diagnosis-Driven Online Reinforcement Learning with Offline Priors

Beyond Monotonic Progress: Retry-Supervised Value Learning for Robot Imitation

Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline)

DF-ExpEnse: Diffusion Filtered Exploration for Sample Efficient Finetuning

Reversal Q-Learning

Receipt and verification

First computed	2026-07-05T02:34:40.161435Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

09e051e21a8ccb6acf005451f37f94b213a0d6416acab0368a6b78956a12c20d

Aliases

arxiv: 2006.09359 · arxiv_version: 2006.09359v6 · doi: 10.48550/arxiv.2006.09359 · pith_short_12: BHQFDYQ2RTFW · pith_short_16: BHQFDYQ2RTFWVTYA · pith_short_8: BHQFDYQ2

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/BHQFDYQ2RTFWVTYAKRI7G74UWI \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 09e051e21a8ccb6acf005451f37f94b213a0d6416acab0368a6b78956a12c20d

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "7edf47029291ba308a8ace796f21d0a8e98d5b5aad6934bfaaa22dee2edff7c1",
    "cross_cats_sorted": [
      "cs.RO",
      "stat.ML"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2020-06-16T17:54:41Z",
    "title_canon_sha256": "38086df4404bc2a7ae68618bd5ae7153e26e52005051f55df074ef5d29987275"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2006.09359",
    "kind": "arxiv",
    "version": 6
  }
}