Pith Number

pith:SY66OVTC

pith:2026:SY66OVTCYN53FBFZDHRLC5CBLU

not attested not anchored not stored refs resolved

Robometer: Scaling General-Purpose Robotic Reward Models via Trajectory Comparisons

Abhishek Gupta, Abrar Anwar, Aditya Shah, Alex S. Huang, Andreea Bobu, Anqi Li, Anthony Liang, Dieter Fox, Erdem Biyik, Jesse Zhang, Jiahui Zhang, Luke Zettlemoyer, Minyoung Hwang, Sidhant Kaushik, Stephen Tu, Yigit Korkmaz, Yu Xiang

Robometer trains generalizable robot reward models by combining frame-level progress with inter-trajectory preferences.

arxiv:2603.02115 v2 · 2026-03-02 · cs.RO · cs.AI · cs.LG

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{SY66OVTCYN53FBFZDHRLC5CBLU}

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

Across benchmarks and real-world evaluations, Robometer learns more generalizable reward functions than prior methods and improves robot learning performance across a diverse set of downstream applications.

C2weakest assumption

That inter-trajectory preference supervision from comparisons imposes reliable global ordering constraints even on ambiguous suboptimal and failure trajectories without introducing significant labeling noise or bias.

C3one line summary

Robometer combines intra-trajectory progress supervision with inter-trajectory preference supervision on a 1M-trajectory dataset to learn more generalizable robotic reward functions than prior methods.

References

166 extracted · 166 resolved · 15 Pith anchors

[1] The relativity of ‘absolute’ judge- ments, 1984

[2] Absolute identification by relative judgment 2005

[3] The effect of relative encoding on memory-based judgments, 2016

[4] Rank2reward: Learning shaped reward func- tions from passive video, 2024

[5] ReWiND: Language-guided rewards teach robot policies without new demonstrations, 2025

Formal links

2 machine-checked theorem links

Cited by

8 papers in Pith

Beyond Pixels: Learning Invariant Rewards for Real-World Robotics From a Few Demonstrations

Reinforcing VLAs in Task-Agnostic World Models

ARM: Advantage Reward Modeling for Long-Horizon Manipulation

DreamAvoid: Critical-Phase Test-Time Dreaming to Avoid Failures in VLA Policies

Reinforcing VLAs in Task-Agnostic World Models

Receipt and verification

First computed	2026-05-17T23:39:15.890639Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

963de75662c37bb284b919e2b174415d2a3fb2bea17f1e521e4c02569ce876cb

Aliases

arxiv: 2603.02115 · arxiv_version: 2603.02115v2 · doi: 10.48550/arxiv.2603.02115 · pith_short_12: SY66OVTCYN53 · pith_short_16: SY66OVTCYN53FBFZ · pith_short_8: SY66OVTC

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/SY66OVTCYN53FBFZDHRLC5CBLU \
  | 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: 963de75662c37bb284b919e2b174415d2a3fb2bea17f1e521e4c02569ce876cb

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "a505b19a79e578f89d121e8e4ec6a3cd4397029daefce162deaa908f5549ae4a",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.LG"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.RO",
    "submitted_at": "2026-03-02T17:38:58Z",
    "title_canon_sha256": "4761d45d1f37188f3842b83e93a2b0103d0ebfc4a4ca08f4246107d12684c2b2"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2603.02115",
    "kind": "arxiv",
    "version": 2
  }
}