Pith Number

pith:DOONHNVW

pith:2026:DOONHNVWYAHAYHTAIGLW4ZFH6N

not attested not anchored not stored refs resolved

Phasor Memory Networks: Stable Backpropagation Through Time for Scalable Explicit Memory

Hwi-yeol Yun, Sangkeun Jung, Sungwoo Goo

Constraining recurrent states to phase rotations on the complex unit circle preserves gradient norms in explicit memory networks.

arxiv:2605.13370 v1 · 2026-05-13 · cs.LG · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{DOONHNVWYAHAYHTAIGLW4ZFH6N}

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

By constraining recurrent state updates to phase rotations on a complex unit circle, PMNet preserves gradient norms and inherently prevents divergence without the need for specialized initialization.

C2weakest assumption

That the unitary phasor constraint and hierarchical anchors will preserve sufficient expressivity and generalize from the synthetic Copy-Paste task to natural language without introducing new failure modes or requiring extensive hyperparameter tuning.

C3one line summary

PMNet uses unitary phasor dynamics and hierarchical anchors to make explicit memory stable for long sequences, matching a 3x larger Mamba model on long-context robustness with a 119M parameter network.

References

21 extracted · 21 resolved · 12 Pith anchors

[1] B., Lozhkov, A., Bakouch, E., von Werra, L., and Wolf, T 2024

[2] Arjovsky, M., Shah, A., and Bengio, Y . (2016). Unitary evolution recurrent neural networks. InInternational conference on machine learning, pages 1120–1128. PMLR 2016

[3] Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models 2026 · arXiv:2601.07372

[4] Generating Long Sequences with Sparse Transformers 2019 · arXiv:1904.10509

[5] G., Le, Q., and Salakhutdinov, R 2019

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-18T02:44:47.982749Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

1b9cd3b6b6c00e0c1e6041976e64a7f353b55c3a853f5747102a22fb5f80f694

Aliases

arxiv: 2605.13370 · arxiv_version: 2605.13370v1 · doi: 10.48550/arxiv.2605.13370 · pith_short_12: DOONHNVWYAHA · pith_short_16: DOONHNVWYAHAYHTA · pith_short_8: DOONHNVW

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/DOONHNVWYAHAYHTAIGLW4ZFH6N \
  | 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: 1b9cd3b6b6c00e0c1e6041976e64a7f353b55c3a853f5747102a22fb5f80f694

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "9fae0ee4f03281c54c1104a233a60b8bbe868caead070ebaaad1b04f0f399d49",
    "cross_cats_sorted": [
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-13T11:28:06Z",
    "title_canon_sha256": "6ff8c4f4a8e34823fb1f56d450862de71f57cf37e5e33576ffdd64a48b427a19"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13370",
    "kind": "arxiv",
    "version": 1
  }
}