Pith Number

pith:KWR3RFXI

pith:2026:KWR3RFXI5LAPRMTISHZ34OC3MP

not attested not anchored not stored refs resolved

LightSplit: Practical Privacy-Preserving Split Learning via Orthogonal Projections

Ahmad-Reza Sadeghi, Alessandro Pegoraro, Antonino Nocera, Mert Cihangiroglu, Phillip Rieger

LightSplit uses a fixed orthogonal random projection at the cut layer to cut transmitted dimensionality by up to 32 times while retaining more than 95% of baseline accuracy in split learning.

arxiv:2605.13265 v1 · 2026-05-13 · 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{KWR3RFXI5LAPRMTISHZ34OC3MP}

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 results show that the method retains more than 95% of the baseline accuracy at up to 32x reduction in transmitted dimensionality while maintaining stable training dynamics.

C2weakest assumption

That a fixed orthogonal random projection alone sufficiently restricts instance-specific information to prevent reconstruction attacks across varying projection dimensions and client scales without requiring additional mechanisms such as sparsification or noise.

C3one line summary

LightSplit uses non-invertible orthogonal projections as an information bottleneck in split learning to reduce transmitted dimensionality by 32x while retaining more than 95% accuracy and limiting reconstruction risk.

References

67 extracted · 67 resolved · 3 Pith anchors

[1] Deep learning with differential privacy 2016

[2] CONTRA: Defending against poisoning attacks in federated learning 2021

[3] The Johnson–Lindenstrauss transform itself preserves differential privacy 2012

[4] Ashish Bora, Ajil Jalal, Eric Price, and Alexandros G. Dimakis. Com- pressed sensing using generative models. InICML, 2017. 13 2017

[5] FLTrust: Byzantine-robust federated learning via trust bootstrapping 2021

Receipt and verification

First computed	2026-05-18T02:44:49.320494Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

55a3b896e8eac0f8b26891f3be385b63f4d784b59e1a81afeef9dcab15b01298

Aliases

arxiv: 2605.13265 · arxiv_version: 2605.13265v1 · doi: 10.48550/arxiv.2605.13265 · pith_short_12: KWR3RFXI5LAP · pith_short_16: KWR3RFXI5LAPRMTI · pith_short_8: KWR3RFXI

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/KWR3RFXI5LAPRMTISHZ34OC3MP \
  | 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: 55a3b896e8eac0f8b26891f3be385b63f4d784b59e1a81afeef9dcab15b01298

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "43bc4468d8efa64bc047a2b1c89fa529f5a80989e9a3944abc7a1c7de90affa9",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-13T09:45:49Z",
    "title_canon_sha256": "a1595972fd3564df05e5d2a8a5afd78df77d3709383f59f2ff1df16250ed3f94"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13265",
    "kind": "arxiv",
    "version": 1
  }
}