pith:WHTEK4GB
MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
Reinforcement learning lets a compressed student model surpass its teacher in multi-agent trajectory prediction.
arxiv:2604.10169 v2 · 2026-04-11 · cs.AI · cs.LG
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{WHTEK4GBH2P6Q2UYQ6LXTQKO5I}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
the framework incorporates reinforcement learning to overcome the imitation ceiling of traditional distillation, enabling the student to verify, refine, and optimize teacher knowledge through dynamic environmental interaction, potentially achieving more robust decision-making than the teacher itself.
That reinforcement learning via environmental interaction will allow the student to surpass the teacher's decision quality in real multi-agent scenarios without introducing instability, reward hacking, or extra latency that violates real-time constraints.
MAVEN-T uses a hybrid-attention teacher, efficient student, multi-granular distillation with curriculum learning, and reinforcement learning to achieve 6.2x model compression and 3.7x faster inference while matching state-of-the-art accuracy on NGSIM and highD datasets.
Receipt and verification
| First computed | 2026-06-02T01:04:16.246451Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
b1e64570c13e9fe86a98879779c14eea295e54f3d7751e6bfa141a8992dea414
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/WHTEK4GBH2P6Q2UYQ6LXTQKO5I \
| 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: b1e64570c13e9fe86a98879779c14eea295e54f3d7751e6bfa141a8992dea414
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "9ca3d6842cc9a7d5a8b0f7600cbeacf7f7fe62d65383ebc641c156ebbb312dff",
"cross_cats_sorted": [
"cs.LG"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.AI",
"submitted_at": "2026-04-11T11:34:13Z",
"title_canon_sha256": "8184cad1d30ed10c53854f8735cbeeeb66b370c0a801bf27fe29f95499f47640"
},
"schema_version": "1.0",
"source": {
"id": "2604.10169",
"kind": "arxiv",
"version": 2
}
}