pith:L3V3E4EC
StarTSE: Towards Streaming Target Speaker Extraction via Chunk-wise Interleaved Splicing of Autoregressive Language Model
Chunk-wise interleaved splicing lets autoregressive models perform stable streaming target speaker extraction.
arxiv:2604.19635 v2 · 2026-04-21 · cs.SD · cs.AI
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{L3V3E4ECXX6BV7XEEICIW47E6O}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
Our approach maintains 100% stability and superior intelligibility while achieving streaming results comparable to or even surpassing offline baselines, with RTF of 0.248.
That the Chunk-wise Interleaved Splicing Paradigm plus historical context refinement fully resolves the training-inference mismatch for autoregressive models without introducing new artifacts at chunk boundaries.
Chunk-wise Interleaved Splicing with historical context refinement lets autoregressive models deliver stable streaming target speaker extraction on Libri2Mix, matching or exceeding offline baselines at RTF 0.248.
Receipt and verification
| First computed | 2026-07-07T02:18:41.021900Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
5eebb27082bdfc1afee422048b73e4f3977b67b0c29b60a373627a733b96bc0f
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/L3V3E4ECXX6BV7XEEICIW47E6O \
| 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: 5eebb27082bdfc1afee422048b73e4f3977b67b0c29b60a373627a733b96bc0f
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "0947e4415e43a1097348e2c32ae3a479d21be1f89aa020d32dd800462577071c",
"cross_cats_sorted": [
"cs.AI"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"primary_cat": "cs.SD",
"submitted_at": "2026-04-21T16:25:22Z",
"title_canon_sha256": "06403645674df1044a0055482b07b8d8a22512ae778fbdcca77f8d19dd5ef91d"
},
"schema_version": "1.0",
"source": {
"id": "2604.19635",
"kind": "arxiv",
"version": 2
}
}